The Evolutionary Ecology of Sexual Conflict and Condition-Dependence in an Insect Mating System

The evolutionary ecology of sexual conflict and condition-dependence in an insect mating system

Jennifer Christine Perry

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Ecology and Evolutionary Biology University of Toronto

The evolutionary ecology of sexual conflict and condition-dependence in an insect mating system

Jennifer Christine Perry

Doctor of Philosophy

Ecology and Evolutionary Biology University of Toronto

2010 Abstract

Sexual conflict and condition-dependent trait expression have emerged as major themes in sexual selection. There is now considerable evidence suggesting that both conflict and condition- dependence can drive the evolution and expression of sexual traits; still, important questions remain concerning the extent to which conflict shapes sexual traits and the role of condition in mediating conflict. Here, I address these two themes in studies of a ladybird mating system. One set of studies investigates the function and economics of potentially antagonistic traits – nuptial gifts and female mating resistance – while another examines condition-dependence in mating resistance and male ejaculate composition.

Nuptial gifts are often considered beneficial to females, but recent thinking suggests they may also allow males to manipulate females, raising the possibility of conflict. I demonstrate that male ladybirds benefit from nuptial feeding by their mates through reduced re-mating frequency. Benefits to female reproduction and lifespan, however, are weak or non-existent. These results show that although males gain from transferring gifts that influence female behaviour, females experience neither harm nor benefit. I next tested the hypothesis that nuptial feeding is maintained – despite an absence of benefits – because female foraging is generally elevated after mating. However, although

females indeed display strongly increased foraging after mating, this response did not increase nuptial feeding.

Recent studies suggest that individual condition may affect the economics of mating and extent of conflict. Female ladybirds vigorously resist mating, and I show that (1) resistance is condition- dependent, with low-condition females displaying more resistance, and (2) resistance functions to minimize superfluous matings (sexual conflict), rather than to select among males (indirectly benefiting females). Resistance generates selection favouring large males; thus, this work demonstrates that ecological circumstances, through influencing condition, affect the strength of sexual selection.

Finally, male condition may influence investment in ejaculate components, but condition- dependence in ejaculate composition is currently poorly understood. I show that, in agreement with theory, males in poor condition transfer smaller ejaculates that nonetheless contain more sperm, but less seminal fluid.

Taken together, this work highlights both the value of economic studies in evaluating sexual conflict, and the significance of condition-dependence for sexual selection.

iii

Acknowledgements

I am grateful to those who ably and cheerfully assisted with the experiments presented in this thesis: Alya Ahsan, Christine Heung, Kevin Sha, Diana Sharpe and Crystal Tse. I gratefully acknowledge experimental animals supplied by Suzanne Lommen and Lee Henry.

Perhaps the most enjoyable aspect of my doctoral research has been stimulating conversation with colleagues. For this I thank Aneil Agrawal, Jay Biernaskie, Amber Budden, Sean Clark, Asher Cutter, Elah Feder, Darryl Gwynne, Don Jackson, Shannon McCauley, Michael Majerus, David Punzalan, Helen Rodd, Mariana Wolfner and Minyoung Wyman, and above all my advisor Locke Rowe.

The manuscripts were vastly improved by feedback from Jay Biernaskie, Asher Cutter, Darryl Gwynne, Lucia Kwan, Tristan Long, David Punzalan, Maxence Salomon and Jessica Ward. Any errors that remain are mine alone.

I acknowledge with gratitude the members of my supervising committee, Asher Cutter and Darryl Gwynne, for their generous attention and useful criticism. I also appreciate the thoughtful feedback of the additional members of my appraisal and defence committees: Helen Rodd, Nina Wedell and Deborah McLennan.

I thank my advisor Locke Rowe, who has been more than I could have hoped for in a mentor. Thank you for your commitment to making me a better scientist.

I was supported financially by a Canada Graduate Scholarship from the Natural Sciences and Engineering Research Council of Canada (NSERC); the Dr. F.M. Hill Ontario Graduate Scholarship in Science and Technology; an Entomological Society of Canada Postgraduate Award; a W. John D. Eberlie Travel Research Grant from the Toronto Entomological Association; and a Ramsay Wright Scholarship provided by the Department of Ecology and Evolutionary Biology at the University of Toronto. My research was also supported by grants to Locke Rowe from NSERC and the Canada Research Chairs program. I thank these agencies.

It is a pleasure to warmly thank to my parents, Brian and Sharon Perry, for their support throughout my postgraduate career and all my life.

Finally I thank my partner Jay Biernaskie. I promise to listen to you carefully, consider your ideas, and be open to the many things I can learn from you.

Statement of contributions

Chapters 2, 3, 5 and 6 have been published as co-authored papers. All of these papers were co- authored with my senior advisor, Locke Rowe. In each case, we developed the study’s aims and experimental methods in collaboration; I conducted the experiments, analyzed the data and was the primary author of the manuscripts. Chapter 5 was also co-authored with Diana Sharpe. Diana refined the experimental methods and we conducted the experiments together.

Table of Contents

Acknowledgments ...... iv

List of Tables ...... ix

List of Figures ...... x

Chapter 1 General introduction ...... 1

1 General introduction ...... 1

1.1 Sexual conflict ...... 1

1.2 Condition ...... 3

1.3 Nuptial gifts and Adalia bipunctata ...... 3

1.4 The evolutionary ecology of sexual conflict and condition-dependence in Adalia bipunctata ...... 5

Chapter 2 Ingested spermatophores accelerate reproduction and increase mating resistance, but are not a source of sexual conflict ...... 8

2 Abstract ...... 8

2.1 Introduction ...... 9

2.2 Methods ...... 12

2.3 Results ...... 15

2.4 Discussion...... 17

2.5 Appendices ...... 22

2.6 Figures ...... 24

2.7 Tables ...... 27

Chapter 3 Neither mating rate nor spermatophore feeding influences longevity in a ladybird beetle ...... 28

3 Abstract ...... 28

3.1 Introduction ...... 28

3.2 Methods ...... 31

3.3 Results ...... 33 vi

3.4 Discussion...... 34

3.5 Figures ...... 37

3.6 Tables ...... 39

Chapter 4 Mating stimulates female feeding: testing the implications for the evolution of nuptial gifts ...... 41

4 Abstract ...... 41

4.1 Introduction ...... 41

4.2 Methods ...... 44

4.3 Results ...... 48

4.4 Discussion...... 51

4.5 Appendix...... 54

4.6 Figures ...... 55

4.7 Tables ...... 62

Chapter 5 Condition-dependent female remating resistance generates sexual selection on male size in a ladybird beetle ...... 64

5 Abstract ...... 64

5.1 Introduction ...... 65

5.2 Methods ...... 67

5.3 Results ...... 70

5.4 Discussion...... 71

5.5 Figures ...... 75

5.6 Tables ...... 76

Chapter 6 Condition-dependent ejaculate size and composition in a ladybird beetle ...... 78

6 Abstract ...... 78

6.1 Introduction ...... 79

6.2 Methods ...... 81

6.3 Results ...... 84

vii

6.4 Discussion...... 87

6.5 Appendix...... 91

6.6 Figures ...... 92

6.7 Tables ...... 97

Chapter 7 General conclusions ...... 101

7 General conclusions ...... 101

7.1 Major findings ...... 101

7.2 Implications and contributions ...... 104

7.3 Directions for future research ...... 106

7.4 Fin ...... 107

7.5 Tables ...... 108

Literature Cited ...... 111

viii

List of Tables

Table 2.1 ...... 27

Table 3.1 ...... 39

Table 3.2 ...... 40

Table 4.1 ...... 62

Table 4.A.1 ...... 63

Table 5.1 ...... 76

Table 5.2 ...... 77

Table 6.1 ...... 97

Table 6.2 ...... 99

Table 6.A.1 ...... 100

Table 7.1 ...... 108

List of Figures

Figure 2.1 ...... 24

Figure 2.2 ...... 25

Figure 2.3 ...... 26

Figure 3.1 ...... 38

Figure 4.1 ...... 57

Figure 4.2 ...... 58

Figure 4.3 ...... 59

Figure 4.4 ...... 60

Figure 4.5 ...... 61

Figure 5.1 ...... 75

Figure 6.1 ...... 92

Figure 6.2 ...... 93

Figure 6.3 ...... 94

Figure 6.A.1 ...... 95

Figure 6.A.2 ...... 96

x 1

Chapter 1 General introduction

Productive use of the idea of functional design, in modern biological research, often takes this form: an organism is observed to have a certain feature, and the observer wonders what good it might be.

George Williams, 1996, p. 14

1 General introduction

Sexual conflict and condition-dependent trait expression have emerged as dominant themes in recent decades of sexual selection research. This thesis presents studies on these themes, using an experimental approach and the ladybird beetle as a new model mating system.

1.1 Sexual conflict

Sexual conflict, characterized by sexually antagonistic selection, occurs when the evolutionary interests of males and females diverge and thus cannot be simultaneously achieved (Arnqvist and Rowe 2005; Parker 1979; Parker 2006; Rowe and Day 2006). In this sense, conflict must be present to some extent in virtually every sexual species. Although sexual conflict was conceptualized early in studies of sexual selection, it was long thought of as side note to the largely cooperative venture of reproduction. Only in the past several decades, beginning with the work of Parker (e.g., Parker 1979) and followed by a resurgence in the 1990s (e.g., Holland and Rice 1998; Rice 1992; Rowe et al. 1994), has the study of sexual conflict and its implications come to form a major field within evolutionary biology (for alternative views, see Cordero and Eberhard 2005; Córdoba-Aguilar and Contreras-Garduño 2003; Eberhard 2004). Studies in this area have since yielded abundant evidence that sexual conflict can lead to truly astonishing phenotypes affecting nearly all aspects of mating and reproduction. Examples include true bugs with hypodermic male genitalia that inject ejaculate directly into the female body cavity (e.g., Tatarnic et al. 2006); male waterfowl engaging in mating

2 attempts so vigorous that they drown their potential mates (McKinney et al. 1983); male ejaculate chemicals in fruit flies that function to inhibit re-mating in their mates even as they decrease their mates’ longevity (Fiumera et al. 2006); and female dance flies with inflatable sacs lining their abdomen that lure males with the illusion of high fecundity (Funk and Tallamy 2000).

An interesting implication of sexual conflict is that it can generate sexually antagonistic coevolution between male and female traits, typically envisioned as male ‘persistence’ and female ‘resistance’ (Parker 1979). Such traits may create selection on the opposite sex: male persistence selects for resistance in females, which in turn selects for increased persistence, ultimately leading to so-called chase-away sexual selection (Gavrilets et al. 2001; Holland and Rice 1998; Parker 1979; Rowe and Day 2006). Although this process is sometimes portrayed as an inevitable outcome of sexual conflict, recent theory suggests that escalating arms races are not the only expected outcome, nor even the most common (reviewed by Hosken et al. 2009); instead, there may be de-escalation of armaments, cyclical evolutionary dynamics, or the evolution of decreased female sensitivity to male persistence traits (Rowe et al. 2005). There are now a small number of well-studied cases of sexually antagonistic coevolution, including male grasping and female anti-grasping modifications in water striders (Arnqvist and Rowe 2002) and diving beetles (Bergsten and Miller 2007); spiny male genitalia and protective female reproductive tissue in flour beetles (Rönn et al. 2007); and male seminal proteins and female sensitivity to them in fruit flies (Chapman 2001; Rice et al. 2006).

Despite the great increase in interest in sexual conflict, important questions remain. Studies are needed on the significance of ecological variation for sexually antagonistic coevolution. Sexually antagonistic arms races are often conceptualized as being directionless (Arnqvist and Rowe 2005), with their outcome depending solely on features within the closed loop of mating interactions. However, ecological variation may well contribute (Chapman et al. 2003b; Fricke et al. 2009). For example, resource availability and population density will set individuals’ resource budgets (i.e., condition; section 1.2) and thus limit investment in the expression of sexual armaments. Population density additionally influences the frequency of male-female interactions, so higher densities may intensify conflict (Kimura and Ihara 2009; Kokko and Rankin 2006; Rowe et al. 1994). Furthermore, we need economic studies that evaluate benefits and costs of a wider range of mating traits and taxa, in order to clarify the importance of conflict in the evolution of mating traits (Arnqvist and Rowe 2005; Chapman et al. 2003b; Fricke et al. 2009; Rowe and Day 2006).

Answers to these questions will be of interest to many evolutionary biologists. Sexual conflict and coevolution have implications for many important topics (Arnqvist and Rowe 2005), including life history evolution (reviewed by Wedell et al. 2006), population persistence (Kokko and Rankin 2006), speciation (Gavrilets and Waxman 2002), and human health and disease (Badcock and Crespi 2008).

1.2 Condition

I follow Rowe and Houle (1996) in defining condition as the pool of resources that individuals have available to allocate to traits. As such, both environmental (e.g., resource availability) and genetic factors (e.g., alleles that influence resource-acquiring ability) contribute to condition. Condition- dependence is a key topic in sexual selection: condition dependent traits may honestly reflect male quality (Iwasa and Pomiankowski 1994; Zahavi 1975), a prediction at the core of good genes models, and condition dependence may provide a resolution to the lek paradox (Rowe and Houle 1996). Condition-dependence is well recognized in many sexual traits (reviewed by Andersson 1994; Cotton et al. 2004), but its implications extend beyond variation in trait expression. For example, in water striders, food-deprived female display more resistance to mating; large males have a mating advantage with food-deprived, resistant females, but not with well-fed, non-resistant females (Ortigosa and Rowe 2002). Thus, condition-dependence in female resistance traits means that sexual selection on male traits is also condition-dependent.

1.3 Nuptial gifts and Adalia bipunctata

This thesis combines the themes of sexual conflict and condition-dependence in studies of the two- spot ladybird, Adalia bipunctata . The next section (1.4) summarizes each chapter; here I describe two key elements in more detail: current ideas about nuptial gifts and the mating system of A. bipunctata .

1.3.1 Current thinking on nuptial gifts

A wide variety of phenotypes falls under the term ‘nuptial gift’, all featuring the transfer of edible (or 1 absorbable ) material from males to females before or during copulation. Gifts include prey items collected by male empidid flies (Preston-Mafham 1999); male body tissue and haemolymph fed on by females during copulation in sagebrush crickets (Sakaluk et al. 2004); and edible seminal products that can be large and specialized (as in katydids; Gwynne 1997) or seemingly unspecialized (as in piophilid flies; Bonduriansky 2003). In many cases nuptial gifts involve the transfer of nutrients that benefit females; for example, saliva gifts enhance fecundity in the scorpion fly Panorpa cognata (Engqvist 2007a).

By contrast, the expected nutritional benefits of gifts are surprisingly difficult to detect in other systems (see section 4.1), and in fact reduced and minimal nutritional value of gifts is predicted by theory under some conditions (e.g., under high rates of polyandry, Kura and Yoda 2001; or when elaborate gifts indicate high male quality, Sozou and Seymour 2005). Furthermore, gifts often influence female behaviour in ways not predicted by nutritional effects, such as by inhibiting female re-mating, an effect that clearly benefits males but may or may not benefit females (Gwynne 2008; Simmons and Parker 1989; Vahed 2007; Wedell et al. 2006). In other species, gifts facilitate the transfer of more ejaculate (Arnqvist and Nilsson 2000; Vahed 2007). In this case, and in the case of edible gifts derived from ejaculate, the possibility of male manipulation of female behaviour via gifts is particularly intriguing because there is a plausible mechanism: male seminal proteins are well known for their influence on female behaviour (Gillott 2003) and are in some cases harmful to females (Chapman et al. 1995). The balance of costs and benefits may be positive, negative or neutral for females, a topic of significant debate (e.g. Gwynne 2008; Sakaluk et al. 2006; Vahed 2007). In most systems, there are simply not enough data available to address this point (but see Wedell et al. 2008): many of the studies that imply fecundity benefits from nuptial feeding have not separated the effects of gifts from those of mating itself, nor have they tested the effect of gifts on a full range of female responses (e.g., fecundity, fertility, re-mating behaviour, longevity).

1 Seminal nuptial gifts are the enlarged ejaculate found in some insects that are absorbed and metabolized after deposition in the female reproductive tract (Simmons and Parker, 1989). I do not discuss them further here because it is difficult to distinguish their effects on females from those of seminal proteins that influence female behaviour.

1.3.2 Study organism

Adalia bipunctata (Coleoptera: Coccinellidae) is an arboreal aphid predator with a Holarctic distribution. Adults become sexually mature approximately five days after eclosion and live 30-60 days (Lanzoni et al. 2004). Females oviposit in batches of 10-30 eggs and can produce >500 eggs in their lifetime (Lanzoni et al. 2004). Both sexes mate multiply (on average 3-4 matings in field conditions; Haddrill et al. 2008), even though one mating can be sufficient to fertilize a female’s lifetime egg production (Hodek 1973). Females frequently resist re-mating by moving rapidly away from approaching males, kicking at or biting the feet of males that have made contact, flipping over in an apparent attempt to dislodge mounted males, and ultimately tucking their abdomen into their elytra, preventing genital contact. Males can remain mounted on a resisting female for hours. One potential cost of mating for A. bipunctata is the risk of infection by a sexually transmitted parasitic mite that reduces female fecundity (Hurst et al. 1995). Males also experience costs of both mating and ejaculate production (Perry and Tse, manuscript in preparation).

Sperm transfer occurs via a spermatophore composed of seminal fluids that solidify into a capsule within the female reproductive tract. Sperm are deposited inside the capsule and subsequently transferred to the spermatheca for storage while copulation is ongoing. Copulation duration varies 2 greatly (typically 60-150 min; the longest copulation I have observed lasted 456 min ), with sperm transfer occurring ~30 minutes into copulation. Patterns of sperm precedence are highly variable, ranging from complete first male to complete second male paternity (de Jong et al. 1998). Females eject the emptied spermatophore capsule after mating (typically within 3 minutes, but occasionally up to several hours) and most often eat it (see section 2.2.1).

1.4 The evolutionary ecology of sexual conflict and condition- dependence in Adalia bipunctata

I present a series of studies that ultimately address the question: for whose benefit are nuptial gifts, and what do they do? In general, my approach to studying mating phenotypes follows the framework outlined by Rowe and Day (2006): it includes functional and economic studies aimed at

2 Unofficial world record.

6 identifying how male and female traits of interest influence ‘shared traits’ (i.e., mating and reproductive phenotypes that the sexes unavoidably share, such as mating rate, reproductive schedule, and copulation duration); and linking those effects on shared traits to effects on male and female fitness (by measuring proxies of fitness such as longevity and reproduction). Where possible, I have measured costs and benefits in multiple contexts (Fricke et al. 2009). To study condition- dependence, I have manipulated condition using food-level treatments that were long enough and distinct enough to induce variation in mass and weight loss or gain.

I explore the potential for conflict in nuptial feeding in A. bipunctata in several studies (Chapters 2, 3, 4) that introduce this species as a new system for the study of nuptial gifts. I first identify effects of gift ingestion on several shared phenotypes in an economic study of gift ingestion (Chapter 2). I show that gifts accelerate oviposition and increase re-mating resistance. The former effect was of similar magnitude in both hungry and satiated females, implying that spermatophores act as signals rather than nutrition. Furthermore, I show that neither female lifespan nor lifetime fecundity was positively or negatively affected by nuptial feeding, implying that the effects of gifts may be neutral for females. A subsequent study (Chapter 3) builds upon these results by considering the economics of gifts and multiple mating under dietary restriction; again, neither nuptial feeding nor mating harmed or helped females.

Next, I explore a sensory bias hypothesis for the maintenance of nuptial feeding when gifts do not benefit females (Chapter 4): females may increase their feeding rate after mating (commonly observed in arthropods; Browne 1995) and may ingest nuptial gifts as a by-product of increased foraging motivation. Although females had strongly increased feeding after mating, stimulated by interactions with males, recently mated females were not more likely to eat spermatophores, indicating that an explanation for the maintenance of nuptial feeding in A. bipunctata must be sought elsewhere.

In a fourth study, I examine conflict and condition-dependence in female re-mating resistance, which in A. bipunctata can be vigorous and prolonged (Chapter 5). One hypothesis for female mating resistance is that resistance allows females to select the best male; alternatively, females may resist to avoid costly and superfluous matings. I use an experimental manipulation of female condition to distinguish between the hypotheses, and my results support the latter hypothesis. A by-product of female resistance is that large males are more successful in obtaining matings; thus, condition- dependent resistance in females influences the strength of sexual selection on male size. The study

7 also has implications for nuptial gift function: hungry females did not seek matings, implying again that nuptial gifts are not valuable as a source of nutrition.

Finally, I present a study of condition-dependence in ejaculate composition (Chapter 6). Ejaculate composition (i.e., allocation to sperm vs. seminal products) can play an important role in female fertility, sperm competition, and female post-mating responses (Gillott 2003), but little is known of the extent of condition-dependence in ejaculate composition. I show that males in poor condition transfer smaller ejaculates that contain less seminal fluid but, counter-intuitively, more sperm compared to the ejaculates of high-condition males. These results also reveal condition-dependence in the spermatophore nuptial gift.

Chapter 2 Ingested spermatophores accelerate reproduction and increase mating resistance, but are not a source of sexual conflict3

2 Abstract

Seminal products transferred during copulation can have substantial effects on females, including accelerated oviposition, decreased mating receptivity and shorter life span. This study addresses two sets of hypotheses about ingested seminal products: (1) whether they act as nutrition or have effects like those of seminal proteins and (2) whether they harm females (implying sexual conflict). I studied the ladybird beetle Adalia bipunctata , the females of which consume a spermatophore after mating. To examine the effect of spermatophore feeding on short-term reproduction, I combined a spermatophore treatment (allowing or preventing ingestion) with a diet manipulation. If spermatophores serve only as food, then low-food females are expected to show the strongest response; if spermatophores contain signalling proteins, the effect should be of similar magnitude across food treatments. Feeding on a single spermatophore affected females in two ways. The ‘allow’ group oviposited significantly faster than the ‘prevent’ group, but this response was independent of female diet, suggesting that spermatophores act as signals rather than as nutrition. In a second experiment, spermatophore consumption increased female remating resistance. In a long-term experiment, the continued ingestion of multiple spermatophores had no detectable effect on female life span, lifetime reproductive success or lifetime remating behaviour. The absence of such costs does not support the hypothesis of sexual conflict over spermatophore ingestion. Overall, the results imply that spermatophores have a signalling function but provide little, if any, nutritional value or long-term effect on fitness. Direct evidence that spermatophore function is shaped by sexually antagonistic coevolution is still lacking.

3 Published as: Perry, J.C. and L. Rowe. 2008. Ingested spermatophores accelerate reproduction and increase mating resistance but are not a source of sexual conflict. Animal Behaviour 76 : 993-1000.

2.1 Introduction

Male seminal proteins, transferred to females during copulation, can have nontrivial effects on female life history, behaviour and physiology (Wolfner 1997; Chapman 2001; Chapman & Davies 2004; Gillott 2003). At least 80 known proteins and peptides originating in male accessory glands are passed in the seminal fluids and influence egg production, oviposition, receptivity to mating and sperm storage. There is also increasing evidence that some accessory gland proteins (ACPs) have negative effects on female fitness (e.g. by decreasing life span; Chapman et al. 1995; Wigby & Chapman 2005). Hence, the evolution of seminal fluids may be influenced by sexual conflict (a conflict in the evolutionary interests of the sexes; Parker 1979, 2006) and some accessory gland products may be sexually antagonistic traits, that is, traits that further the evolutionary interests of one sex at the expense of the other (Arnqvist & Rowe 2005; Rowe & Day 2006). The sex peptide in Drosophila (Acp70A), for instance, is a well-studied case of an accessory gland product that seems to be sexually antagonistic: it elevates egg production and decreases females’ remating receptivity (Wolfner 1997; Liu & Kubli 2003) and, overall, reduces their lifetime reproductive success (Wigby & Chapman 2005).

The potential role of sexual conflict in the evolution of seminal products has motivated an increasing number of studies that re-examine behavioural traits in a framework of sexual conflict (Holland & Rice 1998; Chapman et al. 2003a; Arnqvist & Rowe 2005). As an example, it is perhaps surprisingly common for female arthropods to ingest male seminal products, typically following copulation (Eberhard 1994; Vahed 1998, 2007). The seminal products may be seemingly unspecialized for consumption (as in piophilid flies; Bonduriansky 2003) or specialized nuptial gifts derived from male accessory glands (e.g., the katydid spermatophylax; Gwynne 1997). Given that male accessory glands are the source of powerful stimulants for females, it is possible that ingested ejaculates, including seminal-derived nuptial gifts, produce effects in females similar to those caused by ACPs in the female reproductive tract (Simmons & Parker 1989; Arnqvist & Nilsson 2000; Sakaluk 2000; Sakaluk et al. 2006; Bonduriansky et al. 2005; Engqvist 2007b; Vahed 2007). I refer to such effects as signals, implying neither positive nor negative effects on female fitness. I imagine several possible mechanisms of signal transfer: ingested ejaculates may contain seminal proteins that pass through the gut wall (passage of ACP-sized proteins is physiologically possible, e.g. Jeffers et al. 2005), or they might mimic other foods and thereby induce the regular female responses to food ingestion. Alternatively, the ingested ejaculate may simply transfer calories or nutrients to females.

These two nonexclusive hypotheses – a signalling versus a nutritive function – have differing implications for the coevolution of ejaculate feeding in females and ejaculate traits in males. Sexual conflict, and the possibility of sexually antagonistic coevolution, is more likely if the signalling function exists.

To determine whether sexual conflict plays a role in ejaculate ingestion, it is necessary to test for positive or negative effects of ejaculate ingestion on fitness components such as life span and offspring production (Rowe & Day 2006). The related hypothesis, whether ingested ejaculates tend to act as a signal or as nutrition, requires experiments that manipulate ejaculate ingestion independent of mating itself, while measuring female traits commonly affected by ACPs or by nutrition. In many insects, ACPs typically induce more rapid oviposition, increased egg production, resistance to remating and shorter life span (e.g., Wolfner 1997; Chapman 2001; Chapman & Davies 2004; Gillott 2003). Food consumption is generally expected to increase oviposition, may increase or decrease resistance to remating (e.g. Gwynne 1990; Perry et al. 2009) and may increase or decrease longevity (if longevity declines with increased reproduction; e.g. Stearns 1992). To distinguish an increase in oviposition caused by signals versus nutrition, several authors have manipulated ejaculate ingestion in conjunction with a diet manipulation (Will & Sakaluk 1994; Vahed & Gilbert 1997; Bonduriansky et al. 2005; but see Eberhard 1997), predicting that low-food females will derive greater marginal benefit from any nutrition in ingested ejaculates, compared to high-food females, which will be near satiated. This is in fact the response to extra food in some insects (Moehrlin & Juliano 1998). In contrast, if ingested ejaculates function as signals, then the effect of ejaculate ingestion on female reproduction should be of similar magnitude across food levels, or there should be a greater response in high-food females because these females may have more eggs available to oviposit in response to a stimulus. A second way to distinguish the signalling and nutrition hypotheses is to assess how ejaculate feeding affects lifetime reproductive success. If seminal products represent extra food, reproduction is expected to increase, whereas if they function mostly as signals, no increase is expected.

Few experiments have manipulated the ingestion of seminal products (Vahed 2007; Gwynne 2008). These studies typically measure the response in oviposition or remating behaviour and rarely evaluate measures of lifetime fitness. The most common effect of ingesting seminal products in these studies has been a decreased willingness to remate (Simmons & Gwynne 1991; Sakaluk et al. 2006; Engqvist 2007b; but see Gwynne 1986;Wedell & Arak 1989). Several studies also detected an

11 increase in female reproductive parameters such as fecundity and egg mass following ejaculate ingestion (Gwynne 1984a, 1988; Simmons 1988, 1990; Reinhold 1999; Ono et al. 2004; Engqvist 2007a), whereas other studies detected no effect (Gwynne et al. 1984; Wedell & Arak 1989; Will & Sakaluk 1994; Vahed & Gilbert 1997; Vahed 2003; Bonduriansky et al. 2005; Mondet et al. 2008). There is one report of a positive effect on female longevity (Brown 1997) and other reports of no effect (Wedell & Arak 1989; Bonduriansky et al. 2005). Only a few studies have combined the manipulation of seminal product ingestion with a diet treatment as discussed above. None detected an interaction between the treatments for female reproductive responses (Gwynne 1988; Will & Sakaluk 1994; Vahed & Gilbert 1997; Bonduriansky et al. 2005), although in a katydid, spermatophylax ingestion increased the refractory period in low-food females but not in high-food females (Simmons & Gwynne 1991). Thus, in general, the effects on females seem to be mixed, and no study has combined an assessment of all the responses of interest with measures of lifetime fitness (Vahed 2007).

Here, I report a series of experiments aimed at determining whether ejaculates fed upon by female two-spot ladybird beetles, Adalia bipunctata (Coleoptera: Coccinellidae), act mainly as signals or as nutrition and whether consumption of ejaculates tends to increase or decrease female fitness. Female beetles eject a spermatophore shortly after copulation and consume it (Obata & Johki 1991), but the spermatophore has no obvious specialization for consumption. I first manipulated spermatophore feeding (permitting or preventing it) following a single mating and measured the short-term effects on fecundity, egg fertility, the delay until oviposition and egg mass. In this experiment, I also manipulated the recent feeding history of females to test the nutritional value of spermatophores. In a second experiment, I tested for an effect of spermatophore feeding on female resistance to remating. Finally, I examined the effect of multiple spermatophore ingestion on resistance behaviour and two proxies for female fitness: life span and lifetime fecundity. Results from this experiment provide a second test of the nutrition and signalling hypotheses and allow an assessment of extant sexual conflict over spermatophore ingestion.

2.2 Methods

2.2.1 Experimental animals

Adalia bipunctata is an aphid predator broadly distributed in North America and Europe. Both males and females mate multiply (de Jong et al. 1998). Copulation lasts 140 ± 65 min (SD; J. Perry, unpublished data). During copulation, males transfer sperm as well as seminal fluids that solidify into a spermatophore. Following copulation, most females ( ~90%) eject this spermatophore (median ejection time: 4 min, range 0-68 min), and nearly all females (94%) immediately begin to eat the ejected spermatophore; 81% consume at least half and 67% eat the entire spermatophore (J. Perry, unpublished data). My initial studies indicate that the spermatophore contains protein and some sperm when ejected. I obtained A. bipunctata larvae from Natural Insect Control (Stevensville, ON, Canada) and reared them on a diet of pea aphids ( Acyrthosiphon pisum reared on broad bean, Vicia faba ) and UV-sterilized flour moth eggs ( Ephestia kuehniella ). I reared the larvae and conducted experiments in an environmental chamber on a 16:8 h light:dark cycle at 23 ± 1 °C. During maintenance and experiments, I provided the females daily with moistened cotton as a water source.

2.2.2 Experiment 1: Short-term reproductive responses

This experiment tested the signalling versus nutrition hypotheses for ingested spermatophores with regard to several reproductive responses: the time until first oviposition after mating, egg fertility (% hatch), egg mass and fecundity. I manipulated two factors: food level and spermatophore consumption, in a 2 × 2 design.

I replicated the experiment three times. Each replicate consisted of a 4-day feeding phase, a mating trial, and 3 days of monitoring egg production. During the feeding phase, I fed virgin female A. bipunctata a low (one adult pea aphid) or high (excess pea aphids) food treatment daily for the first replicate; in the second and third replicates, I also included a medium treatment of four adult pea aphids. I established that the feeding treatment affected females by weighing a subset of females in the first replicate over the 4-day feeding period: low-food females lost mass (mean ± SE = -14 ±

2.1%), whereas high-food females maintained their weight (0.5 ± 2.0%; F1,27 = 25.9, P < 0.0001). In addition, more high-food than low-food females oviposited unfertilized eggs before the mating trial

2 (34/38 versus 2/39; χ 1 = 55.0, P < 0.0001).

For the mating trial, I mated females to a virgin male randomly chosen from the laboratory stock. Following copulation, I monitored the females for spermatophore ejection for 1 h. Upon ejection, I allowed the females to eat the spermatophore or immediately removed the spermatophore, according to the assigned treatment. For females in the former treatment, I simulated removing a spermatophore shortly after the females finished eating the spermatophore to control for any slight disturbance caused by the removal. I assigned more females to the ‘allow’ treatment because some females do not attempt to eat the spermatophore, though the likelihood of spermatophore feeding

2 did not depend on the food treatment (χ 1 = 0.29, P = 0.59). When this occurred, I assigned that female to the ‘prevent’ treatment; six females fell into this category, and excluding these females did not affect the results. I discarded females that did not eject a spermatophore within an hour and females assigned to the ‘allow’ treatment that did not eat the entire spermatophore. The total sample size was 137 females.

After the mating trial, I fed medium- and high-food females the same respective diets and low-food females two aphids daily to promote oviposition. I transferred the females to new petri dishes daily to stimulate oviposition. I monitored egg production regularly for the 3 days following mating and removed any eggs. In the first replicate, I permitted the larvae to hatch to test for a treatment effect on egg fertility. In the second and third replicates, I froze the collected egg batches ( -20 °C) and later weighed them to test for a treatment effect on egg mass.

2.2.3 Experiment 2: Remating response

In this pair of experiments, I tested for an effect of spermatophore feeding on female remating behaviour, specifically, whether females resisted males and the extent of that resistance. Resisting females kick at and move quickly away from males; if a male successfully mounts, the female tucks its posterior abdomen into the elytra, preventing genital contact. Females may resist in this manner for up to several hours until mating commences or the male dismounts. I did not manipulate diet in this experiment because the signalling and food hypotheses make opposite predictions about the effect of spermatophore ingestion on female resistance. The signalling hypothesis predicts increased resistance following spermatophore feeding, whereas other studies indicate that females that have been recently fed or maintained on a high-food diet display decreased resistance (Perry et al. 2009).

I fed females excess Ephestia eggs for several days prior to the experiment. To conduct the experiment, I initially mated females with males from the stock population. After copulation, I

14 removed the male from the petri dish and allowed the female to eat or prevented it from eating the spermatophore. I then exposed the female to a second nonvirgin ‘test’ male 1 h later and recorded the amount of time the female spent resisting the male.

The two experiments differed in how long I monitored female resistance. In the first experiment, I paired the test male and the female for 1 h and recorded resistance behaviour. Twenty-five of 29 females mated within an hour; I excluded the four that did not mate. For the second version of the experiment, I altered the experimental design for two reasons: (1) I attempted to standardize the duration of male mating effort the females were exposed to and (2) I was concerned that females could not evade males within the confines of a petri dish and would eventually acquiesce to mating. In this second version, I monitored the mating effort shown by the test male. If the male did not attempt to mount the female within 6 min, or if more than 6 min elapsed between mounting attempts, the test male was replaced. I continued replacing males until mating occurred or until the females experienced a cumulative total of 15 min of mounting attempts from males. At this point, I reasoned that in a natural setting, females might have successfully rejected a male. This difference between the experiments means that resistance times could range from 0 to 60 min in experiment 1 and from 0 to 15 min in experiment 2.

I excluded females from analysis based on several criteria: failure to eject a spermatophore, failure to consume the entire spermatophore in the spermatophore feeding treatment, and oviposition during the period of exposure to the second male. The final sample size was 58.

2.2.4 Experiment 3: Lifetime effects

I tested for an effect of repeated spermatophore ingestion on two proxies for female fitness, lifetime egg production and longevity. I maintained females subjected to the medium food treatment in experiment 1 (replicates 2 and 3) on the medium diet and mated them every 3 or 4 days for their life span. I chose the medium food level so that females had sufficient food for continued oviposition but were not satiated. At each mating, I manipulated spermatophore feeding according to the original assigned treatment. I counted and collected eggs daily. The total sample size was 44.

Conducting this experiment allowed me to assess remating resistance following spermatophore feeding on a longer time scale than was tested in experiment 2, i.e. several days after spermatophore ingestion rather than 1 h. I recorded resistance behaviour for 27 mating sessions.

2.3 Results

2.3.1 Reproductive responses in the short term

Because there were differences in the three replicates of experiment 1 – medium food was not included in the first replicate – I had two options for analysis of the two responses measured in all replicates, the time until first oviposition and fecundity. I could include the low- and high-food treatments of all three replicates or include all food levels but restrict the analysis to replicates 2 and 3. I chose the latter approach as it included more information. None the less, the pattern of results was the same with the former approach (see Appendix 2.A). I included ‘replicate’ as a random blocking variable with the two treatments (food level and spermatophore feeding) in the analyses. Replicate and its interactions were not significant terms in any analysis.

Both elevated food and spermatophore ingestion advanced the time of first oviposition following copulation (square-root-transformed data: food: F 2,94 = 19.2, P < 0.001; spermatophore: F 1,94 = 4.1, P = 0.045; Fig. 2.1). High-food females tended to oviposit within hours of mating (mean of 1.9 h, back-transformed), whereas low-food females took longer than a day (27.4 h), on average. Averaged across food levels, oviposition was advanced several hours in those females that ate a spermatophore compared to females that did not (7.6 h compared to 14.2 h, back-transformed). In contrast to the prediction of the nutrition hypothesis, there was no significant interaction between food level and spermatophore consumption (F 2,94 = 0.8, P = 0.48; Fig. 2.1). To evaluate my ability to detect an interaction, I compared the effect size of the spermatophore ingestion treatment at each food level. Low-food females had the largest absolute response to spermatophores (a decrease of 3.2 h, 95% CI -0.3 to 16.8), compared to 0.2 h (-4.0 to 7.8) for medium food and 0.7 h (-2 to 3.7) for high food. It may, however, be more appropriate to consider the relative effect size. High-food females had the largest relative response, ovipositing 46% faster after eating a spermatophore, compared to medium-food (12%) or low-food females (30%).

Fecundity, measured in the 3 days following copulation, was strongly influenced by food level, but

not by spermatophore feeding (food: F 2,96 = 73.2, P < 0.001; spermatophore: F 1,96 = 0.3, P = 0.59; interaction: F 2,96 = 0.3, P = 0.73; Fig. 2.2). Females produced 79 ± 15 eggs in the high-food treatment and only 15 ± 15 eggs in the low-food treatment. The absence of any effect of

16 spermatophore feeding on fecundity over this 3-day period suggests that its effect was only on the timing of oviposition (see above), not on the quantity oviposited, and that the effect was short-lived.

I did not detect a difference in hatching success caused by either food level or spermatophore ingestion (food: F 1,31 = 1.8, P = 0.19; spermatophore: F 1,31 = 0.6, P = 0.44; interaction: F 1,31 = 0.5, P = 0.47; hatching success was arcsine-transformed and weighted by the number of eggs). In females allowed to eat versus those prevented from eating the spermatophore, the proportion of hatching larvae was 0.73 ± 0.15 versus 0.54 ± 0.20 for low-food females and 0.80 ± 0.14 versus 0.80 ± 0.14 for high-food females. Female A. bipunctata can harbour male-killing bacteria (Hurst et al. 1992) and a phenotypic indicator of infection is a low hatching rate. With this in mind, I performed a second analysis excluding females with a hatch rate of ≤ 60%. The pattern of results did not change (see Appendix 2.B). Similarly, I did not detect a difference in egg mass from either treatment (square-

root transformed: food: F2,75 = 2.9, P = 0.06; spermatophore: F 1,75 = 0.6, P = 0.43; interaction: F 2,75 = 0.1, P = 0.91). In females allowed to eat versus prevented from eating the spermatophore, the mass per egg was 112 ± 0.03 mg versus 108 ± 0.05 mg for low-food females and 114 ± 0.06 mg versus 112 ± 0.03 mg for high-food females.

2.3.2 Remating behaviour

Spermatophore feeding increased the likelihood that a female resisted remating. Fifteen of 19 spermatophore-feeding females resisted remating, compared to 20 of 39 females prevented from 2 spermatophore feeding (χ 1 = 4.1, P = 0.043).

To test the effect of spermatophore feeding on the duration of resistance, I first set those resistance times from the first replicate that exceeded 15 min to exactly 15 min, to make the results from the two versions of the experiment comparable. This procedure makes the test more conservative. I used the nonparametric Wilcoxon test because the residuals from an ANOVA were not normally distributed and could not be corrected by transforming the data. Spermatophore ingestion increased the duration of female remating resistance by more than twofold (T = 2.3, N = 58, P = 0.0248; Fig.

2.3). This pattern remained when I analysed resistance time using parametric statistics (F 1,56 = 4.0, P = 0.0495).

2.3.3 Lifetime effects

I did not detect any effects of multiple spermatophore ingestion on life span (using a proportional hazards model; Table 2.1). I lost two females partway through the experiment and included them as censored data. To address the possibility that a spermatophore effect might accumulate over a female’s lifetime, I conducted a second analysis that excluded females that died relatively early 2 (before the first quartile of life span, 38 days); however, the results did not change (χ 1 = 0.55; P = 0.55). Including the number of eggs as a covariate did not affect the pattern of results in either analysis.

After controlling for life span, I found no effect of the ingestion of multiple spermatophores on lifetime fecundity, whether females that died before the first quartile of life span were included (Table 2.1) or excluded.

Females allowed to eat spermatophores were just as likely to resist matings as females prevented from eating any spermatophores (F 1,42 = 0.8; P = 0.37), after controlling for the number of mating sessions. Females that ate spermatophores showed mating resistance during 9.8 ± 0.4 mating sessions, compared to 9.3 ± 0.4 sessions for females prevented from spermatophore feeding. Females also spent similar amounts of time resisting males, regardless of their treatment group (Table 2.1).I re-analysed these data excluding females that died before the first quartile of life span; again, the pattern of results did not change.

2.4 Discussion

With increasing understanding of the effects of ACPs on females, there is a growing interest in the hypothesis that seminal products ingested by females transfer signals (in addition to or in place of nutrition), raising the possibility that sexual conflict plays a role in their evolution (Simmons & Parker 1989; Arnqvist & Nilsson 2000; Arnqvist & Rowe 2005; Sakaluk 2000; Sakaluk et al. 2006; Bonduriansky et al. 2005; Engqvist 2007b; Vahed 2007). In this study I found that ingesting spermatophores affects female ladybird beetles in two ways: by advancing females’ oviposition schedule (without affecting fecundity) and by increasing remating resistance. Both effects are more consistent with a signalling than with a food function, and the absence of any effects on fecundity in the short or long term likewise challenges the food hypothesis. Although there seems to be a

18 signalling function, I found no evidence supporting the hypothesis that sexual conflict is currently involved in spermatophore ingestion. Below, I evaluate the support that the data provide for these conclusions.

2.4.1 Spermatophores as nutrition versus signals

The ingestion of a spermatophore caused accelerated oviposition, and, importantly, this effect seems independent of female nutritional condition. If there were significant nutritional value to the spermatophore, then the effect of spermatophore feeding should be greatest in low-food females (Bonduriansky et al. 2005); on the other hand, if spermatophores contain signals that influence females in a way distinct from food, then a spermatophore effect should be of similar magnitude across food levels. The strong effect of food treatment on oviposition delay (see Fig. 2.1) does indicate that nutrition limits the oviposition schedule under these experimental conditions. Yet the result that the spermatophore effect was statistically independent of food level is consistent with a signalling function. Clearly, it is important to question my ability to detect an interaction between food level and spermatophore consumption. Although the confidence intervals of the effect size at each food level (that is, the average difference in oviposition delay between the ‘allow’ and the ‘prevent’ groups) overlap, the largest absolute decrease in oviposition time occurred at low food. The largest relative effect size, however, was measured at the high-food level. I acknowledge that it is relatively easy (in a statistical sense) to accept the null hypothesis of no interaction; accordingly, I refrain from basing a conclusion on this result alone. Another important outcome that supports the signalling hypothesis is the immediacy of the spermatophore effect: high-food females oviposited sooner, even within hours of copulation, much sooner than they could have generated eggs from additional food. It is also noteworthy that ACPs are well known to accelerate oviposition in insects (Eberhard 1996; Vahed 1998; Gillott 2003; Chapman & Davies 2004). A similar advance in oviposition schedule has emerged from an ejaculate feeding study of piophilid flies maintained at two food levels (Bonduriansky et al. 2005).

The absence of a spermatophore effect on fecundity, measured in the short term or over a female’s lifetime, provides further evidence that spermatophores do not contribute significant nutritional value. The results of the short-term experiment clearly imply that fecundity is limited by food (see Fig. 2.2); however, ingesting a spermatophore did not permit females to produce more eggs. A potential concern is whether the short-term experiment was sensitive enough to detect an effect, given that females were allowed to eat only one spermatophore. However, in the long-term

19 experiment females ate on average eight spermatophores in the spermatophore feeding treatment versus zero in the prevented treatment. In other systems, increased ejaculate ingestion or nuptial feeding has indeed resulted in increased fecundity when isolated from the effect of mating on fecundity (Gwynne 1984a, 1988; Simmons 1990), although this effect is not universal (Gwynne et al. 1984; Wedell & Arak 1989; Will & Sakaluk 1994; Vahed & Gilbert 1997; Vahed 2003; Ono et al. 2004; Bonduriansky et al. 2005; Mondet et al. 2008).

Females that consumed a spermatophore also showed an immediate decline in receptivity to mating, which is not expected under the food hypothesis. This response is not expected because other studies suggest that food makes female A. bipunctata more willing to mate in both the short and the long term (Perry et al. 2009). The increase in remating resistance is consistent with a variety of other species by which ejaculate-derived material is ingested (Vahed 1998, 2007; Gwynne 2008). It is also consistent with the effects of ACPs transferred during copulation in Drosophila (Wolfner 1997; Chapman & Davies 2004). The decrease in receptivity in the short term was associated with increased female resistance to male mating attempts, rather than a decrease in male attempts to mate, suggesting that males were not reacting to some scent or other signal of female mating status (reviewed by Simmons 2001).

In contrast to the result from the short-term experiment, females that fed on multiple spermatophores throughout their lives were no more or less resistant to remating than females prevented from spermatophore feeding. In this long-term experiment, females were offered another mating opportunity 3 to 4 days after each previous mating; thus, it seems that the influence of spermatophores on remating behaviour diminishes with time. This pattern is consistent with the action of ACPs on Drosophila remating behaviour: ACPs induce refraction in the day following mating, but the effect disappears by the second day (reviewed by Wolfner 1997; Chapman & Davies 2004).

Last, I note that no effect of spermatophore feeding on egg mass or hatch rate was detected, a result that is in accord with prior studies that manipulated ingestion of seminal products (e.g. Wedell & Arak 1989; Mondet et al. 2008). The absence of a response in these traits does not argue as strongly against spermatophores as nutrition as do the fecundity data, because I also observed no effect of the feeding treatment on these variables.

2.4.2 Spermatophore ingestion and sexual conflict

The effects of spermatophore ingestion in the short term – advancement of the oviposition schedule and decreased remating receptivity – are certainly likely to benefit a male that transfers a spermatophore. Faster oviposition will benefit males if females deposit eggs before remating, and this is a genuine risk because female A. bipunctata do mate multiply (Brakefield 1984; de Jong et al. 1998; Webberley et al. 2006) and because stored sperm is mixed in the spermatheca (de Jong et al. 1998). If we assume a male benefit from spermatophore feeding by their mates, the occurrence of sexual conflict depends on the costs of spermatophore ingestion for females. Such costs are plausible, for example, through an energetic cost to faster oviposition (Simmons & Parker 1989) or a decrease in a female’s optimal mating rate (Arnqvist & Nilsson 2000). Ultimately, though, direct evidence for the role of conflict in the maintenance of spermatophore production and feeding should include evidence that spermatophores are antagonistic to female fitness. This is the approach I have taken, and the data do not support the existence of such costs: both female life span and lifetime fecundity were unaffected by multiple spermatophore feeding events. In a separate study, I tested for a spermatophore effect on longevity independent of any effect on reproduction by using nutrient-stressed females (which do not oviposit); again, ingesting spermatophores had no effect on longevity (Perry & Rowe 2008b). My finding that spermatophore ingestion does not lead to increased longevity is consistent with studies of other insects in which consumption of ejaculate material was manipulated directly (Wedell & Arak 1989; Bonduriansky et al. 2005; see also Omkar & Mishra 2005).

These data do not provide evidence that sexual conflict has been involved in the coevolution of the spermatophore and its consumption. But theory and data suggest that sexual conflict may be hidden by cycles of adaptation and counteradaptation between the sexes (Chapman & Partridge 1996; Rice 1998; Arnqvist & Rowe 2005) and so this issue requires further study. Indeed, there is evidence of cryptic sexual conflict in a nuptial-feeding system: the spermatophylax of the house cricket, Gryllodes sigillatus , has no effect on G. sigillatus females but induces decreased mating receptivity when offered to females of related species that do not normally have nuptial feeding (Sakaluk et al. 2006). A similar comparative experimental approach in these ladybirds could prove informative.

2.4.3 Conclusion

There are many species in which females eject and consume seemingly unspecialized ejaculate material. It is noteworthy that in two such species, the piophilid flies (Bonduriansky et al. 2005) and the ladybird beetles studied here, there is now evidence that ingested ejaculates affect females in a manner distinct from nutrition, implying a form of signalling. These systems with ejaculate ingestion, and potentially systems with nuptial gifts, represent previously unknown pathways by which males can access female regulatory systems and influence female behaviour; the extent to which these effects are detrimental to females is still unknown.

2.5 Appendices

2.5.1 Appendix 2.A: Supplementary analysis for experiment 1: the interval between mating and first oviposition

In experiment 1, I tested for an effect of food level and spermatophore ingestion on the interval between mating and first oviposition. The experiment was replicated three times, with two food levels (high and low) in the first replicate and three (including medium) in the second and third replicates. In the main paper, I analyzed the data including the three food levels of replicates 2 and 3, in order to include more information. Here I present a second analysis that tests the effect of high vs. low food only in all three replicates. The term replicate was included in the model as a random blocking variable.

The qualitative pattern of results is similar to that presented in the paper. Both elevated food level and spermatophore ingestion decreased the interval between mating and first oviposition (food,

F1,103 = 89.0, P < 0.0001; spermatophore, F 1,103 = 4.1, P = 0.04). Again, there was no interaction between the treatments (F 1,103 = 0.2, P = 0.62). Oviposition was more rapid in high than low food females (2.9 h vs. 37.2 h, back-transformed least squares means) and in females that ate a spermatophore compared to those that did not (11.6 h vs. 19.1 h). Low-food females experienced a larger absolute advancement in oviposition from spermatophore feeding compared to high-food females (difference in means: low-food, 1.4 h, -0.3 – 8.4 (95% CI); high-food, 0.7 h, -1.0 – 6.2). However, high-food females experienced a larger relative advancement in oviposition from spermatophores (36% compared to 18% for low-food females).

2.5.2 Appendix 2.B: Supplementary analysis for experiment 1: hatching success

I tested for an effect of food level (high or low) and spermatophore ingestion on hatching success. Because female A. bipunctata can harbour male-killing bacteria, which causes approximately 50% hatching failure, I re-analyzed the data excluding females with a hatch rate of ≤ 60%, to be conservative. The proportion of hatching eggs was arcsine-transformed and weighted by the number of eggs.

Thirteen females had fewer than 60% of their eggs hatch overall. When I excluded these females the pattern of results is similar to the initial analysis. Hatching success was not affected by either food or

spermatophores (food, F 1,21 = 1.0, P = 0.32; spermatophore, F 1,21 = 2.0, P = 0.18; interaction, F 1,21 = 0.0, P = 0.95). At low food, the proportion of hatching eggs for females allowed vs. prevented from eating a spermatophore was 0.94 ± 0.11 vs. 0.87 ± 0.17, and 0.89 ± 0.08 vs. 0.81 ± 0.08 for high food females.

2.6 Figures

Figure 2.1

The effect of spermatophore ingestion (allowed or prevented) and food level (low, medium or high) on the interval between mating and first oviposition. Bars represent standard errors.

Figure 2.2

Effects of food level and spermatophore ingestion on oviposition in the 72 h following copulation. Bars represent standard error.

Figure 2.3

Effect of spermatophore ingestion on the extent of female resistance to a second mating. Least- squares means are presented, with bars indicating standard error.

2.7 Tables

Table 2.1

A lifetime of spermatophore ingestion (or prevention from spermatophore ingestion) had little effect on life span, lifetime fecundity or remating behaviour

Spermatophore feeding

Response Allowed Prevented Effect size and CI 1 Statistic P

2 Life span 69 days 59 days 9.1 (-9.8 to 28.1) χ 1 0.39

Lifetime fecundity 460 eggs 454 eggs 5.8 (-138 to 149) F1,41 = 0.0 0.89

Average duration 17 min 21 min -4.1 (-2.7 to 0.4) F1,41 = 0.4 0.54 of remating resistance*

1 CI: confidence interval

2 Analysis performed on square-root-transformed data; back-transformed least-squares means presented.

Chapter 3 Neither mating rate nor spermatophore feeding influences longevity in a ladybird beetle 4

3 Abstract

Females of many species experience costs associated with mating. Seminal products, including nuptial gifts, may mitigate these mating costs or exacerbate them. For example, nuptial gifts derived from male accessory glands may transfer nutrition or potentially harmful seminal proteins to females. In this study, I assay the costs of multiple mating and the consumption of seminal products in a ladybird beetle. I compared longevity in females mated singly or multiply, while allowing or preventing spermatophore consumption at each mating. In order to distinguish a cost of mating per se from a cost of elevated reproduction, I prevented reproduction by using nutrient-stressed females. Mating singly or multiply had no effect on female longevity, nor did spermatophore feeding influence longevity. The results imply, first, that intermediate mating rates do not directly harm females, though females may experience other indirect costs of mating (e.g. reduced foraging efficiency) or costs of reproduction; and second, that spermatophores transfer neither food nor directly harmful substances to female ladybirds.

3.1 Introduction

Much current research in sexual selection is centred on the conflicting evolutionary interests of males and females (Arnqvist & Rowe 2005; Andersson & Simmons 2006; Kokko et al. 2006). Sexual conflict is defined as a conflict between the evolutionary interests of individuals of the two sexes, and this conflict may often result in sexually antagonistic selection on traits that are part of the phenotypes of both sexes (Parker 1979, 2006; Rowe & Day 2006). For example, increasing mating rate may often be favoured in males but disfavoured in females (Arnqvist & Rowe 2005). Several

4 Published as: Perry, J.C. and L. Rowe. 2008. Neither mating rate nor spermatophore feeding influences longevity in a ladybird beetle. Ethology 114 : 504-511. Publisher: Wiley-Blackwell. Copyright © 2008 The Authors.

29 reviews have highlighted the need for more empirical investigation of the economics of shared traits, such as mating rate, so that we can evaluate the extent to which these traits are shaped by sexually antagonistic selection (Chapman et al. 2003b; Zeh & Zeh 2003; Arnqvist & Rowe 2005; Kokko et al. 2006; Rowe & Day 2006; Vahed 2007).

Costs of mating to females have been documented in a number of taxa (reviewed in Arnqvist & Rowe 2005), and may include elevated predation and parasitism, reduced foraging success, genital damage and a variety of life history adjustments caused by signals from male seminal products (e.g. Rowe 1994; Wolfner 1997; Crudgington & Siva-Jothy 2000; Baer et al. 2001; Blanckenhorn et al. 2002). A recent meta-analysis of insect polyandry suggests that there is generally an intermediate optimal mating rate for females, resulting from increased offspring production and decreased longevity with multiple mating, and furthermore that observed mating rates may often exceed the optimum (Arnqvist & Nilsson 2000). However, in those cases where mating includes nuptial gifts, increased mating rates tend to increase reproductive success while having little or a weak positive effect on longevity (Arnqvist & Nilsson 2000). A reasonable inference is that consumption of these seminal products tends to offset any longevity costs associated with mating, and is on average a net benefit to females.

In spite of this broad pattern, another line of evidence suggests that nuptial gifts may impose costs on females. Nuptial gifts are often produced by male accessory glands and there is reason to believe that accessory gland products may not always be beneficial to females. For example, there is increasing evidence that some seminal products delivered to the female reproductive tract reduce female fitness. In Drosophila melanogaster , components of male accessory gland proteins (Acps) in the seminal fluid (e.g. sex peptide) reduce female lifetime reproductive success (Chapman et al. 1995; Wigby & Chapman 2005). Likewise, in species with nuptial gifts derived from male accessory glands, transferred products may carry some costs to females (Simmons & Parker 1989; Sakaluk 2000; Bonduriansky et al. 2005; Sakaluk et al. 2006; Engqvist 2007b). This interesting possibility has led to calls for more study of the fitness consequences of nuptial feeding for females (Vahed 1998, 2007; Arnqvist & Rowe 2005; Gwynne 2008).

Although there is a large number of studies of nuptial feeding on components of female fitness (for reviews, see Vahed 1998, 2007; Gwynne 2008), these studies typically manipulate mating rate and thereby nuptial feeding, rather than nuptial feeding alone. As discussed above, there is a generally positive effect of an increased mating rate in these studies (Arnqvist & Nilsson 2000). There are

30 fewer studies that manipulate nuptial feeding independently of mating, and the results of these do not always lead to the conclusion of a net benefit to females. Although several report positive effects of nuptial feeding on female reproductive parameters (e.g. orthopterans, Gwynne 1984a, 1988; Simmons 1988, 1990; Kasuya & Sato 1998; Reinhold 1999; Ono et al. 2004; scorpionflies, Engqvist 2007a), several other studies report no effect on parameters such as fecundity and egg mass (in orthopterans, Gwynne et al. 1984; Wedell & Arak 1989; Will & Sakaluk 1994; Vahed & Gilbert 1997; Vahed 2003; piophilid flies, Bonduriansky et al. 2005; cockroaches, Mondet et al. 2008; coccinellid beetles, Perry & Rowe 2008a). One study found a positive effect on female longevity (tree cricket, Brown 1997), while others detected no effect (wartbiters, Wedell & Arak 1989; piophilid flies, Bonduriansky et al. 2005; cockroaches, Mondet et al. 2008; coccinellid beetles, Perry & Rowe 2008a). Only one of these studies independently compared the effects of mating and nuptial feeding rates on female fitness (Vahed 2003).

I examined the effect of multiple mating and spermatophore feeding on female longevity in the two- spot ladybird beetle ( Adalia bipunctata ). Males of the two-spot ladybird beetle transfer a proteinaceous spermatophore during copulation, which females eject and consume after mating. Data for two other ladybird species that do not ingest spermatophores suggest a longevity cost of increased number of matings, possibly due to increased oviposition with increased mating (Omkar & Srivastava 2002; Omkar & Mishra 2005; Omkar et al. 2006; but see Haddrill et al. 2007). My previous work on A. bipunctata indicates no long-term benefit to females of repeated spermatophore feeding independent of mating rate (own data). However, that study, like most previous studies of the effect of nuptial feeding on females, was conducted under relatively benign laboratory conditions of sufficient food. There is good reason to expect that any costs (or benefits) of mating and nuptial feeding be exaggerated under stressful conditions. For example, a nutritional effect of nuptial feeding may be detectable only when females are food-limited, when the marginal benefit would be higher. To address this shortcoming, I conducted a study of the economics of mating and nuptial feeding in a highly food-stressed environment. Because I deprived females of food during the experiment, I expected to see costs or benefits of nuptial feeding that might be masked in more benign environments. Moreover, because food deprivation prevented egg production, any observed depression in longevity can be attributed to harmful substances in the gift or ejaculate, or other interactions with the male at mating, rather than to an increase in oviposition induced by mating.

3.2 Methods

3.2.1 Species

Adalia bipunctata (Coleoptera: Coccinellidae) is an aphidophagous predator broadly distributed across temperate habitats. Following most copulations (>90%), females eject a hollow spermatophore and most females (>90%) immediately ingest the spermatophore.

I obtained several hundred first and second instar A. bipunctata (Natural Insect Control, Stevensville, ON, Canada) and reared them to adulthood on a combined diet of pea aphids ( Acyrthosiphon pisum reared on broad bean, Vicia faba ) and UV-sterilized Ephestia kuehniella eggs (Beneficial Insectary, Redding, CA, USA). Larvae and adults were housed in an environmental chamber on a 16:8 h dark:light cycle at 23°C ± 1°C. Adults used in the experiment were at least 12 d post-emergence. Adults were fed aphids ad libitum until 1 d prior to the initial mating trial; on that day each female was fed four adult aphids and males continued on the ad libitum diet. Throughout the experiment, females were housed individually in petri dishes (50 mm × 12 mm) and transferred to new petri dishes daily.

3.2.2 Experimental procedure

Females were randomly assigned to the two crossed experimental treatments: number of matings (one, three or five), and spermatophore feeding (allowed or prevented at each mating). Females were mated once or twice per day until the assigned number of matings was attained, which occurred within 5 d. Although females may mate many more times over a lifetime in a natural setting, the mating rates I used are within the range of rates observed in the field (0.1–0.5 copulations ⁄ d, Webberley et al. 2006; see also Brakefield 1984). The males used in the experiment had not mated for at least 2 d to ensure time to replenish ejaculate stores. If a male failed to mount a female within 1 h, it was discarded and a new male introduced. I ignored copulations that lasted <30 min (19 occurrences out of 433) because I have seldom observed spermatophore transfer in such brief matings. After each copulation ended, I monitored females for spermatophore ejection and, according to the assigned treatment either removed the spermatophore immediately or allowed females to ingest it. For females in the latter treatment, I simulated removing a spermatophore shortly after females finished eating the spermatophore to control for any slight disturbance to females caused by the removal.

After mating, and between matings, females were housed in the environmental chamber under the conditions described and provided with cotton dampened with 50 µl of water daily. Females were deprived of food from the first day of mating, except on day 3, when I fed each female a large pea aphid to stimulate mating receptivity. Beginning on the first day of matings, I checked for survival at 0900, 1300 and 1700 h daily until all females perished. I also noted and removed any eggs laid. Depriving females of food stopped virtually all oviposition by day 3 of the mating portion of the experiment. Forty-eight females oviposited in the first 2 d, whereas only four females oviposited from day 3 onward and only two laid eggs after mating was completed on day 5.

3.2.3 Analyses

To test for an effect of the number of matings, spermatophore feeding, and the interaction between these two factors on female survival, I used a proportional hazards model, with the response variable being hours survived since the beginning of the experiment.

I initially excluded females based on three criteria: (1) failure to mate the number of times assigned; (2) failure to eject spermatophores and (3) failure to eat all of the ejected spermatophore when assigned to the ‘allow’ treatment. The initial sample sizes for the mating treatment levels of 1, 3 and 5 were 29, 31 and 31 females for the ‘allow’ spermatophore feeding treatment and 20, 24 and 20 for the ‘prevent’ treatment. After excluding females, the final sample sizes for those respective treatment groups were 10, 11 and 4 females vs. 14, 24 and 16 females. The sample size was reduced to four for females mating five times and feeding on spermatophores because females rarely ejected and ate five entire spermatophores with five matings. The small sample size in this group reduced my ability to meaningfully test the hypotheses, and for this reason I performed a second analysis in which I grouped females into those that mated singly vs. multiply. This allowed me to include additional females that mated two or four times, and the sample sizes for these pooled data were, for females mating singly vs. multiply, 10 vs. 23 for females allowed to eat the spermatophore and 14 vs. 48 for prevented females. The mean number of matings in the multiply mated group was 3.7 ± 0.1 SE. I report statistics from the first analysis because this was the planned approach, but focus on results from the second test.

Least squares means are reported with SE. Analyses were conducted using JMP 5.0 (SAS Institute).

3.3 Results

One female survived much longer than the others (720 h vs. x¯ = 161 h ± 28). To avoid having an extreme outlier, I set survival for this individual to the next longest survival time (280 h).

When I analysed survival for females that mated one, three or five times, neither feeding on 2 spermatophores nor the number of matings influenced female survival (spermatophore feeding: χ 1 2 2 = 0.03, P = 0.70; number of mates: χ 2 = 0.84, P = 0.66; interaction: χ 2 = 0.9, P = 0.64). The 95% confidence intervals for each treatment group overlap broadly and thus give no indication of a trend (spermatophore feeding allowed, one mating: 139–174 h; three matings: 145–178 h; five matings: 141–196 h; spermatophore feeding prevented, one mating: 148– 178 h; three matings: 152–175 h; five matings: 144– 172 h).

When I grouped females that mated multiply and compared their survival to that of females mated 2 singly, there was again no effect from spermatophore feeding (χ 1 = 0.75, P = 0.39) or the number of 2 2 matings (χ 1 = 0.05, P = 0.82), or their interaction (χ 1 = 0.11, P = 0.74; Fig. 3.1, Table 3.1; results were similar from a reduced model dropping the non-significant interaction term). Because these results were null, it is of interest to estimate the magnitude of treatment effect that I can confidently exclude at the α = 0.05 level, given the data (e.g. Hoenig & Heisey 2001). I did so by constructing confidence intervals for the differences between groups (Table 3.2). For example, at α = 0.05, this experiment could have detected a decrease in longevity larger than 18% or an increase larger than 11%, between singly and multiply mated females that ingested spermatophores (Table 3.2).

All females mated at least once, but some females refused subsequent matings (six females originally assigned to five matings). If mating refusal is governed by female condition and there is a relationship between condition and longevity, then excluding these females could bias the outcome (Rönn et al. 2006). To test for this possibility, I conducted a Student’s t-test on survival times based on whether females completed their assigned number of matings. There was no difference in survival between the groups for females assigned to five matings ( x = 160.1 h ± 6.0 for females that

completed the mating treatment; x = 160.0 h ± 26.1 for females that refused matings; F 1,18 = 0.0, P = 0.99).

3.4 Discussion

The results suggest that there are no detectable costs of either multiple mating or nuptial feeding for female longevity in this ladybird beetle, even when beetles are nutritionally stressed. I deprived females of food during the experiment, an approach that offered two advantages. First, females deprived of food quickly cease oviposition and oogenesis, thus nullifying a potential trade-off between reproduction and longevity. Earlier studies demonstrated that consumption of spermatophores advanced oviposition in time (Perry & Rowe 2008a), and this may itself elevate mortality (Stearns 1992). Second, the chance of detecting either a longevity benefit or cost of nuptial feeding or mating should be maximized under conditions of nutritional stress, and nutritional stress is a common feature of these beetles’ environment (Sloggett & Majerus 2000; Evans 2003). I discuss this puzzling behaviour of nuptial feeding in coccinellid beetles in the context of other studies of nuptial feeding and mating costs.

Given the extent to which nuptial gifts have been characterized as gifts of nutrition from males to females (Vahed 1998, 2007; Gwynne 2008), it is initially surprising that female ladybirds should gain no apparent nutritional benefit from nuptial feeding, even under conditions of extreme nutritional duress. The results suggest that any benefit for female longevity from spermatophore ingestion must be smaller than 12% for a single spermatophore or smaller than 6% for multiple spermatophores (see Table 3.2). Yet, despite the overall weakly positive relation between female longevity and mating rate in nuptial feeding arthropods (Arnqvist & Nilsson 2000), the finding of no apparent nutritional benefit from nuptial feeding is in fact common for studies that manipulate nuptial feeding independent of mating rate (reviewed by Boggs 1995; Vahed 1998, 2007; Arnqvist & Rowe 2005; Gwynne 2008). What does this imply about the fitness value of nuptial feeding for females?

One possible explanation is that nuptial feeding offers minute benefits not detected by experimenters but favoured by selection. With several studies that fail to detect any benefit (reviewed by Vahed 1998, 2007; but see Gwynne 2008), however, this hypothesis appears increasingly unlikely to be general. Sakaluk (2000) proposed an alternative explanation: females are selected in a non-mating context to eat items with certain chemical or physical properties, and nuptial gifts that exploit this ‘gustatory response’ are favoured by selection. The nuptial gift of a decorated cricket appears designed for this purpose: it has a distinct amino acid profile high in free amino acids, which can be tasted, and low in essential amino acids (Warwick et al. 2009), making the spermatophylax nuptial gift an irresistibly delicious, albeit worthless, food item (Sakaluk et al. 2006).

Likewise there is evidence that females’ foraging instinct maintains nuptial feeding in a spider (Bilde et al. 2007). Currently, there is no detailed data available on spermatophore composition in coccinellid beetles, and the hypothesis that the ladybird spermatophore has evolved to exploit a gustatory response remains untested.

I have earlier raised the alternative that spermatophores might contain signals that are detrimental to females, such that nuptial feeding decreases longevity. The data do not support this hypothesis. The possibility remains, though, that nuptial gifts transfer signalling proteins that affect female reproduction or re-mating behaviour, and any costs or benefits of such effects would not have been detected in this study. Previous work has shown that eating a single spermatophore advances oviposition by approx. 1 d, compared to females prevented from eating a spermatophore, but that overall fecundity does not increase from spermatophore feeding (Perry & Rowe 2008a). It is not known whether this shift in reproductive schedule is beneficial or costly to females in the wild.

I have also found that females become more resistant to re-mating in the first day after spermatophore feeding, though the effect disappears within 3 d (Perry & Rowe 2008a). Males are likely to benefit from increased re-mating resistance in their mates if it leads to reduced sperm competition. The current study suggests that a decreased mating rate will have minimal direct costs or benefits to female longevity.

Although I found no indication of a longevity cost for females due to additional copulations (any such cost must have been smaller than 18% for spermatophore-feeding females or 15% for nonfeeding females; see Table 3.2), the experiment was designed to detect only direct costs arising from mating. It remains possible that female A. bipunctata do experience other costs of mating that the study was not designed to detect, such as reduced foraging success or increased predation rate during mating (e.g. Rowe 1994), increased transmission of pathogens (e.g. Hurst et al. 1995) or the cost of replacing stored sperm with sperm from a less desirable mate (Arnqvist & Rowe 2005). Costs like these could account for the fact that females often resist mating vigorously and at length (Majerus 1994b). Finally, because females were deprived of food after the mating treatment, costs arising from a trade-off between reproduction and longevity were removed; there is, in fact, evidence of such a trade-off in two coccinellid species (Omkar & Mishra 2005). If Acps transferred from males induce increased or accelerated oogenesis in females, females may experience decreased longevity as a result, and this study would not detect this type of indirect mating cost.

A recent study suggested that spermatophore feeding might mitigate mating costs in another ladybird. Omkar & Mishra (2005) manipulated mating rate in three ladybirds and found increased fecundity and decreased longevity with increased mating in two species that do not eject spermatophores ( Propylea dissecta and Cheilomenes sexmaculata ) but no effect in the third species (Coccinella septempunctata ). Although the authors propose that spermatophore feeding in C. septempunctata might explain the absence of costs from multiple mating, I think this unlikely for two reasons. First, spermatophore feeding is infrequent (Obata & Johki 1991) or absent (Omkar & Srivastava 2002; own data) in this species. Second, both female and male P. dissecta and C. sexmaculata had depressed longevity from multiple mating; in contrast, longevity was not affected by multiple mating in either male or female C. septempunctata , but the hypothesis that spermatophore feeding accounts for the absence of reduced longevity addresses females only.

An important question is whether the mating rates I tested were high enough to educe mating costs. Although there is no data on the optimum mating rate for female A. bipunctata , I note that the intermediate and elevated mating rates I tested are within the range of rates reported from field observations (Brakefield 1984; Webberley et al. 2006). Furthermore, many features of the laboratory environment in this study were similar to the natural conditions experienced by A. bipunctata : temperature, humidity and the aphid diet provided before the mating treatment. Thus, the experimental setting should have detected realistic direct fitness costs to females from additional copulations.

Nuptial gifts have often been described as male donations with direct benefits to females; in contrast to this view, these results contribute to a growing appreciation of the possibility of little or no direct benefit of nuptial feeding to females in some species. In contrast to related coccinellid species, A. bipunctata females appear to experience no direct longevity costs of mating for moderate mating rates. A next step for understanding nuptial feeding in ladybird beetles is to address the maintenance of spermatophore feeding in spite of its minimal contribution to female fitness.

3.5 Figures

Figure 3.1

Survival of females mated singly (a) or multiply (b), after eating a spermatophore at each mating (solid line) or not (dotted line). Because the lines overlap extensively, the confidence intervals are not presented for visual clarity.

3.6 Tables

Table 3.1

Survival of females mated singly or multiply and either allowed to or prevented from eating a spermatophore at each mating

Treatment combination Survival (h) 1 SE

Single mating, spermatophore eaten 156.7 11.5

Single mating, no spermatophore 162.6 9.7

Multiple matings, spermatophores eaten 151.2 7.6

Multiple matings, no spermatophore 156.3 5.2

1 The survival values are least squares means from a 2 × 2 ANOVA.

Table 3.2

The difference in longevity between females subjected to spermatophore feeding and mating rate treatments, given that neither treatment had a statistically significant effect on longevity (least squares means; negative values indicate a lower mean for the spermatophore feeding and multiple mating groups, compared to non-feeding and singly mating females)

Comparison

Spermatophore feeding Single Multiple (SF) No SF mating mating

Single vs. multiple Single vs. SF vs. no SF vs. no mating multiple mating SF SF

Difference in -5.5 -6.2 -5.9 -5.1 means (h)

Confidence -18%, 11% -15%, 7.3% -19%, 12% -13%, 6% interval 1

1 Any treatment effect outside the confidence interval can be ruled out at the α = 0.05 level (Hoenig & Heisey 2001).

Chapter 4 Mating stimulates female feeding: testing the implications for the evolution of nuptial gifts

4 Abstract

Nuptial gifts provide nutritional benefit to females in some species, but a growing number of studies in other species have failed to detect such benefits. These studies raise an interesting question: what maintains nuptial feeding if gifts do not benefit females? The sensory bias hypothesis proposes that the origin and maintenance of nuptial feeding may be explained by pre-existing sensory responses in females that predispose them to consuming gifts, such as sensitivity to particular compounds, which males can in turn exploit. More generally, recent studies have shown that in some species female feeding rate is elevated after mating and that this response is induced by male seminal proteins. This non-specific feeding response may represent an unexplored aspect of sensory bias for nuptial feeding, if increased motivation to feed leads to increased intake of nuptial gifts along with other foods. I tested these ideas in a beetle in which females ingest a spermatophore after mating. I show that females display a strong post-mating increase in feeding that is stimulated by males. However, there was no evidence for the dose-dependence in feeding in response to additional matings that is predicted by a sexual conflict hypothesis. Moreover, although females ate more after mating, this response cannot explain nuptial feeding in this species: newly mated females were in fact less likely than non-mated females to ingest spermatophore gifts they encountered. I discuss alternative hypotheses for the maintenance of nuptial feeding in this species.

4.1 Introduction

Nuptial gifts are well-known mating phenotypes in which males transfer edible material to females, typically as a form of mating effort (reviewed by Gwynne 2008; Vahed 2007). In some insects, a large and calorie-rich nuptial gift is synthesized by males, and there is evidence that ingestion of these gifts increases female reproductive output (particularly in some orthopterans, e.g. Brown 1997; Gwynne 1984a; Simmons 1988; see also Engqvist 2007a). Yet it is surprisingly difficult to detect the

42 expected benefits of gifts in other species. Examples with no detectable benefits include other orthopteran species (Vahed 2003; Vahed & Gilbert 1997; Wedell & Arak 1989; Will & Sakaluk 1994), as well as cockroaches (Mondet et al. 2008) and coccinellid beetles (Perry & Rowe 2008a). Furthermore, one frequently reported effect of gift ingestion is to increase female re-mating resistance (e.g., Engqvist 2007b; Perry & Rowe 2008a; Simmons & Gwynne 1991), a response which is clearly in the interests of males but which may or may not benefit females. It is therefore of interest to ask what maintains gift ingestion in these species despite an apparent lack of nutritional benefit.

A possible explanation is offered by the sensory bias hypothesis for the evolution of nuptial gifts, which proposes that gifts exploit pre-existing female responses that have evolved by natural selection in another context (Christy 1995; Endler & Basolo 1998; Fuller et al . 2005; West-Eberhard 1984), particularly foraging (Sakaluk 2000; reviewed by Vahed 2007). This hypothesis has largely been discussed with regard to the phagostimulatory properties of gifts (e.g., Bilde et al. 2007; Sakaluk 2000; Sakaluk et al. 2006; Vahed 2007); for example, nuptial gifts in the cricket Gryllodes sigillatus are loaded with free amino acids that are enticing to females rather than nutritious (Warwick et al. 2009; see also Wada-Katsumata et al. 2009). New studies have identified a female response to mating that may also have implications for the evolution of nuptial gifts: in some arthropods, male seminal proteins in the ejaculate stimulate female feeding after mating (Barnes et al. 2008; Carvalho et al. 2006; Kaufman 2007). From the male perspective, increased female motivation to feed may represent an opportunity to entice females to eat a nuptial gift, which in turn would benefit males if gifts affect females in ways that benefit males (e.g., increasing re-mating resistance). Thus, a nonspecific feeding response may increase the likelihood of gift ingestion after mating. This hypothesis has several interesting implications. It is a mechanism that might facilitate the evolution of nuptial gifts in many species because post-mating increases in foraging may be widespread (Browne 1993; Kaufman 2007). It also suggests that nuptial feeding can be maintained by selection favouring the general post-mating feeding response (e.g., increased feeding may be favoured to support reproduction), even if gift ingestion is neutral or even costly for females (see Arnqvist 2006; Arnqvist & Rowe 2005; Vahed 2007). This could occur if it is difficult for selection to separate nuptial feeding after mating from a general motivation to forage.

I investigated the hypothesis that females eat nuptial gifts after mating because mating induces a general feeding response, using a ladybird ( Adalia bipunctata ) that has been well-characterized in

43 terms of the effects of mating and nuptial gifts on females (Haddrill et al. 2007; Majerus 1994a; Perry & Rowe 2008a, b). In this beetle, females eject and generally eat a spermatophore capsule after mating and do not appear to gain increased fecundity or longevity from doing so (Perry & Rowe 2008a, b; Perry et al. 2009). In the present study, I first tested for a post-mating elevation of feeding in the period in which nuptial feeding occurs, in both food-deprived and satiated females. I then examined several proximate factors that may be responsible for increased feeding. Increased feeding might be a consequence of a general increase in activity after mating (e.g., leading to increased encounters with food; see Isaac et al. 2009), or may result from females attempting to recoup energy expended in mating. Alternatively, the link between mating and feeding may be indirect if mating stimulates egg production which in turn stimulates feeding (see Barnes et al. 2008). Lastly, increased feeding may be a response to sperm or seminal proteins transferred during copulation, as observed in Drosophila melanogaster (e.g., Carvalho et al. 2006).

Having found evidence of a post-mating increase in female feeding that was mediated by interactions with males, I tested the dose-dependence of the feeding response to explore two alternative hypotheses for how the response has evolved, by sexual conflict or cooperation. If sexual conflict has played a significant role, the feeding response should exhibit dose-dependence, by the following logic (see Arnqvist 2006; Arnqvist & Nilsson 2000; Arnqvist & Rowe 2005; Engqvist 2007b; Johnstone & Keller 2000). Consider a species in which females regulate their reproduction and re-mating behaviour by an internal stimulus (e.g., a hormonal pathway), and increased female response would increase male fitness. Males would then be selected to provide an extra dose of such a stimulus to boost the female response. If the response was at the optimum level for females before male input, then the male stimulus will move the response above the female optimum. Females will be selected to reduce their sensitivity to the male stimulus, which (it is thought) will create selection on males to provide even greater stimulus. The expected outcome of this process is dose- dependence in the female response. An alternative line of thought suggests that the female responses males induce during mating are beneficial for both males and females. Under this scenario, mating induces reproductive responses that ‘switch on’ and remain on (Ravi Ram & Wolfner 2007), producing a pattern of no dose-dependence with additional matings. These two hypotheses describe simplistic scenarios, and alternatives can certainly be imagined; still, they provide a starting point.

Finally, I tested whether the post-mating general feeding response actually results in increased feeding on spermatophore nuptial gifts.

4.2 Methods

4.2.1 Experimental animals

Adalia bipunctata is an aphid predator with Holarctic distribution. Females can lay 10-30 eggs daily in abundant food conditions (Stewart et al. 1991). Both sexes mate multiply. During mating (x¯ duration 140 min ±65 SD, Perry & Rowe 2008a), males transfer sperm via a spermatophore capsule composed of seminal fluids that solidify within the female reproductive tract. Shortly after mating (typically 2-4 minutes), females usually eject (90% of matings) and ingest the spermatophore capsule (>90% ingest at least a portion, while two thirds ingest the entire capsule; Perry & Rowe 2008a). The beetles used in this study were of the third generation from a laboratory stock initially obtained from a supply company (Natural Insect Control, Stevensville, ON, Canada). Adult beetles were fed excess UV-sterilized flour moth eggs ( Ephestia kuehniella , a standard diet, de Clerq et al. 2005). A previous study has found that A. bipunctata larvae consume these eggs at a rate that is indistinguishable from the rate for aphid prey (Jalali et al. 2009).

4.2.2 General experimental procedures

I conducted six experiments; five involved a mating or non-mating (control) treatment and a feeding trial as described below.

4.2.2.1 Mating and control treatments

Each experiment involved randomly assigning initially virgin females to a mating or non-mating treatment. I made every effort to handle non-mating females identically as for mated females. In most of the experiments, pairs of mating and non-mating females moved through the phases of the experiment as a unit to control for variation in copulation duration (e.g., when a mating female completed copulation and spermatophore ejection, this female and the paired non-mating female were both placed in feeding arenas). To begin each experiment, females were placed in new petri dishes and a male was introduced to the dish of the ‘mating’ female using a fine paintbrush. The action of introducing a male (removing the petri dish lid and brushing a paint brush on the dish) was simulated for the non-mating female. When copulation ended, I immediately removed the male and simulated this removal for the non-mating female. Mated females were monitored for ejection of the

45 spermatophore capsule, which was removed; non-mating females were subjected to a sham spermatophore removal.

4.2.2.2 Measuring feeding

I measured food ingestion by providing individuals with a known mass of flour moth eggs (hereafter ‘food’) and re-weighing the remaining food after a period of feeding. For each group of mating and non-mating individuals, the feeding trial began immediately following spermatophore ejection by the mated female. Each individual was introduced to a feeding arena consisting of an aluminum weighing boat (containing the pre-weighed food) covered with a petri dish lid lined with filter paper. Along with the feeding arenas assigned to the experimental individuals, a control arena with food only was set up to account for changes in food mass in the absence of feeding. I simulated the introduction and removal of a beetle from this control arena. I calculated food ingestion by subtracting the final mass of food from the initial mass, and subtracting from this the mean mass loss of food in control dishes. To examine the relationship between feeding and egg-laying, I recorded the number of eggs laid in each feeding arena. Feeding trials took place at 21°C ± 1 C and 50% humidity. The duration of the trials varied, as described below.

4.2.3 Experiment 1: Post-mating feeding behaviour in males and females

This experiment was designed to establish whether females display a sex-specific post-mating increase in feeding which might lead to feeding on spermatophores, and to test whether any change in feeding represents a specific foraging response or is (1) part of a general post-mating increase in activity, or (2) indirectly stimulated by a post-mating increase in reproduction. I measured food ingestion at two points: 250 min after copulation and spermatophore ejection, and from that point until 24 h after copulation. The first feeding arenas initially contained 5.00 mg ± 10% of food, while the second arenas contained 10.0 mg ± 10%.

I conducted the experiment by setting up groups of two male-female pairs and assigning each pair to the mating or non-mating treatment by a coin toss; 25 groups were tested. Here the non-mating male was handled as for the mating male (e.g., lifted with a paintbrush). During the first period, I assessed beetle activity beginning 10 minutes into the feeding trial and subsequently every 30 minutes, using the following index: 0 (no movement); 1 (slow forward or side to side movement); 2

(slow and steady forward movement); 3 (rapid forward movement). I scored activity blind to the assigned treatments.

4.2.4 Experiment 2: The schedule of the post-mating feeding response in food-deprived and satiated females

Here the goal was (1) to establish the schedule of the post-mating feeding response, and in particular to test whether it occurs immediately after mating when females normally would encounter spermatophores, and (2) to test whether the effect of mating on feeding occurs in hungry females as well as the well-fed females used in experiment 1. I manipulated female state (hungry or satiated) and mating status in three time periods after mating and spermatophore ejection: within 30 minutes, from 30 to 120 minutes, or 120 to 240 minutes. The feeding arenas in each period initially contained 1.0 mg ± 10%, 2.0 mg ±10%, and 2.5 mg ± 10% of food, respectively. To generate hungry and satiated females, I placed initially virgin females on a ‘no food’ or ad libitum diet 48 h before the experiment.

4.2.5 Experiment 3: Is the feeding response an attempt to recoup energy expended in copulation?

I tested whether increased female feeding post-mating is an effort to recoup energy expended in carrying males during copulation, by randomly assigning females to one of three treatments: mating, weight-bearing, or a control. Groups of three females (each assigned to a different treatment) moved through the phases of the experiment as a unit to control for copulation duration. To begin, all females in a group were lightly anesthetized with CO 2. For the weight-bearing treatment, I affixed a lead weight, equivalent in mass to a large male (13.1 mg ± 10%, the third quartile of mass for a sample of laboratory males), to the female’s elytra in a similar position to that of a mating male, using a small piece of double-sided tape. I made every effort to handle the control and mating females exactly as for the weight-bearing treatment. When copulation ended for the mating female within each group, I removed the male and removed the weight and tape from the other treatments within the group (simulating this for the mated female). Each female was transferred to an initial feeding arena containing 5.00 mg ± 10% of food for 250 minutes and then to a second arena containing 10.0 mg ± 10% of food until 24 h total.

4.2.6 Experiment 4: The feeding response and transfer of sperm and seminal proteins

In this experiment I aimed to test whether sperm or seminal protein transfer, if either, causes increased female feeding. Copulation in A. bipunctata is characterized by early transfer of seminal fluids (including those that form the spermatophore; Haddrill 2001; Ransford 1997); seminal proteins can be detected in the female’s reproductive tract after only a minute of copulation (unpublished data). In contrast, sperm transfer begins at least 30 minutes after mating is initiated (Haddrill 2001; Ransford 1997). I randomly assigned females to one of seven treatments: a non- mating treatment, a complete mating, or interrupted mating occurring at 1, 15, 30, 60 or 120 minutes. I monitored mated females for spermatophore ejection for up to 1 h after mating, and after 1 h transferred the females to a feeding arena containing 12.0 mg ± 10% of food for 24 h. As a check on sperm transfer, I retained eggs from each female and monitored them for hatching success.

4.2.7 Experiment 5: Multiple mating and the dose dependence of the feeding response

To test the hypothesis that female feeding is further elevated after a second mating (as predicted if there has been sexual coevolution over control of female feeding), I assigned females to an initial mating or non-mating treatment and then to a second mating or non-mating treatment 9 days later, such that females experienced 0, 1 or 2 matings. This interval between matings is at least as long as natural mating intervals in this species (Haddrill et al. 2008). I measured female feeding in the 24 hours following the second mating (or control) treatment, in feeding arenas containing 13.0 mg ± 10% of food.

4.2.8 Experiment 6: Does increased feeding after mating translate into increased nuptial feeding?

Having detected a post-mating elevation in feeding, I conducted an experiment to directly test the hypothesis that this feeding response explains nuptial feeding on spermatophores. If the hypothesis is correct, then newly mated females will be more likely than non-mated females to ingest spermatophores they encounter. I presented recently mated and non-mated females with a spermatophore (within minutes of spermatophore ejection by the mated female) and monitored their feeding. As above, mated females were not permitted to eat their own ejected spermatophores after mating. To obtain a supply of spermatophores, I collected ejected spermatophores from additional mating pairs immediately before the experiment and stored them in small humidity

48 chambers to prevent drying (on a dry surface within a petri dish containing a saturated salt solution; Winston & Bates 1960). Spermatophores were used in the experiment within 1 h of collection.

To conduct the experiment, at three minutes after spermatophore ejection (to allow recovery from the disturbance of spermatophore removal), I transferred paired mated and non-mated females to individual 1.5 mL microtubes with a small hole in the cap and a fresh spermatophore inside. I monitored females continuously for one hour and recorded whether the spermatophore was wholly or partially eaten.

4.2.9 Statistical analyses

In general, I used one- or two-way mixed model ANOVAs to analyze the effect of mating, along with any other experimental treatments and the interaction term, on food ingestion and the number of eggs laid (by females that laid ≥ 1 egg); most models included the random effect ‘group’. I tested whether continuous responses met the assumptions of parametric statistics and where necessary applied a transformation or used a non-parametric test. I report least squares means ± S.E., back- transformed where appropriate. I analyzed categorical response variables by chi-square tests.

In experiment 1, I tested for a relationship between egg-laying and feeding using a mixed model ANCOVA, including mating, the number of eggs and group. To examine the effect of mating and sex on activity, I conducted a two-way ANOVA using the summed activity scores (0-27) as the response variable. I then tested for a correlation between the activity score and food ingestion.

4.3 Results

4.3.1 Experiment 1: Effects of mating on feeding

I measured food ingestion by mated and non-mated males and females over two time periods. Because results were similar within each time period, I present the total food ingestion.

Both sex and mating had substantial influence on feeding (Figure 4.1a; sex: F1,92 = 31.8, P < 0.0001;

mating: F1,92 = 6.3, P = 0.01; square-root transformed data), and there was a marginally significant

sex x mating interaction effect (F1,92 = 3.9, P = 0.05). Mated females ate 46% more than non-mated females, while there was minimal difference in feeding by mated and non-mated males (Figure 4.1a).

There was no indication that increased feeding was attributable to increased activity after mating: activity rates were similar regardless of sex or mating (Figure 4.1b; sex: F 1,96 = 0.2, P = 0.68; mating:

F1,96 = 0.9, P = 0.34; sex x mating: F 1,96 = 0.7, P = 0.39; square-root transformed data), and there was no correlation between activity rate and feeding (R2 = 0.03, P = 0.09).

In contrast, feeding was strongly and positively correlated with egg-laying (Figure 4.1c; ANCOVA:

F1,44 = 32.9, P < 0.0001), which was itself also stimulated by mating. Mated females were more likely 2 to lay eggs (23/25 vs. 11/25; λ 1 = 12.8, P = 0.0004) and tended to lay more eggs, although this

difference was not statistically significant (22.4 ± 2.8 vs. 14.9 ± 4.1, F 1,32 = 2.9, P = 0.11). The variation in egg-laying could entirely account for the effect of mating on feeding: when variation in the number of eggs was controlled for in an ANCOVA, mating itself no longer had explanatory power (F 1,44 = 2.8, P = 0.10; no significant interaction, F1,44 = 2.3, P = 0.14).

4.3.2 Experiment 2: Female nutritional state and the schedule of the post-mating feeding response

Here I tested the feeding rates of satiated or food-deprived females at three points soon after mating. I found weak evidence of increased feeding by mated females within 30 minutes of mating, and strong evidence of increased feeding by two and four hours after mating (Figure 4.2, Table 4.1). Hungry females ate more than satiated females, but the effect of mating on feeding was similar in both groups (Figure 4.2, Table 4.1).

4.3.3 Experiment 3: Energy expended in copulation and post-mating food ingestion

I tested whether increased female feeding could be explained by energy expended in carrying males during copulation by subjecting females to a mating, weight-bearing or control treatment, and measuring food ingestion over two time periods. Both periods yielded similar results and so the total feeding is presented here. Contrary to the prediction of this hypothesis, there was no evidence that artificially weighted females behaved like mated females (Figure 4.3). Mated females ate more than both weighted and control females, which fed at similar rates (F 1,117 = 6.6, P = 0.002).

4.3.4 Experiment 4: The feeding response and transfer of sperm and seminal proteins

Here I examined whether seminal fluid or sperm transfer could account for increased female feeding by subjecting females to matings of different durations.

Females that mated for any length of time ate more than control females that did not mate, although only matings of 60 minutes or longer resulted in significantly increased feeding (Figure 4.4; F6,281 = 2.7, P = 0.01; Box Cox transformation). As a post hoc test of whether the feeding response is induced after only one minute of mating, I compared the one-minute, non-mating and complete mating treatments together in a separate ANOVA. Female feeding after one minute of mating was indistinguishable from feeding after a complete mating (6.1 ± 1.0 vs. 5.7 ± 1.0), and both were

elevated compared with non-mating females (4.5 ± 1.0 mg; F2,281 = 5.9, P = 0.004).

Mating duration influenced egg fertilization in a manner consistent with previous reports that sperm transfer typically does not begin until 30 minutes into copulation (see section 4.2.6; Appendix 7A, Table 4.A.1).

4.3.5 Experiment 5: Multiple mating and dose dependence in the feeding response

I tested the hypothesis that males can further stimulate female feeding with additional matings by subjecting initially virgin or previously mated females to a second mating or no mating. The results provide little support for this sort of dose dependence: previously mated females mated females had only a slight (10%) and non-significant increase in feeding after a second mating, compared to previously mated females that did not mate a second time (Figure 4.5). This contrasts with the large (65%) and significant increase that initially virgin females experienced after mating compared to non-mating females (Figure 4.5).

4.3.6 Experiment 6: The feeding response and nuptial feeding

I tested whether elevated feeding in recently mated females did in fact increase the likelihood of ingesting spermatophore nuptial gifts. This prediction was not supported; in fact, non-mated females were more likely than recently mated females to eat part or all of the spermatophore (part: 21/23 vs. 2 2 14/23, λ 1 = 6.2, P = 0.01; all: 16/23 vs. 9/23, λ 1 =4.4, P = 0.04). To assess whether this pattern was explained by decreased female activity soon after mating, I examined the activity scores from experiment 1 at 10 minutes after mating; however, mated and non-mated females displayed similar activity levels (respectively, medians: 1 (range 0-2), control 1 (range 0-3); means: 0.8 ± 0.2 vs. 0.9 ± 2 0.2; Wilcoxon test: λ 1 = 0.0, P = 0.83).

4.4 Discussion

I found strong evidence that females experience a rapid and substantial post-mating increase in feeding that is initiated by mating interactions with males. Taken as a whole, the data are consistent with the hypothesis that this feeding response is part of a cascade of reproductive responses (including oviposition) that are switched on by mating, as opposed to alternative hypotheses that males stimulate feeding to a level above the female optimum or that increased feeding is a byproduct of other responses to mating such as increased post-mating activity. However, the results strongly contradict the hypothesis that this post-mating feeding response accounts for spermatophore feeding: recently mated females were in fact less likely to ingest spermatophore nuptial gifts compared with non-mated females. Below, I discuss evidence for each of these claims.

4.4.1 Female feeding is elevated after mating

Increased female feeding after mating was consistent throughout this study (Figures 4.1-4.5; Table 4.1). The response was specific to females: mating did not influence male feeding within 24 hours of mating (Figure 4.1a), a result that is surprising because both mating and ejaculate production are costly for male A. bipunctata (Perry & Tse, unpublished data). Increased feeding began within two hours of mating and possibly as soon as 30 minutes (Figure 4.2); thus, feeding appears to be elevated in the period in which females normally encounter gifts (within minutes to an hour of mating). Mating stimulated feeding independent of female nutritional state (Figure 4.2), suggesting that the response is not simply an artefact of satiated or food-deprived laboratory females. Moreover, the feeding response could not be attributed to energy expended in carrying males during copulation (Figure 4.3) – although it is possible that mating generates more energetic costs than simply bearing a weight – nor to increased female activity after mating (Figure 4.1b), implying that females have a specific foraging response. This response was tightly linked to increased egg-laying after mating (Figure 4.1c); in fact, the increase in egg-laying could completely account for the feeding response (experiment 1). A reasonable hypothesis is thus that mating stimulates egg production, which in turn stimulates feeding; this is the case in D. melanogaster (Barnes et al. 2008).

The results strongly suggest that an element of interactions with males stimulates female feeding (potentially indirectly, as noted above). In particular, it is noteworthy that feeding appeared to be initiated after only one minute of mating (experiment 4) because seminal proteins are detectable in the female reproductive tract after one minute of mating (unpublished data). In turn, receiving

52 sperm cannot explain increasing feeding because sperm is not transferred until 30 minutes into copulation. It is also unlikely that female stretch receptors in the reproductive tract produce the response because a minimum of 20 min of mating is required for transfer of the entire spermatophore (Ransford 1997). The data are consistent with the male stimulus being seminal proteins or physical contact with males, including mechanical stimulation from the aedeagus. Studies from other arthropods suggest that seminal proteins are a likely candidate (Carvalho et al. 2006; Kaufman 2007).

4.4.2 A simple signal or male manipulation?

I found little evidence for the dose dependence predicted by the hypothesis that the feeding response is a result of sexual conflict over control of female reproductive responses (including feeding): feeding levels were indistinguishable in females mated once or twice (Figure 4.5). The results are instead consistent with the hypothesis that a single mating switches on female foraging as part of a suite of post-mating reproductive responses: female feeding and egg-laying both increased after mating and were closely coordinated (Figure 4.1c), and these responses remained elevated as many as 9 days after a single mating (experiment 5, Figure 4.5). A dose-dependent response is predicted when an increased female response is associated with increasing net costs of the response for females (Arnqvist 2006); thus, the finding here of little or no dose-dependence in the feeding response may indicate that the post-mating increase in feeding is not costly for females.

4.4.3 Does a generalized feeding response explain nuptial feeding?

Ingesting spermatophores gives little or no detectable benefit to female A. bipunctata (Perry & Rowe 2008a, b; Perry et al. 2009). The sensory bias hypothesis attempts to reconcile this observation with the maintenance of nuptial feeding by suggesting that nuptial feeding is triggered by female responses that originated in another context (Sakaluk 2000). Yet the results of this study do not support the hypothesis that the post-mating feeding response can account for nuptial feeding in this species. Indeed, mated females were less likely than virgin females to ingest spermatophores presented to them immediately after mating, despite their strongly elevated foraging during this period compared to virgins. Two possible explanations (besides the chance of Type I error) are, first, that mated females have a brief quiescent period after mating. However, this is not supported by the observations that movement levels are similar between mated and non-mated females 10 min after mating and that within 30 minutes of mating mated females tend to eat more than non-mated females (Figure 4.2). Second, given that mated females feed more in general but feed less on

53 spermatophores in particular, an intriguing possibility is that females discriminate against feeding on spermatophores in order to avoid the effects of spermatophore ingestion such as increased resistance to re-mating (Perry & Rowe 2008a). This hypothesis is at odds with the observation that most females (~90%) do in fact ingest at least part of the spermatophore after mating, with 67% of females eating the entire spermatophore (Perry & Rowe 2008a). Yet this observation also indicates that females do not eat the entire spermatophore in one third of opportunities. Experiments testing the effect on female fitness of variation in the responses induced by spermatophore feeding (i.e., advanced oviposition and increased re-mating resistance; Perry and Rowe 2008a; see Gwynne 2008) may help clarify whether selection favours or disfavours nuptial feeding in this species.

4.4.4 Conclusions

This study demonstrates that female ladybirds display a post-mating increase in feeding that is stimulated by males, but this response does not account for nuptial feeding in A. bipunctata . This outcome should not preclude testing the hypothesis in other species characterized by nuptial gifts of little detectable benefit to females (see section 4.1). More broadly, there is growing interest in determining which female post-mating responses are stimulated directly by males (e.g., via seminal proteins) and are part of a cascade of reproductive responses in the mated female (e.g., Barnes et al. 2008; Isaac et al. 2009; Ja et al. 2009). The feeding response may potentially link some of the most striking post-mating responses, such as increased reproduction and decreased longevity. Current research in this area is overwhelmingly based in D. melanogaster , and it appears timely to expand these studies to other taxa.

4.5 Appendix

4.5.1 Appendix 4.A

See Table 4.A.1

4.6 Figures

(a)

(b)

(c)

Figure 4.1

Male and female beetles were assigned to a mating or control (non-mating) treatment to test the effects of sex and mating on (a) food ingestion (± 95% CI) in the 24 h following mating and (b) activity level scored in the 4 h after mating. (c) Feeding was positively correlated with egg-laying 2 during the feeding trial, in both mated (open circles; R = 0.59, β = 0.12 ± 0.02, F 1,20 = 28.7, P < 2 0.0001) and control females (x; R = 0.68, β = 0.21 ± 0.05, F 1,9 = 19.0, P = 0.002). Letters indicate significant differences among groups by a Tukey HSD test.

Figure 4.2

Food ingestion (flour moth eggs, ± 95% CI) by non-mated (open bars) and mated (solid bars) females in each of 3 time periods following mating (see text). Females were deprived of food or satiated before the experiment as indicated.

Figure 4.3

Food ingestion (flour moth eggs, ± 95% CI) over 24 h by females subjected to one of three treatments: mating, artificially weighted to simulate carrying a male during copulation, or a control treatment. Letters indicate significant differences among groups by a Tukey HSD test.

Figure 4.4

Food intake (flour moth eggs, ± 95% CI) in 24 h by females assigned to matings of varying duration or control (non-mated) females, with letters indicating significant differences among groups

Figure 4.5

Food intake (flour moth eggs, ± 95% CI) by previously mated or virgin females assigned to a mating or non-mating treatment, with letters indicating significant differences among groups by a Tukey HSD test. There was a significant interaction effect on feeding between past mating history and the current mating treatment (F1,88 = 5.8, P = 0.02) as well as significant main effects (F1,88 = 26.3, P <

0.0001; F1,88 = 15.6, P = 0.0003, respectively).

4.7 Tables

Table 4.1

Tests of the effect of mating and hunger on food ingestion in three intervals after mating

Interval after mating

0-30 min 30-120 min 120-240 min

Food ingestion 1

P = 0.001 P = 0.02 Mating F1,166 = 3.6, P = 0.06 F1,116 =10.7, F1,166 = 5.7,

P = 0.0002 P = 0.01 Hunger F1,166 = 1.2, P = 0.27 F1,116 = 14.6, F1,166 =6.9,

Mating x Hunger F1,166 = 0.0, P = 0.94 F1,116 =0.0, P = 0.99 F1,166 =0.4, P = 0.53

1 Box Cox transformations were applied

Table 4.A.1

Oviposition behaviour during a feeding trial and egg fertilization after the trial for females subjected to matings of different durations, or control females that did not mate. The results are consistent with previous work on sperm transfer in Adalia bipunctata (see main text) which suggest that little sperm transfer occurs before 30 minutes of mating.

During feeding trial Post feeding trial Mating # of females that laid # of females in which treatment # of eggs laid 1 ≥ 1 egg ≥ 1 egg fertilized 2

No mating 24/42 (57%) 5.7 a 0/39 (0%)

1 min 24/42 (57%) 7.1 ab 0/40 (0%)

15 min 28/44 (64%) 7.2 ab 0/42 (0%)

30 min 29/41 (71%) 8.2 ab 6/41 (15%)

60 min 35/42 (83%) 9.3 ab 36/42 (86%)

120 min 2 18/34 (53%) 11.3 b 27/31 (87%)

Complete 31/44 (70%) 10.3 ab 33/41 (80%) mating 4

1 Least squares means are given for females that laid ≥ 1 egg. S.E. = 0.2 for all treatments. Letters indicate significant differences among groups by Tukey’s HSD test.

2 Sample sizes for egg fertilization may not equal the sample size in each treatment because some females died before laying eggs.

3 Eight females stopped mating before 120 minutes and were excluded from analysis

4 x¯ duration 169 min ± 11.5 (range 37-357 min)

Chapter 5 Condition-dependent female remating resistance generates sexual selection on male size in a ladybird beetle 5

5 Abstract

Behavioural resistance to remating by females is common, but the causes and consequences of resistance are rarely explained. Prominent hypotheses include resistance as a means of avoiding costly and superfluous mating, or as a means of biasing mating towards high-quality males. In species in which males produce nutritious nuptial gifts, females may further modulate resistance according to their need for nutrition. I investigated these hypotheses in the ladybeetle Adalia bipunctata , in which females frequently display vigorous resistance before copulation and ingest a spermatophore after copulation. In two experiments, I manipulated female nutritional state, depriving or satiating females for a short (16 h) or long (96 h) interval before a remating trial. I found that food-deprived females resisted mating more frequently and for longer periods than satiated females and consequently remated less frequently. This condition dependence of resistance supports the hypothesis that resistance functions to reduce superfluous and costly mating. My finding that food-deprived females were more resistant suggests that mating imposes energetic costs, and that nuptial feeding does not offset these costs. In a third experiment, I investigated whether the extent of resistance depended on male size or whether resistance itself biased mating towards large males. The extent of female resistance was independent of male size, but resistance itself resulted in a mating bias towards large males. In summary, these results support the hypotheses that females resist mating simply because it is costly and superfluous, and that a side effect of resistance is sexual selection for large male size.

5 Published as: Perry, J.C., D.M.T. Sharpe and L. Rowe. 2009. Condition-dependent female remating resistance generates sexual selection on male size in a ladybird beetle. Animal Behaviour 77 : 743-748.

5.1 Introduction

Behavioural resistance to mating by females is a common feature of mating systems. It may function to reduce female mating frequency, and in some cases, it biases mating success of males towards phenotypes that can overcome resistance (Arnqvist & Rowe 2005). Forms of resistance range from vigorous struggles with males (e.g. Rowe et al. 1994; Day & Gilburn 1997; Jormalainen 1998; Blanckenhorn et al. 2002) to avoidance of males through habitat switching (e.g. Krupa et al. 1990; Stone 1995; Rowe et al. 1996). In several of these examples, experiments have shown that resistance is costly to females. These costs include physical harm or elevated mortality (e.g. Mesnick & Le Boeuf 1991; Rowe 1994; Mühlhäuser & Blanckenhorn 2002) and missed opportunities such as foraging (Rowe 1992; Stone 1995). Evidence that females may pay a cost for resistance implies that some direct or indirect benefit offsets these costs.

There are several nonexclusive hypotheses that may account for female resistance to mating, yet there have been few attempts to experimentally distinguish among them. First, females may resist simply because additional mating is superfluous for fertilization and is costly. Although there is substantial support for the existence of costs to superfluous matings (reviews in Thornhill & Alcock 1983; Gwynne 1989; Choe & Crespi 1997; Arnqvist & Nilsson 2000), experimental support for the hypothesis that these costs account for female resistance is minimal. Direct support comes from economic studies where the costs (or benefits) of mating to females are manipulated, and the extent of resistance is then monitored (e.g. Lauer 1996; Blanckenhorn et al. 2002; Hosken et al. 2003; Teuschl & Blanckenhorn 2007). For example, in water striders, hungry females tend to increase resistance to mating, as expected because mating conflicts with female foraging (Rowe 1992), and females with stored sperm are more resistant than those depleted of sperm (Ortigosa & Rowe 2003).

The economics of female resistance to mating in species with nuptial gifts may be a particularly interesting case. In these species, it has been argued that some mating occurs as a means of acquiring resources from males (Gwynne 1984b). If so, then one would expect resistance to decline when females are hungry, the opposite of the pattern observed in water striders. The evidence here is mixed. In some species with nuptial gifts, nutritionally deprived females do tend to be less resistant (e.g. Thornhill 1984; Gwynne 1990; Simmons & Bailey 1990; Bilde et al. 2007). However, in other species, nutritional state appears to have no effect on willingness to mate (Engqvist 2007c), suggesting that either nuptial gifts are not valuable to females as food items (see Vahed 1998), or some other factor is determining resistance.

A second hypothesis for female resistance is the male screening hypothesis, which relies on indirect rather than direct selection on resistance (West-Eberhard 1983; Wiley & Poston 1996; Eberhard 2002; Kokko et al. 2003). In short, females may resist males selectively so that mating is biased towards males of high genetic quality; females pay a direct cost of resisting males to obtain the indirect benefit of improved offspring quality. In some species, resistance does favour certain male phenotypes; however, there is little evidence that females modulate their level of resistance based on male phenotype (Arnqvist & Rowe 2005). If female resistance does not depend upon male phenotype, but male phenotype does affect the success of males in overcoming resistance, then biases may simply be a by-product of a general resistance by females to costly mating as per the first hypothesis above. In two well-studied systems (seaweed flies and water striders) the by-product hypothesis is supported: resistance depends on ecological circumstance rather than the phenotype of the male (Crean & Gilburn 1998; Shuker & Day 2001; Ortigosa & Rowe 2002). The by-product hypothesis can account for resistance and biases in male mating success (i.e. direct selection on male phenotypes that help males overcome resisting females), but does not speak to any positive (or negative) indirect selection that may result from these biases (i.e. a good genes process is not necessarily predicted).

In this study I examine each of these hypotheses for female resistance in a species with so-called nuptial gifts, the two-spot ladybird beetle Adalia bipunctata . Following copulation, females eject a spermatophore and consume it (Perry & Rowe 2008a). Nevertheless, females often vigorously resist remating by kicking at or running from males or by bending the abdomen to prevent genital contact. Ladybirds are known to face food-limited conditions in nature (Sloggett & Majerus 2000), and I reasoned that mating would conflict with female foraging. Mating involves females carrying males for several hours, and it is likely that foraging efficiency would decline and energy consumption would be elevated during this period. If females resist because mating interferes with foraging and is energetically costly, then hungry females should resist mating more than satiated females. Alternatively, if energy acquired through consuming spermatophores offsets these costs, then I would expect the opposite effect of hunger on resistance. In two experiments of the current study, I manipulated short-term and long-term female nutritional state to test these predictions.

In a third experiment, I determine whether resistance is dependent on the phenotype of males, and whether resistance tends to bias mating success of males. There are mixed reports of a mating advantage to large males in A. bipunctata , but the origin of these size biases have not been

67 investigated (e.g. Tomlinson et al. 1995; Yasuda & Dixon 2002). One possibility is that females resist less with larger males as a means of biasing mating success towards them (i.e. the male screening hypothesis, Eberhard 2002). Another is that larger males are simply better able to overcome female resistance. To distinguish these hypotheses, I determined whether females altered their level of resistance based on male size, and whether resistance per se favoured larger males.

5.2 Methods

5.2.1 Experimental animals

Adalia bipunctata is an aphid predator widespread in temperate habitats (Omkar & Pervez 2005). Both males and females mate multiply. Females are typically larger than males and there is substantial variation in mass (range: female, 5.22–17.83 mg; male, 5.34–13.95 mg; this study). The beetles used in this study were from the F1 generation reared in the laboratory, from stock obtained from Natural Insect Control (Stevensville, Ontario, Canada). During maintenance periods, animals were provided daily with moistened cotton as a water source and fed pea aphids ( Acyrthosiphon pisum reared on broad bean, Vicia faba ) and UV-sterilized flour moth eggs ( Ephestia kuehniella ).

5.2.2 Assessing female remating resistance

I investigated the effect of food level on female remating behaviour in two experiments. Each experiment consisted of an initial mating, followed by a feeding treatment and then a remating trial in which I monitored mating resistance. Females were housed individually in petri dishes (50 × 12 mm) throughout. For the initial mating, virgin females of similar age were mated once to a male from the laboratory stock. I did not interfere with spermatophore consumption after mating, which meant that most females probably ingested some or all of the spermatophore ( >90%; Perry & Rowe 2008a). Following this mating and before the feeding treatment, females were fed excess flour moth eggs for several days. During this period, I monitored oviposition and discarded females that did not oviposit because it may have indicated a failure of sperm transfer.

For the remating trial, females were paired with a test male from the laboratory stock. All males had mated at least once previously. Males were maintained on excess flour moth eggs and kept isolated from females for at least 2 days before the trial. Males that did not attempt to mount the female within 10 min were replaced. I recorded whether the female resisted a male’s mating attempt, and

when resistance occurred, I measured the duration of resistance behaviour until mating began or until the male was dislodged. Males often remount females immediately after being dislodged. If the male did not remount the female within 1 min, I ended the trial. When males remounted within 1 min and females again resisted, I timed the duration of resistance and added it to the initial resistance time. I repeated this measurement of resistance until mating occurred or the female successfully eluded the male for at least 1 min. This design accounts for the likelihood that, in nature, a male may be able to immediately remount a female if dislodged but, I conjecture, would be less likely to remount if the female puts some distance between them. I separately analysed the remating responses of females considering only the first bout of resistance, but as the results were similar I do not report them here.

5.2.3 Short-term hunger

To test the hypothesis that short-term hunger influences resistance, I began the food treatment 16 h before the mating trial. Females were transferred to new petri dishes, provided with moistened cotton as a water source, and either deprived of food (N = 16) or fed an excess of flour moth eggs (N = 18). Sixteen hours should have been sufficient time for gut clearance, which occurs in 2–12 h (McMillan et al. 2007). After 16 h, females were transferred to a new dish and paired with a male for the remating trial.

5.2.4 Long-term hunger

Here the food treatment began 4 days before the remating trial. Low-food females were fed one adult pea aphid daily, and high-food females were provided excess pea aphids replenished three times a day. Food treatments are often used as surrogates for condition (Cotton et al. 2004; Bonduriansky & Rowe 2005), and there was evidence that the food treatment affected female condition. After 4 days of differential feeding, low-food females oviposited significantly fewer eggs

(mean ± SE = 16 ± 9.1 versus 93 ± 8.0; F1,30 = 40.2, P < 0.0001) and gained less weight, compared to high-food females. High-food females gained mass (3.33 ± 0.42 mg), whereas low-food females maintained or lost mass ( -0.54 ± 0.48 mg), and this difference was significant (F1,30 = 37.1, P < 0.0001). The experiment began with 18 females in each treatment group; four low-food females died before the remating trial.

5.2.5 Resistance and male size

By pairing resistant and less resistant females with small or large males, I tested two hypotheses about the relationship between resistance, male size and mating success. Specifically, I asked first whether either resistant (low condition) or less resistant (high condition) females altered their resistance as a function of male size. Second, I asked whether male size affected the probability of mating at either resistance level.

I generated small and large males by subjecting third-instar larvae to a low(one aphid daily) or high (excess aphids) diet treatment until emergence. Prior to the third instar, all larvae were fed excess flour moth eggs.

Prior to the female feeding treatment, virgin females were fed excess flour moth eggs daily and mated once to a male from the laboratory stock. Following this initial mating, I monitored oviposition and discarded females that did not oviposit before the remating trial. I generated females that were more or less resistant to remating by applying a food treatment after the initial mating: a low (0.1 mg of flour moth eggs for 2 days) or high (excess flour moth eggs) food diet for 8 days before the mating trial.

This food treatment was substantial enough to detect a difference in female condition between treatments. Low-food females experienced a smaller mass increase than high-food females (1.00 ±

0.28 mg versus 3.57 ± 0.29 mg; ANOVA: F1,50 = 40.5, P < 0.0001) and laid fewer eggs before the remating trial (23 ± 8.1 versus 85 ± 8.5; F1,49 = 28.2, P < 0.0001). Three low-food females died before the remating trial. The male food treatment also affected condition. Low-food males emerged as significantly smaller adults compared to high-food males (for a random subset of males: 6.50 ±

0.25 mg versus 7.83 ± 0.26; F 1,61 =13.5, P = 0.0005). All adult males were fed flour moth eggs and pea aphids ad libitum. Despite this, low-food males remained lighter in mass when the experiment

began (7.40 ± 0.11 mg versus 10.04 ± 0.12 mg; F 1,178 = 267.4, P < 0.0001).

For the remating trial, females were randomly paired with a small or large male (low-food, 52 and 39 pairs, respectively; high-food: 47 and 46 pairs). Female resistance behaviour was recorded as described above. The experiment was conducted in two blocks, the first with 53 pairs (28 low-food and 25 high-food) and the second with 131 pairs (63 low-food and 68 high-food).

5.2.6 Analyses

Each experiment generated two categorical and one continuous measure of female resistance: whether resistance occurred, whether remating occurred and the duration of resistance. For the two food-level experiments, the female food treatment was the sole factor in the chi-square goodness-of- fit tests or t-tests. For the food and male size experiment, the categorical responses were analysed with multiple logistic regression and the continuous response by ANOVA. Both types of models included food, male size, their interaction, and experimental replicate. Resistance duration data was analysed only for females that displayed resistance. Where appropriate, I applied a transformation to the resistance duration data to meet the assumptions of parametric statistics; backtransformed least squares means are presented with confidence intervals. The analyses were conducted using JMP 6.0.3 (SAS Institute Inc., Cary, NC, U.S.A.).

5.3 Results

5.3.1 Condition and resistance

Following the long-term hunger treatment, low-food (low condition) females were more likely to display remating resistance, and when they did resist, they resisted longer, with both effects contributing to reduced mating frequency (Table 5.1). Females that had experienced long-term hunger engaged in mating struggles that were seven times as long as those of satiated females, resulting in a 57% decrease in remating frequency. These effects were significant (P < 0.01), while the increase in the frequency of remating resistance was not (Table 5.1). I did not detect a significant difference in mating behaviour from the short-term hunger treatment, although the pattern of responses was consistent with the long-term experiment. Relative to the satiated females, females that had experienced short-term hunger were twice as likely to resist a mating attempt, and when resisting would struggle for over twice as long, resulting in a 28% decrease in mating frequency. This contrast suggests that an effect of short-term hunger was of lower magnitude, but may have been detectable with a larger sample.

5.3.2 Resistance and male size

Female condition had effects on resistance that were consistent with the other experiments. Low- food females were significantly more likely to resist remating (Table 5.2), and when they did resist,

71 they resisted more than two-fold longer than high-food females (backtransformed means: low-food females: 461 s (329, 644), high-food females: 153 s (98, 234); analysis performed on log-transformed data: F 1,118 = 17.3, P < 0.0001; Fig. 5.1).

Male size had no detectable influence on whether females showed remating resistance (Table 5.2) or the duration of resistance when it occurred (F 1,118 = 0.0, P = 0.99; Fig. 5.1; food level × male size interaction: F 1,118 = 0.2, P = 0.68). There was, however, an effect of male size on the frequency of remating that was dependent on female resistance. When males were paired with females that displayed remating resistance, large males were more likely to achieve copulation than small males: 59% of small males copulated, compared to 78% of large males (Table 5.2). In contrast, male size did not influence the mating outcome when males were paired with nonresisting females: 93% of small males copulated, compared to 81% of large males (Table 5.2). These data suggest that female resistance per se favours large males.

The overall level of resistance was higher in the first replicate of this experiment compared to the second, for both resistance frequency (Table 5.2) and resistance duration (F 1,118 = 5.7, P = 0.02).

5.4 Discussion

An understanding of the causes and consequences of costly female resistance to mating is crucial for distinguishing among competing models of sexual selection (Arnqvist & Rowe 2005; Rowe & Day 2006). The data support the hypothesis that female resistance to mating depends on their ecological setting: females that were deprived of nutrition increased their level of resistance, presumably because mating is costlier for low-condition females. This result suggests that any nutritional gain from edible spermatophores that come with mating is not sufficient to offset the cost of mating. The data further suggest that resistance favours large, high-condition males. However, this bias did not occur because females modulated their resistance according to male phenotype; instead, large males simply seem better able to overcome resistance.

5.4.1 Condition dependent remating resistance

Vigorous and lengthy behavioural mating resistance ( >5000 s; unpublished data) presumably carries some energetic costs for female A. bipunctata , and therefore suggests some benefit of resistance to females. Two broad hypotheses may account for resistance: (1) it is a means of reducing the rate of

72 costly and superfluous mating (a direct benefit), or (2) it is a means of screening for high-quality males as fathers (an indirect benefit; West-Eberhard 1983; Eberhard 1996; Wiley & Poston 1996; Cameron et al. 2003; Chapman et al. 2003b; Eberhard & Cordero 2003; Kokko et al. 2003). The results are consistent with the former hypothesis and are not predicted by the latter. In particular, the finding that females in poor nutritional condition displayed increased resistance implies that the costs of mating, and benefits of resisting mating, are greater for low-condition females. Clearly, a potential source for such differential costs is that low-condition females have a more pressing need to forage and mating may entail reduced foraging efficiency. Mating may also impose energetic costs that low-condition females have a reduced ability to tolerate. Both types of costs have been found in other species (e.g. Rowe 1992; Fairbairn 1993; Watson et al. 1998; Plaistow et al. 2003). Furthermore, the need to forage no doubt increases as time passes without food, and in this study, the length of food deprivation influenced the extent of female resistance. Condition-dependent resistance is not predicted by the mate screening hypothesis (Ortigosa & Rowe 2002). In fact, it is more likely that if resistance serves to screen males and is energetically costly, then high-nutrition females should resist more because they are better able to pay these costs. I observed the opposite.

The finding that females in good nutritional condition showed some resistance and did not always accept remating attempts from males may imply other costs of mating in addition to energetic costs, such as a risk of predation or pathogen transmission (e.g. Rowe 1994; Hurst et al. 1995). Direct harm from males is another possible cost, but recent laboratory studies of A. bipunctata have detected no such costs from additional matings to female longevity or reproductive success (Haddrill et al. 2007; Perry & Rowe 2008b). Potential costs (e.g. reduced foraging, increased predation) that are not detectable in the laboratory should be a topic of further investigation.

The extent of mating resistance observed is a joint outcome of female resistance and male persistence. Accordingly, two of the responses I measured, resistance duration and remating frequency, may have been influenced by male behaviour. Males may modulate persistence depending on female quality or on female resistance (and, thus, their chance of success). Both hypotheses predict that males will persist more with high-quality, less-resistant females. However, inspection of the mating frequency data in Table 5.2 indicates that the remating rate among nonresisting females was very similar for low- and high-food females. Furthermore, male persistence is not expected to influence the initial occurrence of female resistance to a mating attempt, and this measure was not substantially different from the other responses.

5.4.2 Resistance and nuptial gifts

I found no support for the hypothesis that hungry females seek additional matings to forage for edible ejaculates, suggesting that any nutritional benefit of the ejaculate does not offset the cost of mating. Past results have been mixed on this issue. In some orthopterans, poor food conditions result in females that are more receptive to mating (Gwynne 1990; Simmons & Bailey 1990). In contrast, female scorpionflies on a low-food diet are no more or less likely to accept nuptial gifts or to copulate with males than are females fed a high-food diet (Engqvist 2007c; but see Thornhill 1984). Previous work has found that female A. bipunctata have little to gain nutritionally from consuming spermatophores (Perry & Rowe 2008a). In fact, the failure to detect any benefit from consuming nuptial gifts is a common outcome in other species, and hypotheses addressing the maintenance of nuptial feeding in the absence of any benefit have been discussed elsewhere (e.g. Sakaluk 2000; Vahed 2007).

I have previously shown that spermatophore ingestion quickly and substantially increases female remating resistance (Perry & Rowe 2008a). It is noteworthy that the present results suggest that providing females with extra food diminishes remating resistance. Taken together, these results suggest that ingested ejaculates are distinct from food in their effect on resistance. An interesting possibility is that spermatophores transfer signals or stimulant proteins from males; the most likely source of such proteins is the male accessory glands, from which spermatophores are derived. Accessory gland proteins are well known to induce remating resistance in several species (Gillott 2003). A similar increase in remating resistance following nuptial gift ingestion has been found in other insects (Sakaluk et al. 2006; Engqvist 2007b). The present study suggests that in A. bipunctata , the increase in resistance from spermatophore ingestion cannot be explained as a food effect.

5.4.3 Sexual selection as a by-product of resistance

Large, high-condition males achieved a mating advantage over small males when females were resistant to mating. This result suggests that large males are better able to overcome the vigorous resistance of females, as has been indicated in other species (Crean & Gilburn 1998; Shuker & Day 2001; Ortigosa & Rowe 2002). Indeed, higher mating success for larger males was reported for A. bipunctata in a study in which heavier and lighter males were placed together with a single female (Tomlinson et al. 1995). My experiment suggests that the basis of the size advantage is in overcoming female resistance rather than success in intrasexual competition. A second study of resistance in A. bipunctata found that small males had an advantage in mounting females when

74 competed against large males, but small males did not achieve more copulations (Yasuda & Dixon 2002).

The mate screening or selective resistance hypothesis suggests that resistance to mating does not arise from a general cost of mating, but from a cost of mating with low-quality males (e.g. West- Eberhard 1983; Wiley & Poston 1996; Eberhard 2002; Kokko et al. 2003). Thus, a large (high- quality) male mating advantage might arise from females selectively screening out the small (low quality) males. Two lines of evidence argue against this hypothesis for these results. First, females did not modulate their resistance with respect to the size of the male attempting to mate. This negates the simplest means of using resistance to selectively screen males: reducing resistance to favoured males. Second, resistance behaviour was condition dependent, and the selective resistance hypothesis cannot account for the finding that high-condition females were less resistant than low- condition females.

In summary, this study suggests an important role for direct selection on female resistance, and points to sexual selection on male size as a consequence of resistance. Future work should investigate the generality of these patterns in other species in which females display conspicuous mating resistance.

5.5 Figures

Figure 5.1

Duration of female resistance to a second mating for female ladybeetles fed low- and high-food diets and paired with a small or large male. Backtransformed least squares means are presented. Bars indicate 95% confidence intervals.

5.6 Tables

Table 5.1

Female remating behaviour in ladybeetles after a short (16 h) or long (4 day) food treatment

Low-food females High-food females Test

Short-term experiment

2 Resistance frequency 9/16 (56%) 5/18 (28%) χ 1 = 2.7, P = 0.10

1,2 Resistance duration (s) 100 (67, 201) 82 (53, 181) t12 = 0.3, P = 0.61

2 Remating frequency 9/16 (56%) 14/18 (78%) χ 1 = 1.7, P = 0.19

Long-term experiment

2 Resistance frequency 11/14 (79%) 9/18 (50%) χ 1 = 2.6, P = 0.11

1,3 Resistance duration (s) 657 (348, 1239) 139 (69, 280) t18 = 10.4, P = 0.005

2 Remating frequency 5/14 (36%) 15/18 (83%) χ 1 =6.8, P = 0.009

1 Least squares mean resistance durations were calculated for females that displayed resistance (i.e. excluding zeroes). Backtransformed means are presented with confidence intervals.

2 Inverse-transformed in the analysis.

3 Log-transformed in the analysis.

Table 5.2

The remating behaviour of female ladybeetles exposed to a low- or high-food diet for 8 days and paired with a small or large male

Resistance frequency Remating frequency

Resisting females Nonresisting females

Treatment group

Low-food female, small male 47/52 (90%) 26/47 (55%) 5/5 (100%)

Low-food female, large male 33/39 (85%) 24/33 (73%) 4/6 (67%)

High-food female, small male 22/47 (47%) 15/22 (68%) 23/25 (92%)

High-food female, large male 21/46 (46%) 18/21 (86%) 21/25 (84%)

Analysis

2 2 2 Food level χ 1=31.2, P < 0.0001 χ 1=2.5, P=0.11 χ 1=0.0, P=0.94

2 2 2 Male size χ 1=0.6, P=0.46 χ 1=4.5, P=0.03 χ 1=0.0, P=0.93

2 2 2 Food level × male size χ 1= 0.6, P=0.45 χ 1=0.1, P=0.74 χ 1=0.0, P=0.93

2 2 2 Experimental replicate χ 1=4.1, P=0.04 χ 1=2.8, P=0.09 χ 1=2.2, P=0.14

Chapter 6 Condition-dependent ejaculate size and composition in a ladybird beetle 6

6 Abstract

Sexually selected male ejaculate traits are expected to depend on the resource state of males. Theory predicts that males in good condition will produce larger ejaculates, but that ejaculate composition will depend on the relative production costs of ejaculate components and the risk of sperm competition experienced by low- and high-condition males. Under some conditions, when low condition leads to poorer performance in sperm competition, males in low condition may produce ejaculates with higher sperm content relative to their total ejaculate and may even transfer more sperm than high-condition males in an absolute sense. Previous studies in insects have shown that males in good condition transfer larger ejaculates or more sperm, but it has not been clear whether increased sperm content represents a shift in allocation or simply a larger ejaculate, and thus the condition dependence of ejaculate composition has been largely untested. I examined condition dependence in ejaculate by manipulating adult male condition in a ladybird beetle in which males transfer three distinct ejaculate components during mating: sperm, non-sperm ejaculate retained within the female reproductive tract, and a spermatophore capsule that females eject and ingest following mating. I found that high condition males indeed transferred larger ejaculates, potentially achieved by an increased rate of ejaculate transfer, and allocated less to sperm compared with low- condition males. Low-condition males transferred ejaculates with absolutely and proportionally more sperm. This study provides the first experimental evidence for a condition-dependent shift in ejaculate composition.

6 Published as: Perry, J.C. and L. Rowe. 2010. Condition-dependent ejaculate size and composition in a ladybird beetle. Proceedings of the Royal Society of London B. Published online 23 June 2010. (doi: 10.1098/rspb.2010.0810).

6.1 Introduction

A recurrent prediction for sexually selected traits is that they should exhibit heightened condition dependence relative to other traits (e.g. Alatalo et al. 1988; Rowe & Houle 1996; Cotton et al. 2004; Bonduriansky & Rowe 2005). This expectation is based on the assumption that individuals in good condition (i.e. with larger pools of resources available to allocate) have higher marginal benefits from increased investment in the expression and maintenance of sexually selected traits (Iwasa et al. 1991; Rowe & Houle 1996; Proulx et al. 2002; Getty 2006). The extent of condition dependence in sexually selected traits is a key issue for sexual selection research (Zahavi 1975; Iwasa & Pomiankowski 1994; Rowe & Houle 1996). Following a series of environmental manipulations of condition, there is now considerable empirical evidence for condition dependence in several types of sexually selected traits, including secondary sexual morphology, ornaments and pigmentation, and acoustic signals (Andersson 1994; Cotton et al. (2004) and references therein; Bonduriansky & Rowe 2005; Punzalan et al. 2008). Recent evidence suggests that male ejaculates are also sexually selected (Eberhard & Cordero 1995; Ramm et al. 2007; Simmons & Kotiaho 2007; Martin-Coello et al. 2009; Wigby et al. 2009) and costly (e.g. Dewsbury 1982). However, the extent of condition dependence in ejaculate traits is currently less well studied than these other sets of traits.

Whenever high-condition males have lower marginal costs of ejaculate production than low- condition males, they are expected to transfer larger ejaculates at mating (even as they invest less reproductive effort per ejaculation; Parker 1990; Tazzyman et al. 2009). Empirical studies involving experimental manipulations of condition generally support this prediction, reporting that high- condition males produce larger ejaculates (Gwynne 1990; Delisle & Bouchard 1995; Watanabe & Hirota 1999; Jia et al. 2000; Ferkau & Fisher 2006; Lewis & Wedell 2007; Blanco et al. 2009; but see Wedell 1993), transfer more sperm (Fedina & Lewis 2006; McGraw et al. 2007; Perez-Staples et al. 2008), and produce more ejaculate-derived nuptial gifts ( Jia et al. 2000; but see Wedell 1993).

Although current theory does not directly address condition dependence in ejaculate composition (the relative allocation to its components), there are two mechanisms by which condition might affect composition. First, if production costs (or benefits) vary among ejaculate components, more costly components should be disproportionately condition-dependent and present in greater concentrations in the ejaculates of high-condition males. A second mechanism links condition dependence to sperm competition theory, which predicts that under some conditions, males that are disfavoured in sperm competition should invest more in sperm relative to their overall ejaculate

80 expenditure, and in many cases should transfer more sperm (in an absolute sense) than favoured males (Cameron et al. 2007). One scenario in which this is expected is in those cases where investment in non-sperm components of the ejaculate tends to elevate short-term egg production. In some species, low-condition males may be consistently disfavoured in sperm competition (e.g. when female re-mating rate is elevated after matings with low-condition males, Chapman et al. 2003c; Pitcher et al. 2003; or when females bias sperm storage towards high quality males, Vermeulen et al. 2008). Although previous studies have reported increased sperm transfer by high- condition males (referenced above), it is generally not clear whether this represents an increased allocation to sperm relative to non-sperm ejaculate components or an increase in total ejaculate, or even a decreased allocation to sperm if non-sperm components increase more than sperm. Consequently, the degree of condition dependence in ejaculate composition remains largely unknown.

Here I investigate the condition dependence of ejaculate size and composition in the two-spot ladybird beetle Adalia bipunctata . During copulation, male beetles transfer ejaculate via a spermatophore. Females eject and then ingest the emptied spermatophore capsule after copulation; notably, this ingestion causes an acceleration of egg production and a marked increase in female resistance to subsequent matings (Perry & Rowe 2008a). Thus, three distinct ejaculate components can be distinguished: the spermatophore capsule; sperm; and non-sperm seminal fluids that are retained within the female after mating (hereafter, ‘retained ejaculate’). I predict that (i) high- condition males will transfer a larger total ejaculate mass than low condition males (Parker 1990), and (ii) male condition will influence ejaculate composition. As discussed above, differences in composition could result from differential costliness of ejaculate components (unknown for these beetles) or from condition-dependent differences in the level of sperm competition. The latter may arise, for example, if production of the spermatophore capsule itself is condition-dependent: reduced capsule production should increase female re-mating (Perry & Rowe 2008a) and lead to increased sperm competition for low-condition males; if so, those low-condition males should increase the sperm content of their ejaculates (Cameron et al. 2007).

6.2 Methods

6.2.1 Experimental animals

Adalia bipunctata are aphid predators with a broad Holarctic distribution. Both males and females mate multiply (field estimates: 2–4 matings per female; Haddrill et al. 2008), with copulations lasting up to 6 h. I have never observed sperm transfer in copulations lasting less than 30 min, and in this study I excluded these shorter matings from analysis. Both mating and ejaculate production appear costly for male A. bipunctata : males are limited in their ability to produce sequential ejaculates and males that mate a single time have reduced survival compared with non-mating males (Perry and Tse, manuscript in preparation). The ladybirds used in these experiments were of the second and third generation reared in the laboratory from a founding population obtained from a biocontrol company (Natural Insect Control, Stevensville, Ontario, Canada). Ladybirds were maintained on pea aphids ( Acyrthosiphon pisum , reared on broad bean, Vicia faba ) and eggs of the Mediterranean flour moth ( Ephestia kuehniella ), a standard artificial diet (de Clerq et al. 2005).

6.2.2 Experimental design

To investigate the effect of male condition on ejaculates, I randomly assigned adult males to a high- or low-food treatment for several days prior to assaying ejaculate size, composition and sperm content. Condition, or an organism’s accumulated resources for allocation, is expected to have both genetic and environmental components (Rowe & Houle 1996), and food level treatments are commonly used to manipulate the environmental component (reviewed by Cotton et al. 2004). I expect that, on average, males assigned to the low-food treatment will have accumulated fewer resources for allocation to ejaculate traits than males assigned to high-food.

Each male was mated once to a female from the stock population before the feeding treatment began. The experiment was replicated three times. In the first replicate, the feeding treatment consisted of adult pea aphids provided daily for 9 days at low (one aphid) or ad libitum levels, while in subsequent replicates males were fed flour moth eggs at low (15–25 eggs) or ad libitum levels for 10 days. I confirmed that the food treatments differentially affected a measure of condition (mass gain). Initial male mass did not differ between the high- and low-food groups (8.83 mg ± 0.20 versus 8.91 ± 0.21, respectively; mixed model with the fixed factor ‘feeding treatment’ and the random

factor ‘replicate’: F 1,159 = 0.1, P = 0.75). However, within each replicate low-food males lost more weight, or gained less weight, than high-food males (mean ± s.e., first replicate: low-food males:

-0.671 mg ± 0.074, high-food males: -0.172 mg ± 0.076, t 38 = 4.7, P < 0.0001; second replicate: low:

-0.090 mg ± 0.187, high: 0.866 mg ± 0.130, t 44 = 4.2, P < 0.0001; third replicate: low: -1.242 mg ±

0.089, high: -0.085 ± 0.100, t 75 = 8.4, P < 0.0001). Ten of 96 high-food males and 29 of 108 low- food males died before the experiment began.

To begin the mating trial, female beetles were placed individually into Petri dishes for an hour to acclimate before a low- or high-food male was introduced. Virgin females were used to minimize the possibility that females would reject low-condition males, as I have never observed mating resistance by reproductively mature virgins from this population. In the first and second replicates, I recorded whether mating occurred within 2 h, copulation duration, the time until spermatophore capsule ejection, and spermatophore capsule mass immediately after ejection. In the third replicate, I expanded these measurements to include male mass immediately before and after copulation, dry spermatophore capsule mass, sperm transfer, and male post-mating survival after an imposed physiological stressor (described below).

To obtain an estimate of the mass of retained ejaculate, I used the difference between male mass before and after mating and corrected this difference for the per minute mass loss experienced by non-mating males (unpublished data: 1.587 µg min -1 ± 0.264; N = 14); finally, I subtracted the mass of the ejected spermatophore capsule. This ‘retained ejaculate’ mass probably reflects the non-sperm component of ejaculate because sperm contributes little to ejaculate mass in many species (Eberhard & Cordero 1995; Simmons 2001; Owen & Katz 2005). I determined the water content of spermatophore capsules by re-weighing them after 65 h in a drying oven (60 °C) and subtracting this mass from initial wet mass. Females do not appear to eject additional ejaculate apart from the spermatophore capsule.

6.2.2.1 Quantifying sperm transfer

To assess sperm transfer, I froze the experimental females (-20°C) 1 h after mating and quantified sperm from the dissected female reproductive tracts following Arnaud et al. (2003). I transferred the bursa copulatrix and spermatheca (the sites of sperm deposition and storage, Ransford 1997) and ovarian tubes to a cavity slide containing 100 µl Ringer’s solution. I ruptured the reproductive tract, crushed the spermatheca, and teased apart the tissues with fine forceps. I then washed the solution from the slide with 2 ml Ringer’s solution into a microcentrifuge tube and vortexed the solution for 2 min. I pipetted two 20 µl samples from each solution on to a glass slide and allowed them to dry

83 under a dust cover. Sperm were counted under a dark field phase contrast microscope at 400× magnification. I summed the number of sperm detected in both samples and multiplied by the dilution factor to estimate the total number of sperm transferred. The number of sperm was highly repeatable between the two samples (r = 0.86, P < 0.0001).

6.2.2.2 Male post-mating survival

Following the mating trial, I maintained the experimental males on the assigned low- or high-food treatment and assayed their survival after exposure to a physiological stress. On the 6th day after the mating trial, each male was placed in a 1.5 ml microcentrifuge tube containing a small air hole and placed in a 40 °C water bath for 1 h (stressful conditions for these beetles, Acar et al. 2005). Males were returned to room temperature and provided with 75 µl distilled water daily. Male survival was monitored three times daily and mortality was recorded when I could no longer provoke a response by gentle prodding.

6.2.2.3 Statistical analyses

I analysed categorical responses by χ2-tests and continuous responses by fitting mixed models with the fixed factor ‘feeding treatment’ and the random factor ‘replicate’ or by a one-way ANOVA for the responses measured in the third replicate only. For two responses—the mass of retained ejaculate and sperm number—I tested for an effect of copulation duration by an ANCOVA with feeding treatment, copulation duration and their interaction. I tested for an effect of the feeding treatment on male survival using a proportional hazards model. Non-significant interactions were dropped from the final models. I transformed continuous responses to meet the assumptions of parametric tests and conducted non-parametric Wilcoxon tests if no transformation could make the data compatible.

To assess the correlation structure among aspects of ejaculate and mating, I conducted two principal components analyses (PCAs). The first analysis included three variables which represented the most complete dataset: copulation duration, the mass of retained ejaculate and number of sperm transferred. The second analysis included these variables and spermatophore capsule mass, which was available for a subset of the data. I then examined how the composite variables generated by the PCA related to male post-mating survival by testing for a correlation between survival and the principal component scores, using linear regression. Because the results from both PCAs were very similar, I present the second analysis here and the initial PCA in the appendix (Appendix 6.A).

6.3 Results

6.3.1 Univariate analyses

Male condition, as determined by a low- or high-food diet, significantly impacted male mating behaviour. Compared with males provided with plentiful food, low-food males were less likely to mate (80% versus 95%) and copulated for shorter periods when they did mate (Table 6.1). I next detail how male condition impacted ejaculate composition and the quantity of each ejaculate component.

6.3.1.1 Retained ejaculate

The portion of ejaculate retained within the female—i.e. the bulk of seminal fluids (see 6.2.2)—was smaller following matings with low-food males compared with high-food males (F 1,75 = 20.1, P < 0.0001, Fig. 6.1a). Furthermore, this retained ejaculate component represented a smaller fraction of the total body mass for low- compared with high-food males (Table 6.1). However, retained

ejaculate mass was not influenced by copulation duration (F 1,73 = 0.1, P = 0.76) or by the interaction

between copulation duration and the food treatment (F 1,73 = 1.1, P = 0.30); these terms were therefore dropped from the final model. There was no significant difference in the rate of retained ejaculate transfer between low- and high-food males, though there was a trend of decreased transfer rate in low-food males (Table 6.1).

6.3.1.2 Sperm

Low- and high-food males were equally likely to transfer sperm, given a successful mating (Table 6.1). However, when sperm transfer occurred, low-food males transferred over twice as much sperm as high-food males (F 1,23 = 7.5, P = 0.01, Fig. 6.1c). Low-food males also transferred sperm at a higher rate, in contrast to the absence of a treatment effect on the transfer rate of the retained ejaculate component reported above (Table 6.1).

There was little detectable effect of copulation duration on sperm transfer. Longer copulations did not increase the likelihood of sperm transfer; in fact, there was a weak negative relationship between 2 2 copulation duration and the likelihood of sperm transfer (R =0.17, β = -0.02 ± 0.01, N = 73, χ 1 = 16.9, P < 0.0001). Similarly, there was no correlation between copulation duration and sperm

quantity (F 1,22 = 0.9, P = 0.34), and no significant interaction effect on sperm quantity (F 1,22 = 3.7, P

= 0.07; no significant correlation between copulation duration and feeding treatment within either treatment group: low-food males: R 2 = 0.14, β = 38.6 ± 32.0, P = 0.26; high-food males: R 2 = 0.16, β = -12.6 ± 8.1, P = 0.15).

I controlled for a strong threshold relationship between spermatophore ejection latency and sperm transfer (Fig. 6.A.1). When ejection latency exceeded 11 min, sperm transfer was very low or nil with only one exception; I therefore excluded cases where ejection latency exceeded this threshold.

6.3.1.3 The spermatophore capsule

Male condition significantly influenced both the size and composition of the spermatophore capsule. Females that mated with low-food males were less likely to eject a spermatophore capsule after mating (Table 6.1); moreover, capsules ejected after matings with low-food males were smaller (F 1,80 = 12.6, P < 0.001, Fig. 6.1b) and contained more water (31% versus 14%; table 1). However, both low- and high-food males transferred spermatophore capsules that represented a similar proportion of ejaculate mass (Table 6.1).

Spermatophore capsule mass was not influenced by copulation duration (low-food males: R 2 = 0, P = 0.81; high-food males: R 2 = 0, P= 0.83) or by the latency until spermatophore ejection (low-food: R2 = 0.04, P = 0.30; high-food: R 2 = 0, P = 0.89). Male condition did not influence the latency until spermatophore ejection (Table 6.1). I excluded 17 spermatophores from analysis because they were partially consumed by females before I could remove them.

6.3.1.4 Ejaculate composition

There was a conspicuous shift in ejaculate composition between low- and high-food males. Specifically, low-food males transferred ejaculates that contained a higher concentration of sperm (a

2.5-fold increase, F 1,23 = 4.5, P = 0.04; Fig. 6.1d). In contrast, high-food males transferred a higher concentration of non-sperm ejaculate components, indicated by their larger retained ejaculates (Fig. 6.1a); and as noted above, they transferred spermatophore capsules with decreased water content (Table 6.1). However, the proportion of ejaculate made up of the spermatophore capsule was not sensitive to male condition (Table 6.1).

6.3.1.5 Male survival

I obtained survival data for 22 low-condition males and 18 high-condition males. Improved male condition greatly increased male survival after exposure to a physiological stress (by over threefold; Fig. 6.2 and Table 6.1), as would be expected if the low-food treatment stressed males.

6.3.2 Multivariate analyses

I used PCA to summarize the correlation structure among four variables: the mass of retained ejaculate, sperm transfer, spermatophore capsule mass and copulation duration. The first two principal components (PCs) from this analysis captured 65% of the variation (Table 6.2). The first PC axis reflected positive correlations among all four aspects of responses, while the second axis captured a negative correlation between ejaculate mass and sperm transfer, indicating a change in the concentration of sperm that appears unrelated to copulation duration or spermatophore capsule mass (Table 6.2).

The male feeding treatment influenced both PC1 and PC2 (Fig. 6.3; PC1: F 1,38 = 5.8, P = 0.02; PC2:

F1,38 = 8.0, P < 0.01), but not PC3 or PC4 (PC3: F 1,38 = 0.7, P = 0.40; PC4: F 1,38 = 1.6, P = 0.21). Males in good condition had, on average, positive scores on PC1 (0.23 ± 0.22), indicating that they had high values for all four mating variables, while males in poor condition had negative PC1 scores on average (-0.55 ± 0.24), indicating decreased values for all variables. Similarly, males in good condition had on average positive scores on PC2 (0.24 ± 0.18) while poor condition males had negative scores (-0.50 ± 0.19); on this axis, though, these scores meant that high-condition males transferred ejaculates with a reduced concentration of sperm, and conversely poor condition males transferred ejaculates with a high sperm concentration (Table 6.2). These results are in agreement with the univariate analyses above.

I next investigated how male survival related to the composite of male mating responses described by PC1 and PC2. Male survival was positively correlated with PC1 scores (R 2 = 0.28, N = 17, β = 1568+647, P = 0.03), indicating an association between male survival and increased values for the four mating variables. In contrast, survival was not correlated with PC2 (R 2 = 0.06, N = 17, β = 1052+1114, P = 0.36).

6.4 Discussion

There is a growing appreciation for the potential influence of resource availability on ejaculate expenditure (Williams et al. 2005; Tazzyman et al. 2009). My finding that ejaculate traits are differentially sensitive to male condition supports this emerging view. Males in good condition produced larger ejaculates, consistent with studies in other species (discussed below). However, males in poor condition strongly increased allocation to sperm despite decreased overall ejaculate size, representing a condition- dependent shift in ejaculate composition. Below, I discuss the evidence and implications for each of these findings, and examine two hypotheses for the mechanism underlying condition dependent ejaculate composition.

6.4.1 Ejaculate size

My finding that males in good condition produced larger non-sperm ejaculates than males in poor condition is consistent with both theory (Parker 1990) and previous results from two insect groups (Lepidoptera: Delisle & Bouchard 1995; Watanabe & Hirota 1999; Ferkau & Fisher 2006; Lewis & Wedell 2007; Blanco et al. 2009; Orthoptera: Jia et al. 2000, but see Wedell 1993). Although high- condition males had longer copulations than low-condition males, there was no evidence that longer copulations were the factor generating increased ejaculate transfer because copulation duration was not correlated with ejaculate mass. Instead, I found weak evidence that high-condition males transfer non-sperm ejaculate at a faster rate, although this difference was not statistically significant.

Relative to high-condition males, low-condition males decreased their investment in spermatophore capsules: their spermatophore capsules were smaller, and their mates were less likely to eject a capsule at all. Previous reports on the condition dependence of ingested ejaculate products are mixed, with high-condition males transferring larger ejaculate gifts in some insects (Jia et al. 2000; Lehmann & Lehmann 2009) but not others (Tuckerman et al. 1993; Wedell 1993; Simmons et al. 1999). For A. bipunctata , the reduction in spermatophore capsule ejection, and thus in capsule ingestion by females, probably did not represent a loss of nutrition for females because spermatophore capsules transfer little or no nutrition and do not improve female fitness (Perry & Rowe 2008a,b; Perry et al. 2009). Instead, ingesting spermatophores increases female re-mating resistance and accelerates short-term egg production (Perry & Rowe 2008a). The first factor will put low-condition males at a disadvantage in sperm competition through re-mating, and the second factor will lead to a reduced number of eggs that they are in competition for.

6.4.2 Ejaculate composition

I found evidence for three condition-dependent changes in ejaculate composition. Most notably, low-condition males transferred ejaculates that were higher in absolute sperm content and in sperm concentration, compared with high-condition males. Low-condition males transferred more sperm despite their reduced copulation duration by transferring sperm at a faster rate than high-condition males, and this occurred despite the trend that high-condition males transferred retained (non- sperm) ejaculate at a faster rate (Table 6.1). In contrast to this result, previous experimental studies in flies and flour beetles have found that improved male condition leads to increased absolute sperm transfer (Fedina & Lewis 2006; McGraw et al. 2007; Perez-Staples et al. 2008); however, these studies did not report sperm concentration and as a result it is not possible to determine whether and how ejaculate composition differed. An alternative explanation for my result is that females mating with low- or high-condition males differed in sperm storage or usage in the hour following mating. However, this is unlikely to have affected my measure of sperm transfer because I examined the entire female reproductive tract for sperm; furthermore, there is no reason to expect females to preferentially retain sperm from low-condition males.

In contrast to the result for sperm, high-condition males transferred larger retained ejaculates than low-condition males, an outcome that is consistent with increased allocation to seminal proteins because seminal proteins make up the bulk of ejaculate in many species (see 6.2.2 ). Indeed, the results show that even greatly increased sperm numbers contribute little to ejaculate mass (Fig. 6.1).

I discussed earlier two mechanisms by which this sort of condition dependence in ejaculate composition might arise (see 6.1 ). First, more costly ejaculate component are expected to be more condition-dependent. The finding of condition dependence in sperm and non-sperm ejaculate components may be consistent with this prediction if seminal proteins have greater production costs than sperm in A. bipunctata . There is currently no data available on production costs in this or indeed in many species, making it impossible to predict the direction of condition dependence by this mechanism a priori .

A second mechanism stems from a model of ejaculate composition with regard to sperm competition (Cameron et al. 2007). It predicts that if seminal proteins function to influence female fecundity or the outcome of sperm competition, then males with an advantage in sperm competition should allocate a larger portion of their ejaculate budget to seminal proteins whereas disfavoured

89 males should allocate more to sperm. Moreover, the model describes a broad parameter space where disfavoured males should produce not only a proportionally greater allocation to sperm but an absolutely greater quantity than favoured males (Cameron et al. 2007). The results fit these predictions if, as in many other animals (e.g. Chapman et al. 2003c; Pitcher et al. 2003; Vermeulen et al. 2008), low-condition male A. bipunctata are disfavoured in sperm competition. There is no direct evidence available on this point, but the finding that low-condition males transfer fewer and smaller spermatophore capsules—which induce female re-mating resistance—makes it quite plausible that low-condition males are indeed disfavoured.

The composition of the spermatophore capsule itself varied with male condition: capsules produced by low-condition males had a higher water content relative to those of high-condition males. A possible explanation is that low-condition males that transfer small ejaculates attempt to maintain spermatophore capsule volume by increasing the water content, if doing so stimulates stretch receptors in the female reproductive tract (Ferkau & Fisher 2006). An additional and non-exclusive hypothesis is that females benefit from the water content of spermatophore capsules, based on recent evidence of such benefit in the seed beetle Callosobruchus maculatus (Edvardsson 2007; Ursprung et al. 2009). However, this hypothesis is unlikely to apply because A. bipunctata feeds on a water-rich prey: a single adult pea aphid contains over 25 times more water (289 mg ± 24; N = 3; unpublished data) than a spermatophore capsule (Table 6.1).

6.4.3 Two aspects of condition dependence

Condition may influence mating behaviour in two distinct ways. First, reduced condition means a smaller pool of resources, and this may mean a decreased budget for the expression of sexual traits. Second, reduced condition may alter the optimal allocation of resources to traits and thus relative trait expression.

I suggest that the multivariate analysis presented here captures these two dimensions of condition dependence. In this sense, PC1 reflects a change in the size of the resource pool because (i) increasing PC1 scores indicate increasing values for copulation duration and three ejaculate variables (Table 6.2) and (ii) PC1 is positively correlated with post-mating survival, which suggests that it reflects a male’s resources available for somatic maintenance. By contrast, PC2 reflects a shift in allocation strategy: increasing scores here signify a shift in allocation to sperm and away from other

90 ejaculate components, and PC2 shows no correlation with post-mating survival, indicating no relationship to overall male somatic quality.

6.4.4 Conclusion

Characterizing condition dependence in ejaculate traits is an exciting challenge because these traits are important arenas for sperm competition, cryptic female choice and sexual conflict (Chapman & Davies 2004; Eberhard 2009). I have provided the first evidence, to my knowledge, supporting the prediction that condition influences allocation to distinct ejaculate components. Future work is needed to investigate the condition dependence of both ejaculate and sperm quality, traits that are highly plastic in other contexts (e.g. Cornwallis & O’Connor 2009). With increasing empirical evidence of condition and context-dependent variation in ejaculates, there is also a need for theoretical work addressing adaptive ejaculate composition (e.g. Alonzo & Pizzari 2010).

6.5 Appendix

6.5.1 Appendix 6.A: Supplementary analysis

In the main paper, we used PCA to summarize correlations among four traits: the mass of retained ejaculate, the number of sperm transferred, copulation duration and the mass of the ejected spermatophore capsule (Table 6.2). Spermatophore capsule mass data was available for a smaller subset of the overall data (e.g., because some females do not eject a capsule after mating; further, 17 spermatophore capsules were excluded because female beetles ingested part of the capsule before we removed it). We conducted a second analysis using only the first three variables above, in order to confirm that the results we obtained (Table 6.2) are consistent when a larger dataset is analyzed.

The first two principal components (PCs) from this reduced PCA captured 76% of the variation. The results mirror those reported in the main paper. As with the PCA incorporating all four variables, the first PC axis in the reduced PCA reflected positive correlations between copulation duration, ejaculate mass and sperm transfer, while the second axis captured a negative correlation between sperm transfer and ejaculate mass (Table 6.A.1); i.e., a decrease in sperm concentration that appears uncorrelated to copulation duration.

These new PC axes showed similar relationships with the feeding treatment to those reported in the main paper: the feeding treatment influenced both PC1 and PC2 (Fig. 6.A.2, PC1: F 1,71 = 10.1, P <

0.01; PC2: F 1,71 = 14.5, P < 0.001), but not PC3 (F 1,71 = 0.2, P = 0.69). Males in good condition had, on average, positive scores on PC1 (0.18 ± 0.17), indicating that they had higher values for copulation duration, sperm content and ejaculate mass, while males in poor condition had, on average, negative PC1 scores (-0.55 ± 0.16), indicating decreased values for these three variables (Fig. 6.A.2). Similarly, males in good condition had on average positive scores on PC2 (0.43 ±0.16) while poor condition males had negative scores (-0.42 ± 0.15; Fig. 6.A.2). On this second axis, these scores again reflected a pattern of decreased sperm concentration for high food males, and increased sperm concentration for low-food males (Table 6.A.1).

As in the analysis reported in the main paper, male survival was positively correlated with PC1 (R2 = 0.28, N = 40, β = 1683 ± 437, P < 0.001) and not correlated with PC2 (R2 = 0.03, N = 40, β = 563 ± 512, P = 0.28).

6.6 Figures

Figure 6.1

Mean effect of a low- or high-food treatment for males on allocation to ejaculate components. (a) mass of the ejaculate retained within the female after copulation; (b) mass of the spermatophore capsule ejected after mating; (c) number of sperm in the female reproductive tract after mating; (d) sperm concentration. 95% CI are indicated.

Figure 6.2

The effect of a low- or high-food treatment (solid and dashed lines, respectively) on male survival following exposure to a physiological stressor (40 °C for 1 h), with confidence intervals indicated.

Figure 6.3

Principal component scores for low- (circles) and high- (diamonds) food males, from an analysis summarizing variation among males in four mating variables: the mass of retained ejaculate, number of sperm transferred during mating, mass of the spermatophore capsule and copulation duration. Eigenvalues and eigenvectors are given in Table 6.2.

95 Number of sperm transferred

Figure 6.A.1

The threshold relationship between the number of sperm transferred to females during mating and the latency period until the ejection of the spermatophore capsule after mating. Low-food males are indicated by circles and high-food males by diamonds.

3.0

2.5

2.0

1.5

1.0

0.5

0.0

-0.5

-1.0

-1.5

-2.0

-2.5

-3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 Component 1

Figure 6.A.2

Principal component scores for low (circles) and high (diamonds) food males, from an analysis of the mass of retained ejaculate, number of sperm transferred during mating and copulation duration. Table 6.A.1 provides eigenvalues and eigenvectors.

6.7 Tables

Table 6.1

Mating behaviour, ejaculate variables and pre-mating mass of males assigned to a low- or high-food treatment prior to mating. Means are given with s.e., or confidence intervals where means are back- transformed; an exception is that medians are given for male post-mating survival in the last row. (Mixed models included the fixed factor ‘feeding treatment’ and the random factor ‘replicate’. Means for copulation duration are back-transformed from square-root. ‘Retained ejaculate’ refers to ejaculate retained within the female, as opposed to the ejected spermatophore capsule. I was unable to obtain sperm data for four low-food males.) Significant P-values are given in italic. The table continues onto the following page.

Male food treatment

Response Low food High food Model Test P-value statistic

Male pre-mating mass 8.27 mg ± 9.03 mg ± mixed model F1,161 =12.2 <0.0001 0.28 0.28

2 Likelihood of mating 66/83 82/86 chi-square χ 1=9.7 <0.01

Copulation duration 99 min (82, 137 min mixed model F1,144 =17.4 <0.0001 116) (120, 157)

Retained ejaculate

Mass as a proportion of 0.029 ± 0.045 ± ANOVA F1,75 =12.3 <0.0001 male mass before 0.003 0.003 mating

-1 -1 Rate of ejaculate transfer 2.23 µg min 3.37 µg min ANOVA F1,75 =3.7 0.06 (1.59, 2.97) (2.48, 4.41)

Male food treatment

Response Low food High food Model Test P-value statistic

Sperm transfer

2 Likelihood of sperm 22/39 22/34 chi-square χ 1=0.5 0.47 transfer

Rate of transfer (number 0.74 ± 0.13 0.36 ± 0.12 ANOVA F1,26 =4.7 0.04 of sperm per minute)

Ejected spermatophore capsules

2 Likelihood of ejection 38/63 63/82 chi-square χ 1=5.8 0.02

Latency until ejection 13 ± 3 min 12 ± 2 min Wilcoxon test Z=0.8 0.40

Mass as a proportion of 0.16 ± 0.02 0.14 ± 0.02 ANOVA F1,40 =0.9 0.36 total ejaculate mass

Water content 10.8 µg ± 1.5 6.4 µg ± 1.3 ANOVA F1,26 =5.0 0.03

2 Male post-mating survival 1610 min 6670 min proportional χ 1=46.5 <0.0001 (1217, (5931, hazards 2637) 7799)

Table 6.2

Eigenvalues and eigenvectors for the principal components (PCs) from an analysis of male mating performance. Mass of retained ejaculate refers to ejaculate components retained by females after mating (e.g. seminal fluids).

PC1 PC2

Eigenvalue 1.6 1.0

% variation explained 38.9 26.1

Variable

Mass of retained ejaculate 0.355 0.785

Number of sperm transferred 0.482 -0.619

Spermatophore capsule mass 0.590 0.001

Copulation duration 0.541 0.035

100

Table 6.A.1

Eigenvalues and eigenvectors for the principal components (PCs) generated by an analysis of three male mating variables for males. Retained ejaculate refers to ejaculate material retained within the female reproductive tract after copulation.

PC 1 PC 2

Eigenvalue 1.25 1.04

Percent of variation explained 41.7 34.7

Variable:

Mass of retained ejaculate 0.402 0.782

Number of sperm transferred 0.548 -0.622

Copulation duration 0.733 0.036

101

Chapter 7 General conclusions

Is it not pleasing to know that nature is full of examples of such beautifully effective machinery as the human eye and hand and the pony fish's light to hide by? And is it not satisfying to know that human ingenuity can explore and understand these marvels?

George C. Williams, 1996, p. 16

7 General conclusions

This thesis evaluates the potential for conflict and condition-dependence in the sexual traits of Adalia bipunctata . Three of the studies represent a detailed investigation of nuptial feeding by females on spermatophores. Two others examine the extent and implications of condition dependence in male and female traits (ejaculate composition and re-mating resistance, respectively). Taken together, these studies highlight (1) the value of economic and functional analyses for assessing sexual conflict over mating and reproduction, and (2) the significance of condition-dependence for sexually selected traits.

7.1 Major findings

7.1.1 The economics of nuptial feeding in Adalia bipunctata

The collection of studies presented here introduces a new system – ladybird beetles – for the study of nuptial gifts (a system with exciting potential; see section 7.3.1). Together, these studies represent a comprehensive assessment of the economics of gifts in A. bipunctata . I found that spermatophore ingestion induces changes in females that are not consistent with a nutritional function, but are consistent with the effects of seminal proteins: increased re-mating resistance and accelerated oviposition (with no change in lifetime fecundity or longevity) (Chapters 2, 3). These effects are expected to increase male fitness. However, the data show that female fitness, as measured here, is

102 not increased (as predicted by cooperation) or decreased (as predicted by conflict) by spermatophore ingestion (Chapter 2). Thus, spermatophore gifts enhance male fitness with little clear benefit or cost to females (as the benefits or costs to females from accelerated oviposition and increased re-mating resistance are unknown). A natural question is: what maintains feeding on gifts, given that females seemingly do not benefit?

In economic studies, the experimenter has the challenge of ensuring that she has measured all relevant costs and benefits, with sufficient statistical power. This is a nearly hopeless task. The best one can do is to measure effects on fitness components and those phenotypes expected to be important for fitness, and to conduct experiments in a range of contexts expected to influence costs and benefits. The experiments presented here did not assess indirect benefits to females from ingesting spermatophores. Such benefits might occur if females gain information about male quality from spermatophores (Sozou and Seymour 2005) and if females can bias sperm usage towards high quality males. Indeed, spermatophore size and composition does reflect male condition (Chapter 6). However, other aspects of male mating performance are also condition-dependent (e.g., copulation duration and ejaculate mass; Chapter 6), so it is not easy to imagine that any extra information about male quality gleaned from spermatophores provides enough benefit to explain the evolution of spermatophore feeding.

Another type of economic study I conducted was to experimentally ask females how much they value gifts as nutrition, by testing for differences in mating rate between food-deprived and satiated females (Gwynne 2008). Chapter 5 takes this approach. Hungry females were more likely to resist mating, implying that they do not value spermatophores as nutrition, which supports the conclusions of Chapters 2 and 3. This evidence does not eliminate the possibility of a small nutritional benefit to spermatophores: if female foraging efficiency is hampered during copulation, hungry females might profit more from foraging encumbered by a male than from seeking a mating as a source of food. Nonetheless, no benefit is suggested.

The results from Chapters 2 and 3 raise a natural question: what maintains feeding on gifts, if females do not benefit nutritionally? In Chapter 4 I test a sensory bias hypothesis for feeding on gifts that do not benefit females: females may eat spermatophores after mating as a by-product of a general increase in foraging after mating, possibly induced by mating itself. The results show that female A. bipunctata do display a generalized increase in feeding in response to mating; however, this response cannot account for nuptial feeding because recently mated females appear to bias their

103 feeding away from consuming spermatophores. The study thus leaves open the question of why female A. bipunctata ingest spermatophores. Hypotheses include (1) minute nutritional benefits not detected in experiments (Chapters 2, 3) and not reflected in the mating behaviour of hungry females (Chapter 5); (2) benefits to females from the increase in re-mating resistance and accelerated oviposition caused by ingesting gifts, or an alternative form of benefits not detectable by these experiments (see above); and (3) a different form of sensory bias, such as a spermatophore capsule that appeals to female taste perception (Sakaluk 2000), for which there is some preliminary support in A. bipunctata (unpublished data) as well as evidence from other species (Wada-Katsumata et al. 2009; Warwick et al. 2009).

Beyond nuptial feeding, the post-mating elevation in feeding is itself an interesting phenotype of potentially broad relevance. Increased feeding after mating may provide a link between nutritional ecology and sexual selection research (see Morehouse et al. 2010), and characterizing the feeding response in agriculturally relevant insects (including A. bipunctata and other coccinellids) and disease vectors may suggest management approaches (Sirot et al. 2006; Wong and Wolfner 2007).

7.1.2 Condition dependence in male and female traits

Chapter 5 demonstrates the condition-dependent nature of female re-mating resistance, and supports the hypothesis that the function of female resistance is to reduce mating rate rather than to select among males. This conclusion is based on the assumption that mating is costly in A. bipunctata . While my study of multiple mating did not detect longevity costs of mating (Chapter 3), mating may incur other costs for females, particularly if the population harbors sexually transmitted parasite mites known to reduce female fecundity (Hurst et al. 1995). Chapter 5 also showed that female mating resistance generates selection favouring large males as a byproduct: these males were better able to overcome resistant females, but did not have a mating advantage with non-resistant females. Thus, condition-dependence in female resistance generated condition-dependent sexual selection on male size. This result underscores the need for experiments to resolve the basis of male mating advantages: a mating advantage for larger or higher condition males does not necessarily result from active female preference for those males.

Chapter 6, in turn, provides evidence for a condition-dependent shift in ejaculate composition, with males in poor condition transferring ejaculates with higher sperm but lower seminal protein content. This outcome is consistent with two lines of thought: seminal proteins may be more costly to

104 produce than sperm, or there may be a negative correlation between individual condition and the extent of sperm competition faced by a male’s ejaculate (Cameron et al. 2007). Support for the latter hypothesis comes from the observation that low-condition males produce fewer and smaller spermatophore gifts, which increase female re-mating resistance (Chapter 6); further study is needed to resolve these hypotheses. Regardless, the results show that surprising patterns that can emerge from condition-dependent trait expression – it is certainly counter-intuitive that low-condition males should transfer more sperm!

7.2 Implications and contributions

7.2.1 Sexual conflict and nuptial feeding in Adalia bipunctata

Ultimately, what can be said about the extent of conflict in shaping spermatophore feeding in A. bipunctata ? The studies that I have presented show that spermatophore ingestion affects female behaviour in males’ favour, and that females may gain little or nothing from spermatophore feeding. While other benefits and costs to females are conceivable, my studies covered a wide range of expected costs and benefits. With the absence of detectable harm or benefit to females, the results give a picture of neutral effects of spermatophore feeding for females, and thus no extant conflict. This outcome does not preclude the possibility that conflict has influenced the evolution of spermatophore feeding; indeed, the most plausible scenarios for the evolution of nuptial gifts in general invoke conflict at some juncture (Arnqvist and Nilsson 2000; Gwynne 1997; Sakaluk 2000; Simmons and Gwynne 1991; Simmons and Parker 1989). It is possible that females evolve resistance to harmful effects of male traits, including gifts, such that only neutral or positive effects remain (Arnqvist and Rowe 2002; Cordero 1998; Partridge and Hurst 1998; Rice 2000), and there is some evidence for the evolution of such resistance in at least one nuptial gift system (Sakaluk et al. 2006).

7.2.2 Sexual conflict and nuptial gifts

The role of sexual conflict in the origin and maintenance of nuptial gifts has been a topic of considerable debate in recent years (see reviews by Gwynne 2008; Vahed 2007). It has long been recognized that conflict has probably influenced the origin of gifts, and that some features of gifts are undoubtedly shaped by conflict: as examples, male hanging flies offer a prey item to females but attempt to take it back when copulation ends (Thornhill 1976), and the gummy spermatophylax of

105 some male orthopterans (Heller et al. 1998) seems designed to maximize female handling time during copulation. It was thought, though, that even if gift design could conceivably be improved from a female’s perspective, gifts provided females with nutrition and did not have other detrimental effects, and thus, on the whole, gift ingestion increased female fitness.

Increasing interest in sexual conflict prompted research that questioned this view. Studies in Drosophila melanogaster showed that male accessory gland proteins had strong and sometimes harmful effects on females (Chapman 2001). This work emphasized that males could influence female behaviour via chemicals, and that not all female responses to males were in females’ interests. Moreover, empirical studies of nuptial gifts began to yield results that were not consistent with the hypothesis that the effects of gifts on females were nutritional and only nutritional. An increasing number of experimental studies failed to detect the expected nutritional benefits to females from gifts (although it should be emphasized that some studies clearly show such benefits in other species; section 2.1); and more studies began to report effects of gifts that clearly benefited males (e.g., decreased female re-mating rate, advanced reproductive schedule) but were of questionable benefit to females. Some of my work (Chapters 2, 3) falls into these categories.

These patterns suggest some interesting questions. Are these non-nutritional effects of gifts harmful to females? If so, are gifts on the whole beneficial or harmful to females? Two points are worth considering here. First, nuptial gifts have evolved independently in a huge range of taxa. They are not a set of homologous phenotypes and there is no particular reason to expect they will be functionally homologous either. Thus, comparative studies such as Arnqvist and Nilsson’s (2000) meta-analysis, while providing much useful insight, probably capture much variation in function. That study in particular found positive effects of multiple mating in nuptial feeding species, on average, but this does not mean that gifts will improve female fitness as a rule. An alternative possibility is that the net effect of gifts on females varies across species. Second, many studies assessing the costs and benefits of gifts do not isolate the effect of gifts on females. For example, some studies do not separate the effects of gifts from the effects of mating and ejaculate receipt (section 2.1). For many species we need direct tests of the effects of gifts on females, and we need data on a more complete range of female responses, including lifespan, re-mating behaviour, offspring size, egg fertility and reproductive schedule. Very few species have been fully evaluated in this way. The picture suggested by my own results (acknowledging that other interpretations are

106 possible) is that spermatophores are neutral for females and that males take advantage of female spermatophore ingestion to alter female behaviour in their favour.

7.2.3 Condition-dependence in sexual traits

Condition-dependence has a long history in sexual selection. This thesis contributes to the view that condition-dependence can contribute to variation in selection on sexual traits and thus, potentially, the maintenance of variation (at the phenotypic level, at least). Results from Chapter 6 also exemplify that condition-dependence can generate surprising shifts in mating phenotypes.

7.3 Directions for future research

7.3.1 Mating and spermatophore feeding in ladybird beetles

Two topics will be particularly interesting to investigate further in the coccinellid mating system. First, the potential for gifts to influence post-copulatory processes is intriguing. In other species, gifts transferred during mating are known to influence sperm transfer (e.g., in the scorpionfly Panorpa similis , Kullman and Sauer 2009), but I am not aware of any data addressing the hypothesis that gift size or quality influences how females store and utilize sperm from multiple mating partners. Such an effect would be of interest as a mechanism of post-copulatory sexual selection, and preliminary data for A. bipunctata indicates that ingesting spermatophores may have some influence on paternity (unpublished data).

Second, the incredible diversity of ejaculate feeding behaviours in the Coccinellidae (Table 7.1) deserves attention. Comparative work may elucidate how each stage of the behaviour has evolved – spermatophore formation, spermatophore or ejaculate ejection, and spermatophore feeding – as well as the evolution of female responses to spermatophore feeding. Of particular interest would be multiple origins of these behaviours (e.g., Gwynne 1997), as well as transitions from feeding on ejaculate to the absence of feeding, which would suggest that spermatophore feeding provides females with no direct or indirect benefits or is even harmful. The ease of working with coccinellids in the lab, along with extensive variation in behaviours and current research towards a detailed phylogeny (e.g., Magro et al. 2009), makes them a promising system for future comparative research in mating behaviour evolution.

107

7.3.2 The evolution of nuptial gifts

In addition to the need for studies directly testing the effects of gifts on multiple female responses identified above (section 7.2.2), two lines of investigation would be especially interesting. First, a finding of manipulative or signaling proteins in a nuptial gift itself would be a most exciting result. It would represent a new pathway by which males can gain access to female regulatory systems, potentially gaining influence over female reproduction and life history attributes through the digestive tract. Such a route is quite plausible – proteins in the size class of seminal proteins pass through the gut wall in active form in other insects (e.g. Jeffers and Roe 2008). There have been no biochemical studies of gifts that tackle this question, to my knowledge.

Second, given the number of studies that have detected no benefit to nuptial feeding (section 2.1), identifying the selection pressures that maintain nuptial feeding in these species becomes more pressing. Without an answer to this question, it is difficult to say that we truly understand the evolution and function of nuptial feeding.

7.4 Fin

The studies of nuptial gifts in this thesis contribute to a growing view that both conflict and cooperation can play a role in nuptial gift function. These studies are the first to consider spermatophore feeding by coccinellid beetles in an evolutionary ecological perspective. They set a foundation for continuing research in this group, and further study has great potential for additional insights. Overall, this thesis emphasizes that economic studies that directly assess costs and benefits are a useful approach for investigating the function of sexual traits, and that experimental manipulations of individual condition can yield novel insights into both trait function and sexual selection.

108

7.5 Tables

Table 7.1

Ejaculate ejection and feeding behaviours in the Coccinellidae, based on published sources and my unpublished data. There are three entries for Coccinella septempunctata because there are three dissimilar reports for this species, possibly indicating variation among populations. Blank cells indicate information that was not reported or is not applicable.

Species Material ejected Latency until Feeding Source ejection behaviour

Chilocorinae 1

Chilocorus discoideus spermatophore capsule 24 h none Fisher 1959

Coccidulinae 1

Lindorus lophanthae None unpublished data

Coccinellinae

Adalia bipunctata spermatophore capsule several very frequent section 2.2.1 minutes (> 90%)

Coccinella spermatophore capsule infrequent Obata and Johki septempunctata (1/9) 1991

spermatophore capsule 60-180 min none Omkar and Srivastava 2002

None unpublished data

Coccinella trifasciata None unpublished data

109

Species Material ejected Latency until Feeding Source ejection behaviour

Coleomegilla maculata gelatinous material several very frequent unpublished data minutes

Cycloneda munda None unpublished data

Halyzia 16-guttata gelatinous material Majerus 1994b

Harmonia axyridis spermatophore capsule 15-60 min very frequent Obata 1987, unpublished data

Hippodamia convergens None unpublished data

Hippodamia glacialis None unpublished data

Hippodamia variegata None unpublished data

Menochilus sexmaculatus None Obata and Johki 1991

Neoharmonia venusta spermatophore capsule very frequent unpublished data

Propylea japonica Liquid 70-389 min very frequent Obata and Johki 1991

Propylea gelatinous material 1-20 min very frequent unpublished data quatuordecimpunctata

Psyllobora None unpublished data vigintimaculata

Scymninae 1

Brachiacantha ursina spermatophore capsule 3-52 min very frequent unpublished data

110

Species Material ejected Latency until Feeding Source ejection behaviour

Cryptolaemus clear or cloudy liquid, or several very frequent unpublished data montrouzieri spermatophore capsule minutes

Stethorus punctillum None unpublished data

Sticholotidinae

Delphastus catalinae None unpublished data

1 A recent phylogeny has shown that these subfamilies are paraphyletic (Magro et al. 2009)

111

Literature Cited

Acar, E. B., D. D. Mill, B. N. Smith, L. D. Hansen, and G. M. Booth. 2005. Comparison of respiration in adult Harmonia axyridis Pallas and Hippodamia convergens Guerrin-Manaville (Coleoptera: Coccinellidae). Environmental Entomology 34:241-245. Alatalo, R. V., J. Höglund, and A. Lundberg. 1988. Patterns of variation in tail ornament size in birds. Biological Journal of the Linnean Society 34:363-374. Alonzo, S. H., and T. Pizzari. 2010. Male fecundity stimulation: conflict and cooperation within and between the sexes: model analyses and coevolutionary dynamics. American Naturalist 175:174-185. Andersson, M. 1994. Sexual selection: Monographs in Behavior and Ecology. Princeton, NJ, Princeton University Press. Andersson, M., and L. W. Simmons. 2006. Sexual selection and mate choice. Trends in Ecology and Evolution 21:296-302. Arnaud, L., Y. Spinneux, and E. Haubruge. 2003. Preliminary observations of sperm storage in Adalia bipunctata (Coleoptera: Coccinellidae): sperm size and number. Applied Entomology and Zoology 38:301-304. Arnqvist, G. 2006. Sensory exploitation and sexual conflict. Philosophical Transactions of the Royal Society of London B 361:375-386. Arnqvist, G., and T. Nilsson. 2000. The evolution of polyandry: multiple mating and female fitness in insects. Animal Behaviour 60:145-164. Arnqvist, G., and L. Rowe. 2002. Comparative analysis unveils antagonistic coevolution between the sexes in a group of insects. Nature 415:787-789. —. 2005. Sexual conflict: Monographs in behavior and ecology. Princeton, New Jersey, Princeton University Press. Badcock, C., and B. Crespi. 2008. Battle of the sexes may set the brain. Nature 454:1054-1055. Baer, B., E. D. Morgan, and P. Schmid-Hempel. 2001. A nonspecific fatty acid within the bumblebee mating plug prevents females from remating. Proceedings of the National Academy of Sciences 98:3926-3928. Barnes, A. I., S. Wigby, J. M. Boone, L. Partridge, and T. Chapman. 2008. Feeding, fecundity and lifespan in female Drosophila melanogaster . Proceedings of the Royal Society of London B 275:1675-1683.

112

Bergsten, J., and K. B. Miller. 2007. Phylogeny of diving beetles reveals a coevolutionary arms race between the sexes. PLoS ONE 2:e522. Bilde, T., C. Tuni, R. Elsayed, S. Pekar, and S. Toft. 2007. Nuptial gifts of male spiders: sensory exploitation of the female's maternal care instinct or foraging motivation? Animal Behaviour 73:267-273. Blanckenhorn, W. U., D. J. Hosken, O. Y. Martin, C. Reim, Y. Teuschl, and P. I. Ward. 2002. The costs of copulating in the dung fly Sepsis cynipsea . Behavioral Ecology 13:353-358. Blanco, C. A., M. G. Rojas, A. T. Groot, J. Morales-Ramos, and C. A. Abel. 2009. Size and chemical composition of Heliothis virescens (Lepidoptera: Noctuidae) spermatophores. Annals of the Entomological Society of America 102:629-637. Boggs, C. L. 1995. Male nuptial gifts: phenotypic consequences and evolutionary implications, Pages 215—242 in S. R. Leather, and J. Hardie, eds. Insect Reproduction. Boca Raton, FL, CRC Press. Bonduriansky, R. 2003. Layered sexual selection: a comparative analysis of sexual behaviour within an assemblage of piophilid flies. Canadian Journal of Zoology 81:479-491. Bonduriansky, R., and L. Rowe. 2005. Sexual selection, genetic architecture, and the condition dependence of body shape in the sexually dimorphic fly Prochyliza xanthostoma (Piophilidae). Evolution 59:138-151. Bonduriansky, R., J. Wheeler, and L. Rowe. 2005. Ejaculate feeding and female fitness in the sexually dimorphic fly Prochyliza xanthostoma (Diptera: Piophilidae). Animal Behaviour 69:489-497. Brakefield, P. M. 1984. Ecological studies on the polymorphic ladybird Adalia bipunctata in the Netherlands. 1. Population biology and geographical variation of melanism. Journal of Animal Ecology 53:761-774. Brown, W. D. 1997. Courtship feeding in tree crickets increases insemination and female reproductive life-span. Animal Behaviour 54:1369-1382. Browne, L. B. 1993. Physiologically induced changes in resource-oriented behavior. Annual Review of Entomology 38:1-25. —. 1995. Ontogenic changes in feeding behaviour, Pages 307-342 in R. F. Chapman, and G. de Boer, eds. Regulatory Mechanisms in Insect Feeding. London, Chapman & Hall. Cameron, E., T. Day, and L. Rowe. 2003. Sexual conflict and indirect benefits. Journal of Evolutionary Biology 16:1055-1060. —. 2007. Sperm competition and the evolution of ejaculate composition. American Naturalist 169:E158-E172.

113

Carvalho, G. B., P. Kapahi, D. J. Anderson, and S. Benzer. 2006. Allocrine modulation of feeding behavior by the sex peptide of Drosophila . Current Biology 16:692-696. Chapman, T. 2001. Seminal fluid-mediated fitness traits in Drosophila . Heredity 87:511-521. Chapman, T., G. Arnqvist, J. Bangham, and L. Rowe. 2003a. Response to Eberhard and Cordero, and Córdoba-Aguilar and Contreras-Garduño: sexual conflict and female choice. Trends in Ecology and Evolution 18:440-441. —. 2003b. Sexual conflict. Trends in Ecology and Evolution 18:41-47. Chapman, T., J. Bangham, G. Vinti, B. Seifried, O. Lung, M. F. Wolfner, H. K. Smith et al. 2003c. The sex peptide of Drosophila melanogaster : female post-mating responses analyzed by using RNA interference. Proceedings of the National Academy of Sciences 100:9923-9928. Chapman, T., L. F. Liddle, J. M. Kalb, M. F. Wolfner, and L. Partridge. 1995. Cost of mating in Drosophila melanogaster females is mediated by male accessory-gland products. Nature 373:241- 244. Chapman, T., and L. Partridge. 1996. Female fitness in Drosophila melanogaster : an interaction between the effect of nutrition and of encounter rate with males. Proceedings of the Royal Society of London B 263:755-759. Chapman, T. W., and S. J. Davies. 2004. Functions and analysis of the seminal fluid proteins of male Drosophila melanogaster fruit flies. Peptides 25:1477-1490. Choe, J. C., and B. J. Crespi. 1997. The Evolution of Mating Systems in Insects and Arachnids. Cambridge, Cambridge University Press. Christy, J. H. 1995. Mimicry, mate choice, and the sensory trap hypothesis. American Naturalist 146:171-181. Cordero, C. 1998. Chemical ornaments of semen. Journal of Theoretical Biology 192:581-584. Cordero, C., and W. G. Eberhard. 2005. Interaction between sexually antagonistic selection and mate choice in the evolution of female responses to male traits. Evolutionary Ecology 19:111-122. Córdoba-Aguilar, A., and J. Contreras-Garduño. 2003. Sexual conflict. Trends in Ecology and Evolution 18. Cornwallis, C. K., and E. A. O'Connor. 2009. Sperm: seminal fluid interactions and the adjustment of sperm quality in relation to female attractiveness. Proceedings of the Royal Society of London B 276:3467-3475.

114

Cotton, S., K. Fowler, and A. Pomiankowski. 2004. Do sexual ornaments demonstrate heightened condition-dependent expression as predicted by the handicap hypothesis? Proceedings of the Royal Society of London B 271:771-783. Crean, C. S., and A. S. Gilburn. 1998. Sexual selection as a side-effect of sexual conflict in the seaweed fly, Coelopa ursina (Diptera: Coelopidae). Animal Behaviour 56:1405-1410. Crudgington, H. S., and M. T. Siva-Jothy. 2000. Genital damage, kicking and early death - the battle of the sexes takes a sinister turn in the bean weevil. Nature 407:855-856. Day, T. H., and A. S. Gilburn. 1997. Sexual selection in seaweed flies. Advances in the Study of Behavior 26:1-57. de Clerq, P., M. Bonte, K. Van Speybroeck, K. Bolckmans, and K. Deforce. 2005. Development and reproduction of Adalia bipunctata (Coleoptera: Coccinellidae) on eggs of Ephestia kuehniella (Lepidoptera: Phycitidae) and pollen. Pest Management Science 61:1129-1132. de Jong, P. W., P. M. Brakefield, and B. P. Geerinck. 1998. The effect of female mating history on sperm precedence in the two-spot ladybird, Adalia bipunctata (Coleoptera, Coccinellidae). Behavioral Ecology 9:559-565. Delisle, J., and A. Bouchard. 1995. Male larval nutrition in Choristoneura rosaceana (Lepidoptera: Tortricidae): an important factor in reproductive success. Oecologia 104:508-517. Dewsbury, D. A. 1982. Ejaculate cost and male choice. American Naturalist 119:601-610. Eberhard, W. G. 1994. Evidence for widespread courtship during copulation in 131 species of insects and spiders, and implications for cryptic female choice. Evolution 48:711-733. —. 1996. Female control: sexual selection by cryptic female choice. Princeton, NJ, Princeton University Press. —. 1997. Sexual selection by cryptic female choice in insects, Pages 32-57 in J. C. Choe, and B. J. Crespi, eds. Social Competition and Cooperation in Insects and Arachnids. I. Evolution of Mating Systems. Cambridge, Cambridge University Press. —. 2002. The function of female resistance behavior: intromission by male coercion vs. female cooperation in sepsid flies (Diptera: Sepsidae). Revista de Biologia Tropical 50:485-505. —. 2004. Rapid divergent evolution of sexual morphology: comparative tests of antagonistic coevolution and traditional female choice. Evolution 58:1947-1970. —. 2009. Postcopulatory sexual selection: Darwin's omission and its consequences. Proceedings of the National Academy of Sciences 106:10025-10032.

115

Eberhard, W. G., and C. Cordero. 1995. Sexual selection by cryptic female choice on male seminal products - a new bridge between sexual selection and reproductive physiology. Trends in Ecology and Evolution 10:493-496. —. 2003. Sexual conflict and female choice. Trends in Ecology and Evolution 18:438-439. Edvardsson, M. 2007. Female Callosobruchus maculatus mate when they are thirsty: resource-rich ejaculates as mating effort in a beetle. Animal Behaviour 74:183-188. Endler, J. A., and A. L. Basolo. 1998. Sensory ecology, receiver biases and sexual selection. Trends in Ecology and Evolution 13:415-420. Engqvist, L. 2007a. Nuptial food gifts influence female egg production in the scorpionfly Panorpa cognata . Ecological Entomology 32:327–332. —. 2007b. Nuptial gift consumption influences female remating in a scorpionfly: male or female control of mating rate? Evolutionary Ecology 21:49-61. —. 2007c. Sex, food and conflicts: nutrition dependent nuptial feeding and pre-mating struggles in scorpionflies. Behavioral Ecology and Sociobiology 61:703-710. Evans, E. W. 2003. Searching and reproductive behaviour of female aphidophagous ladybirds (Coleoptera: Coccinellidae): a review. European Journal of Entomology 100:1-10. Fairbairn, D. J. 1993. Costs of loading associated with mate-carrying in the waterstrider, Aquarius remigis . Behavioral Ecology 4:224-231. Fedina, T. Y., and S. M. Lewis. 2006. Proximal traits and mechanisms for biasing paternity in the red flour beetle Tribolium castaneum (Coleoptera : Tenebrionidae). Behavioral Ecology and Sociobiology 60:844-853. Ferkau, C., and K. Fischer. 2006. Costs of reproduction in male Bicyclus anynana and Pieris napi butterflies: effects of mating history and food limitation. Ethology 112:1117-1127. Fisher, T. W. 1959. Occurrence of spermatophores in certain species of Chilocorus (Coleoptera: Coccinellidae). Pan Pacific Entomologist 35:205-208. Fiumera, A. C., B. L. Dumont, and A. G. Clark. 2006. Natural variation in male-induced ‘cost-of- mating’ and allele-specific association with male reproductive genes in Drosophila melanogaster . Philosophical Transactions of the Royal Society of London B 361:355–361. Fricke, C., J. Perry, T. Chapman, and L. Rowe. 2009. The conditional economics of sexual conflict. Biology Letters 5:671-674. Fuller, R. C., D. Houle, and J. Travis. 2005. Sensory bias as an explanation for the evolution of mate preferences. American Naturalist 166:437-446.

116

Funk, D. H., and D. W. Tallamy. 2000. Courtship role reversal and deceptive signals in the long- tailed dance fly, Rhamphomyia longicauda . Animal Behaviour 59:411-421. Gavrilets, S., G. Arnqvist, and U. Friberg. 2001. The evolution of female mate choice by sexual conflict. Proceedings of the Royal Society of London B 268:531-539. Gavrilets, S., and D. Waxman. 2002. Sympatric speciation by sexual conflict. Proceedings of the National Academy of Sciences 99:10533-10538. Getty, T. 2006. Sexually selected signals are not similar to sports handicaps. Trends in Ecology and Evolution 21:83-88. Gillott, C. 2003. Male accessory gland secretions: modulators of female reproductive physiology and behaviour. Annual Review of Entomology 48:163-184. Gwynne, D. T. 1984a. Courtship feeding increases female reproductive success in bush crickets. Nature 307:361-363. —. 1984b. Sexual selection and sexual differences in Mormon crickets (Orthoptera: Tettigoniidae, Anabrus simplex ). Evolution 38:1011-1022. —. 1986. Courtship feeding in katydids: investment in offspring or in obtaining fertilisations? American Naturalist 128:342-352. —. 1988. Courtship feeding and the fitness of female katydids (Orthoptera: Tettigoniidae). Evolution 42:545-555. —. 1989. Does copulation increase the risk of predation? Trends in Ecology and Evolution 4:54-56. —. 1990. Testing parental investment and the control of sexual selection in katydids: the operational sex ratio. American Naturalist 136:474-484. —. 1997. The evolution of edible 'sperm sacs' and other forms of courtship feeding in crickets, katydids and their kin (Orthoptera: Ensifera), Pages 110-129 in B. J. Crespi, and J. C. Choe, eds. Mating Systems in Insects and Arachnids. —. 2008. Sexual conflict over nuptial gifts in insects. Annual Review of Entomology 53:5.1-5.19. Gwynne, D. T., B. J. Bowen, and C. G. Codd. 1984. The function of the katydid spermatophore and its role in fecundity and insemination (Orthoptera: Tettigoniidae). Australian Journal of Zoology 32:15-22. Haddrill, P. R. 2001. The development and use of molecular genetic markers to study sexual selection and population genetics in the 2-spot ladybird, Adalia bipunctata (L.). PhD thesis, University of Cambridge, Cambridge.

117

Haddrill, P. R., D. M. Shuker, W. Amos, M. E. N. Majerus, and S. Mayes. 2008. Female multiple mating in wild and laboratory populations of the two-spot ladybird, Adalia bipunctata . Molecular Ecology 17:3189-3197. Haddrill, P. R., D. M. Shuker, S. Mayes, and M. Majerus. 2007. Temporal effects of multiple mating on components of fitness in the two-spot ladybird, Adalia bipunctata (Coleoptera: Coccinellidae). European Journal of Entomology 104:393-398. Heller, K.-G., S. Faltin, P. Fleischmann, and O. v. Helversen. 1998. The chemical composition of the spermatophore in some species of phaneropterid bushcrickets (Orthoptera: Tettigonioidea). Journal of Insect Physiology 44:1001-1008. Hodek, I. 1973. Biology of Coccinellidae. Prague, Academia. Hoenig, J. M., and D. M. Heisey. 2001. The abuse of power: the pervasive fallacy of power calculations for data analysis. The American Statistician 55:19-24. Holland, B., and W. R. Rice. 1998. Perspective: chase-away sexual selection: antagonistic seduction versus resistance. Evolution 52:1-7. Hosken, D. J., O. Y. Martin, J. Born, and F. Huber. 2003. Sexual conflict in Sepsis cynipsea : female reluctance, fertility and mate choice. Journal of Evolutionary Biology 16:485-490. Hosken, D. J., P. Stockley, T. Tregenza, and N. Wedell. 2009. Monogamy and the battle of the sexes. Annual Review of Entomology 54:361-378. Hurst, G. D. D., M. E. N. Majerus, and L. E. Walker. 1992. Cytoplasmic male killing elements in Adalia bipunctata (Linnaeus) (Coleoptera: Coccinellidae). Heredity 69:84-91. Hurst, G. D. D., R. G. Sharpe, A. H. Broomfield, L. E. Walker, T. M. O. Majerus, I. A. Zakharov, and M. E. N. Majerus. 1995. Sexually transmitted disease in a promiscuous insect, Adalia bipunctata . Ecological Entomology 20:230-236. Isaac, R. E., C. Li, A. E. Leedale, and A. D. Shirras. 2009. Drosophila male sex peptide inhibits siesta sleep and promotes locomotor activity in the post-mated female. Proceedings of the Royal Society of London B 277:65-70. Iwasa, Y., and A. Pomiankowski. 1994. The evolution of mate preferences for multiple sexual ornaments. Evolution 48:853-867. Iwasa, Y., A. Pomiankowski, and S. Nee. 1991. The evolution of costly mate preferences. II. The ‘handicap’ principle. Evolution 45:1431-1442. Ja, W. W., G. B. Carvalho, B. M. Zid, E. M. Mak, T. Brummel, and S. Benzer. 2009. Water- and nutrient-dependent effects of dietary restriction on Drosophila lifespan. Proceedings of the National Academy of Sciences 106:18633-18637.

118

Jalali, M. A., L. Tirry, and P. de Clercq. 2009. Food consumption and immature growth of Adalia bipunctata (Coleoptera: Coccinellidae) on a natural prey and a factitious food. European Journal of Entomology 106:193-198. Jeffers, L. A., and R. M. Roe. 2008. The movement of proteins across the insect and tick digestive system. Journal of Insect Physiology 54:319-332. Jeffers, L. A., D. M. Thompson, D. Ben-Yakir, and R. M. Roe. 2005. Movement of proteins across the digestive system of the tobacco budworm, Heliothis virescens . Entomologia Experimentalis et Applicata 117:135-146. Jia, Z., Z. Jiang, and S. K. Sakaluk. 2000. Nutritional condition influences investment by male katydids in nuptial food gifts. Ecological Entomology 25:115-118. Johnstone, R. A., and L. Keller. 2000. How males can gain by harming their mates: sexual conflict, seminal toxins, and the cost of mating. American Naturalist 156:368-377. Jormalainen, V. 1998. Precopulatory mate guarding in crustaceans: male competitive strategy and intersexual conflict. Quarterly Review of Biology 73:275-304. Kasuya, E., and N. Sato. 1998. Effects of the consumption of male spermatophylax on the oviposition schedule of females in the decorated cricket, Gryllodes sigillatus . Zoological Science 15:127-130. Kaufman, W. R. 2007. Gluttony and sex in female ixodid ticks: How do they compare to other blood-sucking arthropods? Journal of Insect Physiology 53:264-273. Kimura, M., and Y. Ihara. 2009. Replicator-dynamics models of sexual conflict. Journal of Theoretical Biology 260:90-97. Kokko, H., R. Brooks, M. Jennions, and J. Morley. 2003. The evolution of mate choice and mating biases. Proceedings of the Royal Society of London B 270:653-664. Kokko, H., M. D. Jennions, and R. Brooks. 2006. Unifying and testing models of sexual selection. Annual Review of Ecology and Systematics 37:43-66. Kokko, H., and D. J. Rankin. 2006. Lonely hearts or sex in the city? Density-dependent effects in mating systems. Philosophical Transactions of the Royal Society of London B 361:319-334. Krupa, J. J., W. R. Leopold, and A. Sih. 1990. Avoidance of male giant water striders by females. Behaviour 115:247-253. Kullman, H., and K. P. Sauer. 2009. Mating tactic dependent sperm transfer rates in Panorpa similis (Mecoptera; Panorpidae): a case of female control? Ecological Entomology 34:153-157. Kura, T., and K. Yoda. 2001. Can voluntary nutritional gifts in seminal flow evolve? Journal of Ethology 19:9-15.

119

Lanzoni, A., G. Accinelli, G. G. Bazzocchi, and G. Burgio. 2004. Biological traits and life table of the exotic Harmonia axyridis compared with Hippodamia variegata , and Adalia bipunctata (Col., Coccinellidae). Journal of Applied Entomology 128:298-306. Lauer, M. J. 1996. Effect of sperm depletion and starvation on the mating behaviour of the water strider, Aquarius remigis . Behavioral Ecology and Sociobiology 39:89-96. Lehmann, G. U. C., and A. W. Lehmann. 2009. Condition-dependent spermatophore size is correlated with male’s age in a bushcricket (Orthoptera: Phaneropteridae). Biological Journal of the Linnean Society 96:354-360. Lewis, Z., and N. Wedell. 2007. Effect of adult feeding on male mating behaviour in the butterfly, Bicyclus anynana (Lepidoptera: Nymphalidae). Journal of Insect Behavior 20:201-213. Liu, H. F., and E. Kubli. 2003. Sex-peptide is the molecular basis of the sperm effect in Drosophila melanogaster . Proceedings of the National Academy of Sciences 100:9929-9933. Magro, A., E. Lecompte, F. Magné, J.-L. Hemptinne, and B. Crouau-Roy. 2009. Phylogeny of ladybirds (Coleoptera: Coccinellidae): are the subfamilies monophyletic? Molecular Phylogenetics and Evolution 54:833-848. Majerus, M. E. N. 1994a. Female promiscuity maintains high fertility in ladybirds (Col., Coccinellidae). Entomologist's Monthly Magazine 130:205-209. —. 1994b. Ladybirds. London, HarperCollins. Martin-Coello, J., H. Dopazo, L. Arbiza, J. Ausio, E. R. S. Roldan, and M. Gomendio. 2009. Sexual selection drives weak positive selection in protamine genes and high promoter divergence, enhancing sperm competitiveness. Proceedings of the Royal Society of London B 276:2427- 2436. McGraw, L. A., A. C. Fiumera, M. Ramakrishnan, S. Madhavarapu, A. G. Clark, and M. F. Wolfner. 2007. Larval rearing environment affects several post-copulatory traits in Drosophila melanogaster . Biology Letters 3:607-610. McKinney, F., S. R. Derrickson, and P. Mineau. 1983. Forced copulation in waterfowl. Behaviour 86:250-294. McMillan, S., A.-K. Kuusk, A. Cassel-Lundhagen, and B. Ekbom. 2007. The influence of time and temperature on molecular gut content analysis: Adalia bipunctata fed with Rhopalosiphum padi . Insect Science 14:353-358. Mesnick, S. L., and B. J. Le Boeuf. 1991. Sexual behavior of male northern elephant seals: II. Female response to potentially injurious encounters. Behaviour 117:262-280.

120

Moehrlin, G. S., and S. A. Juliano. 1998. Plasticity of insect reproduction: testing models of flexible and fixed development in response to different growth rates. Oecologia 115:492-500. Mondet, C., D. Abed-Vieillard, P. Gautier, and J.-P. Farine. 2008. Could male tergal secretions be considered as a nuptial gift in the Madeira cockroach? Animal Behaviour 75:451-460. Morehouse, N. I., T. Nakazawa, C. M. Booher, P. D. Jeyasingh, and M. D. Hall. 2010. Sex in a material world: why the study of sexual reproduction and sex-specific traits should become more nutritionally-explicit. Oikos 119:766-778. Mühlhäuser, C., and W. U. Blanckenhorn. 2002. The costs of avoiding matings in the dung fly Sepsis cynipsea . Behavioral Ecology 13:359-365. Obata, S. 1987. Mating behavior and sperm transfer in the ladybird beetle, Harmonia axyridis Pallas (Coleoptera: Coccinellidae). Applied Entomology and Zoology 22:434-442. Obata, S., and Y. Johki. 1991. Comparative study on copulatory behaviour in four species of aphidophagous ladybirds, Pages 207-212 in L. Polgár, R. J. Chambers, A. F. G. Dixon, and I. Hodek, eds. Behaviour and impact of Aphidophaga. The Hague, The Netherlands, SPB Academic Publishing. Omkar, and G. Mishra. 2005. Mating in aphidophagous ladybirds: costs and benefits. Journal of Applied Entomology 129:432-436. Omkar, G. Mishra, and S. K. Singh. 2006. Optimal number of matings in two aphidophagous ladybirds. Ecological Entomology 31:1-4. Omkar, and A. Pervez. 2005. Ecology of two-spotted ladybird, Adalia bipunctata : a review. Journal of Applied Entomology 129:465. Omkar, and S. Srivastava. 2002. The reproductive behaviour of an aphidophagous ladybeetle, Coccinella septempunctata (Coleoptera: Coccinellidae). European Journal of Entomology 99:465- 470. Ono, T., C. Ando, and Y. Kondo. 2004. Effect of metanotal secretion ingestion on oviposition in a tree cricket, Truljalia hibinonis (Orthoptera: Gryllidae). Entomological Science 7:9-13. Ortigosa, A., and L. Rowe. 2002. The effect of hunger on mating behaviour and sexual selection for male body size in Gerris buenoi . Animal Behaviour 64:369-375. —. 2003. The role of mating history and male size in determining mating behaviours and sexual conflict in a water strider. Animal Behaviour 65:851-858. Owen, D. H., and D. F. Katz. 2005. A review of the physical and chemical properties of human semen and the formulation of a semen simulant. Journal of Andrology 26:459-469.

121

Parker, G. A. 1979. Sexual selection and sexual conflict, Pages 123-166 in M. S. Blum, and N. A. Blum, eds. Sexual Selection and Reproductive Competition in Insects. New York, Academic Press. —. 1990. Sperm competition games: raffles and roles. Proceedings of the Royal Society of London B 242:120-126. —. 2006. Sexual conflict over mating and fertilization: an overview. Philosophical Transactions of the Royal Society of London B 361:235-259. Partridge, L., and L. D. Hurst. 1998. Sex and conflict. Science 281:2003-2008. Perez-Staples, D., A. M. T. Harmer, S. R. Collins, and P. W. Taylor. 2008. Potential for pre-release diet supplements to increase the sexual performance and longevity of male Queensland fruit flies. Agricultural and Forest Entomology 10:255-262. Perry, J. C., and L. Rowe. 2008a. Ingested spermatophores accelerate reproduction and increase mating resistance but are not a source of sexual conflict. Animal Behaviour 76:993-1000. —. 2008b. Neither mating rate nor spermatophore feeding influences longevity in a ladybird beetle. Ethology 114:504-511. —. 2010. Condition-dependent ejaculate size and composition in a ladybird beetle. Proceedings of the Royal Society of London B. doi:10.1098/rspb.2010.0810 Perry, J. C., D. M. T. Sharpe, and L. Rowe. 2009. Condition-dependent female remating resistance generates sexual selection on male size in a ladybird beetle. Animal Behaviour 77:743-748. Pitcher, T. E., B. D. Neff, F. H. Rodd, and L. Rowe. 2003. Multiple mating and sequential mate choice in guppies: females trade up. Proceedings of the Royal Society of London B 270:1623-1629. Plaistow, S. J., L. Bollache, and F. Cezilly. 2003. Energetically costly precopulatory mate guarding in the amphipod Gammarus pulex : causes and consequences. Animal Behaviour 65:683-691. Preston-Mafham, K. G. 1999. Courtship and mating in Empis (Xanthempis ) trigramma Meig., E. tessellata F. and E. ( Polyblepharis ) opaca F. (Diptera: Empididae) and the possible implications of `cheating' behaviour. Journal of Zoology 247:239-246. Proulx, S., T. Day, and L. Rowe. 2002. Older males signal more reliably. Proceedings of the Royal Society of London B 269:2291-2299. Punzalan, D., M. Cooray, F. H. Rodd, and L. Rowe. 2008. Condition dependence of sexually dimorphic colouration and longevity in the ambush bug Phymata americana . Journal of Evolutionary Biology 21:1297-1306.

122

Ramm, S. A., P. L. Oliver, C. P. Ponting, P. Stockley, and R. D. Emes. 2007. Sexual selection and the adaptive evolution of mammalian ejaculate proteins. Molecular Biology and Evolution 25:207-219. Ransford, M. O. 1997. Sperm competition in the 2-spot ladybird, Adalia bipunctata , University of Cambridge, Cambridge. Ravi Ram, K., and M. F. Wolfner. 2007. Seminal influences: Drosophila Acps and the molecular interplay between males and females during reproduction. Integrative and Comparative Biology 47:427–445. Reinhold, K. 1999. Paternal investment in Poecilimon veluchianus bushcrickets: beneficial effects of nuptial feeding on offspring viability. Behavioral Ecology and Sociobiology 45:293-299. Rice, W. R. 1992. Sexually antagonistic genes: experimental evidence. Science 256:1436-1439. —. 1998. Male fitness increases when females are eliminated from gene pool: implications for the Y chromosome. Proceedings of the National Academy of Sciences 95:6217-6221. —. 2000. Dangerous liaisons. Proceedings of the National Academy of Sciences 97:12953-12955. Rice, W. R., A. D. Stewart, E. H. Morrow, J. E. Linder, N. Orteiza, and P. G. Byrne. 2006. Assessing sexual conflict in the Drosophila melanogaster laboratory model system. Philosophical Transactions of the Royal Society of London B 361:287-299. Rönn, J., M. Katvala, and G. Arnqvist. 2006. The costs of mating and egg production in Callosobruchus seed beetles. Animal Behaviour 72:335-342. —. 2007. Coevolution between harmful male genitalia and female resistance in seed beetles. Proceedings of the National Academy of Sciences 104:10921-10925. Rowe, L. 1992. Convenience polyandry in a water strider: foraging conflicts and female control of copulation frequency and guarding duration. Animal Behaviour 44:189-202. —. 1994. The cost of mating and mate choice in water striders. Animal Behaviour 48:1049-1056. Rowe, L., G. Arnqvist, A. Sih, and J. J. Krupa. 1994. Sexual conflict and the evolutionary ecology of mating patterns: water striders as a model system. Trends in Ecology and Evolution 9:289- 293. Rowe, L., E. Cameron, and T. Day. 2005. Escalation, retreat, and female indifference as alternative outcomes of sexually antagonistic coevolution. American Naturalist 165:S5-S18. Rowe, L., and T. Day. 2006. Detecting sexual conflict and sexually antagonistic coevolution. Philosophical Transactions of the Royal Society of London B 361:277-285. Rowe, L., and D. Houle. 1996. The lek paradox and the capture of genetic variance by condition dependent traits. Proceedings of the Royal Society of London B 263:1415-1421.

123

Rowe, L., J. J. Krupa, and A. Sih. 1996. An experimental test of condition-dependent mating behavior and habitat choice by water striders in the wild. Behavioral Ecology 7:474-479. Sakaluk, S. K. 2000. Sensory exploitation as an evolutionary origin to nuptial food gifts in insects. Proceedings of the Royal Society of London B 267:339-343. Sakaluk, S. K., R. L. Avery, and C. B. Weddle. 2006. Cryptic sexual conflict in gift-giving insects: chasing the chase-away. American Naturalist 167:94-104. Sakaluk, S. K., M. T. H. Campbell, A. P. Clark, J. C. Johnson, and P. A. Keorpes. 2004. Hemolymph loss during nuptial feeding constrains male mating success in sagebrush crickets. Behavioral Ecology 15:845-849. Shuker, D. M., and T. H. Day. 2001. The repeatability of a sexual conflict over mating. Animal Behaviour 61:755-762. Simmons, L. W. 1988. The contribution of multiple mating and spermatophore consumption to the lifetime reproductive success of female field crickets ( Gryllus bimaculatus ). Ecological Entomology 13:57-69. —. 1990. Nuptial feeding in tettigoniids: male costs and the rates of fecundity increase. Behavioral Ecology and Sociobiology 27:43-47. —. 2001. Sperm competition and its evolutionary consequences in the insects. Princeton, New Jersey, Princeton University Press. Simmons, L. W., and W. J. Bailey. 1990. Resource influenced sex roles of zaprochiline tettigoniids (Orthoptera: Tettigoniidae). Evolution 44:1853-1868. Simmons, L. W., L. Beesley, P. Lindhjem, D. Newbound, J. Norris, and A. Wayne. 1999. Nuptial feeding by male bushcrickets: an indicator of male quality? Behavioral Ecology 10:263-269. Simmons, L. W., and D. Gwynne. 1991. The refractory period of female katydids (Orthoptera: Tettigoniidae): sexual conflict over the remating interval? Behavioral Ecology 2:276-282. Simmons, L. W., and J. S. Kotiaho. 2007. Quantitative genetic correlation between trait and preference supports a sexually selected sperm process. Proceedings of the National Academy of Sciences 104:16604-16608. Simmons, L. W., and G. A. Parker. 1989. Nuptial feeding in insects: mating effort versus paternal investment. Ethology 81:332-343. Sirot, L. K., S. L. Lapointe, R. Shatters, and M. Bausher. 2006. Transfer and fate of seminal fluid molecules in the beetle, Diaprepes abbreviatus : Implications for the reproductive biology of a pest species. Journal of Insect Physiology 52:300-308.

124

Sloggett, J. J., and M. E. N. Majerus. 2000. Habitat preferences and diet in the predatory Coccinellidae (Coleoptera): an evolutionary perspective. Biological Journal of the Linnean Society 70:63-88. Sozou, P. D., and R. M. Seymour. 2005. Costly but worthless gifts facilitate courtship. Proceedings of the Royal Society of London B 272:1877-1844. Stearns, S. C. 1992. The evolution of life histories. Oxford, Oxford University Press. Stewart, L. A., A. F. G. Dixon, Z. Ruzicka, and G. Iperti. 1991. Clutch and egg size in ladybird beetles. Entomophaga 36:329-333. Stone, G. N. 1995. Female foraging responses to sexual harassment in the solitary bee Anthophora plumipes . Animal Behaviour 50:405-412. Tatarnic, N. J., G. Cassis, and D. F. Hochuli. 2006. Traumatic insemination in the plant bug genus Coridromius Signoret (Heteroptera: Miridae). Biology Letters 2:58-61. Tazzyman, S. J., T. Pizzari, R. M. Seymour, and A. Pomiankowski. 2009. The evolution of continuous variation in ejaculate expenditure strategy. American Naturalist 174:E71-E82. Teuschl, Y., and W. U. Blanckenhorn. 2007. The reluctant fly: what makes Sepsis cynipsea females willing to copulate? Animal Behaviour 73:85-97. Thornhill, R. 1976. Sexual selection and nuptial feeding behavior in Bittacus apicalis (Insecta: Mecoptera). American Naturalist 110:529-548. —. 1984. Alternative female choice tactics in the scorpionfly Hylobittacus apicalis (Mecoptera) and their implications. American Zoologist 24:367-383. Thornhill, R., and J. Alcock. 1983. The Evolution of Insect Mating Systems. Cambridge, Mass., Harvard University Press. Tomlinson, I. P. M., P. W. E. Kearns, and C. J. Veltman. 1995. Nonrandom mating in the 2-spot ladybird ( Adalia bipunctata ) - the influence of weight on mating success. Behavior Genetics 25:467-474. Tuckerman, J. F., D. T. Gwynne, and G. K. Morris. 1993. Reliable acoustic cues for female mate preference in a katydid ( Scudderia curvicauda , Orthoptera: Tettigoniidae). Behavioral Ecology 4:106-113. Ursprung, C., M. den Hollander, and D. T. Gwynne. 2009. Female seed beetles, Callosobruchus maculatus , remate for male-supplied water rather than ejaculate nutrition. Behavioral Ecology and Sociobiology 63:781-788. Vahed, K. 1998. The function of nuptial feeding in insects: a review of empirical studies. Biological Reviews 73:43-78.

125

—. 2003. Increases in egg production in multiply mated female bushcrickets Leptophyes punctatissima are not due to substances in the nuptial gift. Ecological Entomology 28:124-128. —. 2007. All that glisters is not gold: sensory bias, sexual conflict and nuptial feeding in insects and spiders. Ethology 113:105-127. Vahed, K., and F. S. Gilbert. 1997. No effect of nuptial gift consumption on female reproductive output in the bushcricket Leptophyes laticauda Friv. Ecological Entomology 22:479-482. Vermeulen, A., S. Engels, and K. P. Sauer. 2008. Mating effort and cryptic sperm choice in scorpionflies: male investment strategy vs. female control. Ethology 114:1166-1172. Wada-Katsumata, A., M. Ozaki, F. Yokohari, M. Nishikawa, and R. Nishida. 2009. Behavioral and electrophysiological studies on the sexually biased synergism between oligosaccharides and phospholipids in gustatory perception of nuptial secretion by the German cockroach. Journal of Insect Physiology 55:742-750. Warwick, S., K. Vahed, D. Raubenheimer, and S. J. Simpson. 2009. Free amino acids as phagostimulants in cricket nuptial gifts: support for the ‘Candymaker’ hypothesis. Biology Letters 5:194-196. Watanabe, M., and M. Hirota. 1999. Effects of sucrose intake on spermatophore mass produced by male swallowtail butterfly Papilio xuthus L. Zoological Science 16:55-61. Watson, P. J., G. Arnqvist, and R. R. Stallman. 1998. Sexual conflict and the energetic costs of mating and mate choice in water striders. American Naturalist 151:46-58. Webberley, K. M., J. Buszko, V. Isham, and G. D. D. Hurst. 2006. Sexually transmitted disease epidemics in a natural insect population. Journal of Animal Ecology 75:33-43. Wedell, N. 1993. Mating effort or paternal investment? Incorporation rate and cost of male donations in the wartbiter. Behavioral Ecology and Sociobiology 32:239-246. Wedell, N., and A. Arak. 1989. The wartbiter spermatophore and its effect on female reproductive output (Orthoptera: Tettigoniidae, Decticus verrucivorus ). Behavioral Ecology and Sociobiology 24:117-125. Wedell, N., T. Tregenza, and L. W. Simmons. 2008. Nuptial gifts fail to resolve a sexual conflict in an insect. BMC Evolutionary Biology 8:204. West-Eberhard, M. J. 1983. Sexual selection, social competition, and speciation. Quarterly Review of Biology 58:155-183. West-Eberhard, M. J. 1984. Sensory bias as an explanation for the evolution of mate preferences, Pages 283-384 in T. Lewis, ed. Insect Communication. New York, Academic Press.

126

Wigby, S., and T. Chapman. 2005. Sex peptide causes mating costs in female Drosophila melanogaster . Current Biology 15:316-321. Wigby, S., L. K. Sirot, J. R. Linklater, N. Buehner, F. C. F. Calboli, A. Bretman, M. F. Wolfner et al. 2009. Seminal fluid protein allocation and male reproductive success. Current Biology 19:751-757. Wiley, R. H., and J. Poston. 1996. Perspective: indirect mate choice, competition for mates, and coevolution of the sexes. Evolution 50:1371-1381. Will, M. W., and S. K. Sakaluk. 1994. Courtship feeding in decorated crickets: is the spermatophylax a sham? Animal Behaviour 48:1309-1315. Williams, G. C. 1996. Plan and purpose in nature. Weidenfeld & Nicolson, London. Williams, P. D., T. Day, and E. Cameron. 2005. The evolution of sperm-allocation strategies and the degree of sperm competition. Evolution 59:492-499. Winston, P. W., and D. H. Bates. 1960. Saturated solutions for the control of humidity in biological research. Ecology 41:232-237. Wolfner, M. F. 1997. Tokens of love: functions and regulation of Drosophila male accessory gland products. Insect Biochemistry and Molecular Biology 27:179-192. Wong, A., and M. F. Wolfner. 2007. Sexual behavior: a seminal peptide stimulates appetites. Current Biology 16:R256-R257. Yasuda, H., and A. F. G. Dixon. 2002. Sexual size dimorphism in the two spot ladybird beetle Adalia bipunctata : developmental mechanism and its consequences for mating. Ecological Entomology 27:493-498. Zahavi, A. 1975. Mate selection - a selection for a handicap. Journal of Theoretical Biology 53:205- 214. Zeh, J. A., and D. W. Zeh. 2003. Toward a new sexual selection paradigm: polyandry, conflict and incompatibility. Ethology 109:929-950.