Jump to ContentJump to Main Navigation
The Evolution of Insect Mating Systems$

David Shuker and Leigh Simmons

Print publication date: 2014

Print ISBN-13: 9780199678020

Published to Oxford Scholarship Online: October 2014

DOI: 10.1093/acprof:oso/9780199678020.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 25 September 2017

The evolution of polyandry

The evolution of polyandry

Chapter 9 The evolution of polyandry
The Evolution of Insect Mating Systems

Rhonda R. Snook

Oxford University Press

Abstract and Keywords

Polyandry is widespread, but its origin and maintenance remains enigmatic. Direct benefits arising from polyandry are easy to understand, but the recent emphasis on sexual conflict over mating decisions suggests that polyandry to gain indirect benefits either through intrinsic male quality or through genetic compatibility is limited. Meta-analyses have found strong support for direct benefits and weak support for indirect benefits in insects, but some individual studies show that indirect benefits can outweigh both direct costs and benefits. A quantitative genetics framework must recognize that the economy of polyandry is context dependent, changing in response to spatio-temporal variation in ecology. Experimental evolution and genomics techniques have improved our understanding of polyandry, but they are currently limited to only a few model insect organisms. One challenge for future research is to integrate these techniques and observations into a framework that predicts the cost:benefit relationship of female multiple mating

Keywords:   polyandry, direct benefits, indirect benefits, intrinsic male quality, genetic compatibility, quantitative genetics, genomics, experimental evolution

The Evolution of Insect Mating Systems. Edited by David M. Shuker and Leigh W. Simmons.

© The Royal Entomological Society 2014. Published 2014 by Oxford University Press.

9.1 Introduction

We saw in Chapter 3 how insect mating systems encompass monogamy through to polyandry. Historically, monogamy was the presumed predominant mating pattern in females, but close observations and the subsequent advent of molecular techniques to allocate parentage has demonstrated that polyandry is rampant (Birkhead and Møller 1998). This is perhaps the single most important empirical advance since Thornhill and Alcock’s (1983) volume. Classifying mating systems is rife with terminology, including several definitions for polyandry (Snook 2013; Chapter 3). However, the definition most widely used today is from Thornhill and Alcock (1983); females have more than one male as a mate during a breeding season (pp. 81–82). While polyandry is now recognized as being widespread, its origin and maintenance remain enigmatic due to variation between the sexes in the costs and benefits of reproduction. Central to the problem are theories on the evolution of anisogamy, Bateman’s principles (Bateman 1948) and parental investment (Trivers 1972). Anisogamy is the defining feature of males and females, wherein males produce small but numerous gametes and females produce large, nutritious but generally rather few gametes. This fundamental difference between the sexes tends to limit female reproduction more strongly than male reproduction, with further delineation (or indeed reversal) of these ‘sex roles’ arising through subsequent variation in parental investment associated with ecological circumstances (Trivers 1972). Bateman’s principles derive from observations, including Bateman’s original experiments on Drosophila, which show that male fitness increases with an increasing number of mates whereas female fitness does not. These sex differences predict that male reproductive success will exhibit greater variance than female reproductive success. As a consequence, sexual selection will act more strongly on males than females. These patterns will be opposite in sex-role-reversed species, where males become the limiting sex and there is greater variance in success in competition for mates among females than males. Additionally, whereas sex is costly for both sexes, females appear to bear the sharp end of this expense. Aside from the energy and time expenditure required to engage in copulation (Thornhill and Alcock 1983), polyandrous females may, for example, experience greater predation (Arnqvist 1989), be injured (e.g. Eberhard 1996), and have reduced longevity as a consequence of the receipt of caustic male seminal fluids (e.g. Chapman et al. 1995). Sexual conflict over mate number can have profound influences on the behaviour, morphology, and physiology of insects (Arnqvist and Rowe 2005). Together, these data indicate that typically males should benefit most from (p.159) (p.160) multiple mating whereas females should benefit most from, and thus are predicted to exhibit, monogamy. However, females are rarely monogamous.

When predictions (females should be monogamous) do not match empirical data (females mate with more than one male), alternative explanations need to be sought. One such alternative is that the entire edifice (the trinity of anisogamy—Bateman’s principles—parental investment) on which the predictions are made is false. Recently, several researchers have suggested this alternative, criticizing the data and interpretation of Bateman’s own experiment (Tang-Martinez 2010; Gowaty et al. 2012). Some of these criticisms are justified—for example, problems with the genetic strains used—but others, such as a biblical-like literal interpretation that females do not benefit at all from multiple mating, are not. Whereas male fitness generally increases linearly with increasing number of mates, female fitness can increase asymptotically with number of mates, an empirical result found by Bateman (1948). Regardless of the experimental sins of Bateman, sufficient additional studies across different taxa by different researchers have supported the underlying principles—the non-limiting sex tends to gain more from multiple mating than the limiting sex—such that rejecting the entire idea is unwarranted.

However, it is likewise true that not all systems studied have conformed to Bateman’s principles (Snyder and Gowaty 2013), that males can be choosy and females compete for males (Clutton-Brock 2009), and that polyandry can change the relative strength of sexual selection on each sex. For example, if females benefit from polyandry, then females may compete for access to multiple males resulting in increased sexual selection on females while polyandry simultaneously limits the male’s ability to monopolize females, potentially reducing the strength of sexual selection acting on males (for review, see Kvarnemo and Simmons 2013). Thus, sex differences in competition and preference are more nuanced than the traditional Bateman view accommodates.

Regardless of the underlying evolutionary rationale for polyandry, the consequences of female multiple mating are indisputable. As discussed in Chapters 10 and 11, polyandry gives rise to post-copulatory sexual selection in both sexes and sexual conflict between the sexes, resulting in both competition within and co-evolution between the sexes. Indeed, the raison d’etre for Thornhill and Alcock’s book was to ‘. . . attempt to explain characteristics bordering on the bizarre (the fantastic horns of certain male beetles, the complex penis structure of damselflies, the presentation of food gifts to females by male scorpionflies, and week-long copulations in walking sticks) as the products of social [sexual] selection for traits useful in an intensely competitive and coevolving sexual environment.’ (Thornhill and Alcock, 1983, p. 54). Thirty years later, we understand that insect polyandry has additional impacts on, and is affected by, a variety of different evolutionary processes: the maintenance of genetic variability (Jennions and Petrie 2000 for review); sociality (Hughes et al. 2008; see Chapter 14); sex allocation (Ratnieks and Boomsma 1995); the spread of selfish genetic elements (Price et al. 2008a; see Chapter 13); speciation (Martin and Hosken 2003; Bacigalupe et al. 2007); and inbreeding (Michalczyk et al. 2011).

The widespread influence of polyandry on a variety of different evolutionary phenomena highlights just how important it is for us to understand this fundamental phenomenon. Studies on the evolutionary causes and consequences of polyandry have accumulated rapidly over the past 30 years, revealing a tremendous amount about the raison d’etre for those bizarre behaviours and structures that so fascinated Thornhill and Alcock and their early students. The goal of this chapter is to discuss a variety of recent developments regarding our understanding of polyandry, including perspectives that were nascent in Thornhill and Alcock’s book, and to draw attention to areas of growing interest to which a new (p.161) student of polyandry can contribute. To that end, this chapter includes (i) a brief consideration of the rise of studies of sexual conflict over mating decisions between the sexes that affect the economics of polyandry, (ii) how adoption of a quantitative genetics framework can elucidate the relative contribution of different indirect benefits and why this is important, (iii) the use of modern techniques, from experimental evolution through to genomics, for understanding the underlying phenotypic and genetic responses and potential constraints of polyandry, and (iv) how the economics of polyandry within and between species is affected by a variety of context-dependent factors. Consideration of these topics may open up new research areas for elucidating the evolutionary significance of polyandry. Such understanding hinges on the relatively poorly studied underlying genetic causes and consequences of female multiple mating.

9.2 Quantifying polyandry

Three different aspects of female reproduction can contribute to the effective levels of polyandry: the proportion of females in a population that remate, the number of males that an individual female mates with (or at the population level, the average number of males), and the (average) number of sires. Various different techniques can assess one or more of these components and may be applied to studies of both natural and laboratory populations. For example, in wild populations of lepidopterans, the number of times a female has mated can be determined by dissecting the female reproductive tract and counting the number of spermatophore remnants, as these remain in the female for her lifetime (for review, see Simmons 2001). However, this technique cannot determine whether these multiple matings were with the same individual or different males or whether, and the extent to which, sperm from each mating are used.

A second method applied in both wild and laboratory populations is to subject focal individuals to a continuous supply of the opposite sex to estimate the maximum number of mates within a given time. For example, in the seaweed fly, Coelopa frigida, whose mating system is characterized as convenience polyandry (see Section 9.3.2), Blyth and Gilburn (2006) collected wild males and females and measured in the field the number of pairs in which a male mounted a female within 5 min. These data suggested that, on average, males will mount a female approximately every 9 min; taking this into account together with rates of both male and female rejection of the mounts, an individual female was estimated to mate more than 30 times a day.

Various genetic techniques have been used to estimate the number of males a female has mated with. These originally included allozymes but have now moved on to microsatellite markers, including competitive polymerase chain reaction (PCR) techniques applied to sperm stored by females (Bretman and Tregenza 2005). The latter technique allows for an extensive evaluation of not only whether females are polyandrous but also whether there is any sperm sorting across sperm storage organs. For example, in the yellow dung fly, Scathophaga stercoraria, wild caught females were collected across a spring season and frozen upon collection, with the three spermathecae later dissected individually and the stored sperm extracted. Competitive PCR was then used to assign the number of males that females had mated with and whether sperm from these males were distributed equally between the three spermathecae. The researchers found that the proportion of females multiply mated increased sharply at the beginning of the season, remaining high, until shortly before the last sampling day. The average number of ejaculates ranged from 2.47 to 3.33 based on assumptions underlying the estimation process. Additionally, the number (p.162) of ejaculates stored differed between the spermathecae, with the singlet spermatheca storing generally fewer sperm than either the inner or outer doublet (Demont et al. 2011). Results from the wild population mainly conformed to previous results based on laboratory populations (Demont et al. 2011).

A new technique to examine polyandry and its consequences in the wild is the combined use of DNA profiling and video monitoring of wild populations (Rodríguez-Muñoz et al. 2010). Two generations of a Spanish population of the flightless field cricket, Gryllus campestris, were observed with videos used to assess such factors as dominance, number of mates, and calling behaviour, combined with DNA profiling of the subsequent generation to assign parentage. Males had significantly greater variance in offspring number relative to females, confirming Bateman’s principle in a wild population, but there was no difference between the sexes in the variance in the number of mates, contradicting Bateman’s principle. The study also found that both males and females benefit from multiple mating by increasing the number of offspring, although what generates this increase is unclear. Females could benefit directly either through ejaculate donations or sperm replenishment, or indirectly either via good genes or via compatibility genes (see Sections 9.3.4 and 9.3.5).

As the ability to use more sophisticated techniques in the field improves, our understanding of what drives polyandry can only gain. Currently, many studies examining the evolutionary significance of female multiple mating rely on laboratory populations. While these are useful in allowing controlled conditions under which a researcher can isolate and assess the relative importance of one aspect of reproductive performance, these results may not always reflect what happens in the multivariate space of natural populations where both natural and sexual selection operate. For example, in S. stercoraria, patterns of polyandry and sperm storage in the field matched what was found in laboratory populations, although fertilization patterns and costs and benefits of polyandry have only been estimated in the laboratory (Demont et al. 2011). Hence, the relationships between polyandry, sperm storage and fitness cannot be compared between the field and laboratory populations. In contrast, the positive effect of dominance and male singing on male reproductive success in laboratory populations of Gryllus was not seen in the wild populations of G. campestris, for various reasons (Rodríguez-Muñoz et al. 2010) which can be followed up through experimental manipulations in a laboratory setting. A more integrated approach between field and laboratory studies of the evolutionary significance of polyandry is a crucial endeavour to generate far-reaching insights into sexual selection acting on natural populations of insects (e.g. Bretman and Tregenza 2005; Demont et al. 2011).

9.3 The evolutionary causes of polyandry

9.3.1 Origin versus maintenance of polyandry

A number of hypotheses have been proposed to explain the origin of polyandry: that is, why selection should favour the evolution of females multiply mating with different males (Table 9.1). However, most research has aimed primarily at studying the current utility of the phenomenon. While the origin and maintenance of polyandry may be closely linked, contemporary selection may not reflect historical selection, shifting the functional significance of polyandry across the evolutionary history of a species.

Table 9.1 Some direct and indirect benefits associated with the origin and maintenance of polyandry

Type of benefit

Evolutionary rationale for polyandry




Fertility assurance

Ensure lifetime sperm supply

Increased number of fertilized eggs compared to singly mated females or females mated multiply to same male

Paternal donations

Transfer of nutrients

Increased fecundity and/or longevity as a result of donated nutrients

Paternal donations

Transfer of chemicals

Decreased predation risk of remating females

Convenience polyandry

Females acquiesce to remating

Increased longevity compared to females that are courted by males but do not remate


Trading up

Females remate with partners of better genetic quality

Offspring with increased viability or increased reproductive success

Genetic diversity/genetic bet-hedging

Females guard against future environmental change by increasing genetic diversity of offspring

Increased probability of some offspring surviving

Good sperm/intrinsic male quality

Females remate with males to select superior fertilizing sperm which reflects male genetic quality

Increased offspring viability

Sexy sperm

Females remate with males to promote increased fertilization efficiency (?) arising from sperm competition.

Sons with increased fertilization success and daughters that promote sperm competition

Fertilization success?

Genetic compatibility

Females multiply mate to increase genetic heterozygosity and avoid inbreeding through beneficial maternal and paternal genome combinations

Increased offspring fitness

The origin of traits can be addressed using comparative methods, but, given the near ubiquity of polyandry, historical analyses may not be exceptionally informative. An (p.163) alternative is to study species that are either monogamous or in which polyandry has evolved relatively recently, in order to understand what forces promote the evolution of a particular mating system. Studying these cases may shed light on the extent to which the origin and maintenance of the phenomenon are aligned, although currently this insight is limited.

For example, in the jewel wasp, Nasonia vitripennis, wild females are generally monogamous and laboratory females are reluctant to remate (Burton-Chellew et al. 2007). However, the longer the period of time in which strains are kept in the laboratory, the more frequently females mate with multiple males (Burton-Chellew et al. 2007). Additional work has shown that this response is heritable, although largely driven by non-additive effects (Shuker et al. 2007). The costs and benefits to polyandry in this system have not yet been studied, so the reason why polyandrous behaviour increases in laboratory culture remains unknown.

(p.164) In Drosophila subobscura, females are resolutely monogamous (Smith 1956; Holman et al. 2008), yet this species shows nuptial feeding—a hallmark of direct benefits (Section 9.3.4) for mating multiply. Indeed, whereas direct benefits are dependent on nutritional status of both the male and female, females can increase fecundity from the nuptial gift and experience no identified mating cost (Immonen et al. 2009). Thus, in D. subobscura, females should benefit from multiple mating, yet they do not mate multiply.

Mating systems, despite often being described categorically, are actually a continuum encompassing spatio-temporal variation in ecology within and among populations of the same species. Thus, the general rule is likely to be that neither exclusive monandry nor polyandry is observed within most species. Indeed, in many insect groups, populations differ in the extent to which females engage in polyandry; and this may be heritable, as in N. vitripennis (Shuker et al. 2007). Thus, many species may have a plastic mating system in which the costs and benefits of mating are altered according to the particular number or quality of those partners. This understanding requires studying the conditions under which such plasticity is manifest, keeping in mind that the origin and maintenance of polyandry can be conflated.

9.3.2 Non-adaptive explanations for polyandry: the rise of sexual conflict

One early explanation for polyandry was that female mating rate was genetically correlated with male mating rate and was thus non-adaptive for females (Halliday and Arnold 1987). Although plausible, artificial selection experiments in which proxies for mating frequency were manipulated in Drosophila melanogaster (e.g. Gromko and Newport 1988) found no genetic correlation between male and female mating rate. Two subsequent selection experiments on the stalk-eyed fly Cyrtodiopsis dalmanni directly selecting for mating frequency (Grant et al. 2005), and on the adzuki bean beetle Callosobruchus chinensis selecting for female receptivity (Harano and Miyatake 2007a), also found no genetic correlation in mating frequency between the sexes. Thus, overall there is currently no support for polyandry arising as a correlated response to selection on increased male mating frequency, albeit with relatively limited data.

Thornhill and Alcock (1983) recognized the phenomenon of convenience polyandry in which females acquiesce to matings to alleviate constant and costly male harassment. In this case, females are making the ‘best of a bad situation’ and do not gain any benefits from remating. For example, in the seaweed fly, C. frigida, males do not perform any courtship behaviour and simply attempt to mount any female they come into contact with (for references see Blyth and Gilburn 2006). Females resist mating by shaking, kicking, and curling their abdomen, an energetically costly rejection response. Mating decreases longevity and wild females may experience more than 30 matings per day (Blyth and Gilburn 2006).

Convenience polyandry represents one manifestation of sexual conflict—an area of evolutionary biology that has been integral to recent studies of polyandry but not well-incorporated at the time of Thornhill and Alcock because its importance was not widely appreciated at that time (see Arnqvist and Rowe 2005). As discussed in Chapter 2, sexual conflict occurs when the fitness optima of the sexes differ (Parker 1979). A vast literature is now accumulating that documents the existence of sexual conflict over mating, including evidence from both comparative and experimental data (for review, see Arnqvist and Rowe 2005). Such evidence includes costs of mating to females, conflicts between the sexes over mating frequency, and the reproductive tactics employed by the sexes either to persuade or to resist mating. For example, mating itself may be differentially costly to females by increasing predation probability, as in Gerris water striders (Arnqvist 1989), or decreasing (p.165) longevity due to physical damage to the female during mating, as seen in the bean weevil Callosobruchus maculatus (Crudgington and Siva-Jothy 2000), or chemical damage to the female after mating as seen in Drosophila melanogaster (e.g. Chapman et al. 1995). It is likely that many of these costs are collateral consequences of male–male competition; in other words a negative side-effect on females due to a conflict-driven adaptation in males, such as persistent courtship of females as seen in the drowning of female yellow dung flies, Scathophaga stercoraria, in the dung pats during mating struggles between males (see Chapter 10; reviewed by Arnqvist and Rowe 2005). Females are expected to evolve counter-adaptations to limit fitness costs, following which males may evolve a more powerful form of the costly trait, or selection may favour a different manipulative trait (Parker 1979, Holland and Rice 1998).

One recent example of such sexually antagonistic co-evolution can be found in the red-backed water strider, Gerris gracilicornis. Males of many Gerris species forcibly mate females whose genitalia are exposed. In G. gracilicornis, females have evolved concealed genitalia preventing male intromission (Figure 9.1; Han and Jablonski 2009). Females allow males access for mating only after the male performs tapping behaviour with his mid-legs, producing ripples on the water surface, which the authors interpret as a ‘courtship signal’. Thus, females in this species have evolved a structure that prevents male manipulation. Subsequent work, however, found that the male courtship signal was actually a form of intimidation of the female. The tapping attracted predatory backswimmers, Notonecta triguttata (p.166) (Figure 9.2a), which targeted female but not male water striders as prey. Females reduce predation risk by allowing males to copulate (Figure 9.2b; Han and Jablonski 2010).

The evolution of polyandry

Figure 9.1 Genital segments of Gerris lacustris (a), a typical Gerris, and in G. gracilicornis (b), in which females have hidden genitalia from Segment 7 (S7). The vulvar opening (Vo) is normally hidden by the gonocoxa 1 (Gx1) which is partially inflated here (c). Scale bar: 0.1 mm. Modified from Han and Jablonski (2009).

The evolution of polyandry

Figure 9.2 Female response to predator experience and male signals in Gerris gracilicornis. (a) Mating pair of water striders, female on the bottom, with a backswimming notonectid predator (from http://blogs.discovermagazine.com/notrocketscience/2010/08/10/male-water-striders-summon-predators-to-blackmail-females-into-having-sex/#.UXUX76LCZ8G). (b) Females were exposed to four fixed-order treatments and their latency to protrude their genitalia was measured. Both predator experience and male signalling influenced latency to protrude genitalia. Genitalia protrusion latency decreased when females had experienced a predator (Control vs Treatment 1; *P 〈 0.05) but latency substantially increased if males were experimentally manipulated so that they could not tap their legs (Treatment 1 vs Treatment 3; ***P 〈 0.001). Boxes indicate the 25% and 75% quartiles with the solid line indicating the median and the dotted line the mean. Black dots represent outliers. Modified from Han and Jablonski (2010). Reproduced with permission from Nature Publishing Group.

Conflict over mating rate is a form of interlocus conflict (see Chapter 2). However, when examining the functional significance of polyandry, intralocus conflict may also affect the economics of female multiple mating. This is because the measures of reproductive success and hence fitness will reflect net selection acting on both sons and daughters, which obviously share some loci associated with offspring fitness, such as longevity and performance. Indeed, studies in insects have shown that there may be negative genetic correlations between son and daughter fitness with high-performing sires giving rise to high-performing sons but low-fitness daughters (e.g. Chippindale et al. 2001). In D. melanogaster, there was a positive correlation between male and female juvenile survival, but adult reproductive success was negatively correlated. While overall total fitness was not correlated between the sexes, substantial genetic × gender interactions were found (Figure 9.3). These results suggest pervasive sexual antagonism between the sexes at the genetic level; negative genetic correlations between the sexes, especially for shared fitness components, are fairly common (Poissant et al. 2010).

The evolution of polyandry

Figure 9.3 Significant interaction between relative total fitness rank, combining juvenile survival and adult reproduction, of male and female genomes in Drosophila melanogaster. Relative fitness between the genders frequently reversed, indicating that genomes producing high fitness in one sex can produce low fitness in the other sex. From Chippindale et al. (2001). Reproduced with permission from the US National Academy of Sciences.

Intralocus sexual conflict poses a significant challenge for indirect-benefit models to explain the evolution of polyandry since mating with high-quality males may generate the production of low-quality daughters, cancelling any indirect benefit. Indeed, theory predicts that indirect benefits are likely to be smaller than direct costs of multiple mating (Kirkpatrick and Barton 1997; Cameron et al. 2003). In D. melanogaster, a number of studies from the Rice laboratory have found no evidence for indirect benefits but rather substantial mating costs (e.g. Stewart et al. 2005, 2008).

(p.167) 9.3.3 Overview of adaptive hypotheses for polyandry

A number of adaptive hypotheses that focus on the currency of the fitness advantage have been proposed to explain why selection should favour the evolution of polyandry (Table 9.1). Thornhill and Alcock recognized that polyandry was easily explained if females gained directly from multiple mating either through increased longevity or increased production of offspring. The mechanisms of such benefits are often quite evident (i.e. nuptial feeding in bushcrickets) and relatively easy to study and manipulate (Section 9.3.4).

The conundrum over the evolutionary rationale for polyandry lies in species in which males make no material contribution to female reproduction. In many species, males apparently donate nothing more than the ejaculate, such that material benefits (beyond sperm and seminal fluid) are unlikely. In these cases, alternative indirect (‘genetic’) benefits have been suggested, such that females gain fitness from mating with different partners through increased performance of progeny. This general class of hypotheses has been more difficult to study than material benefits, as both the action and effects of indirect benefits are less conspicuous. Very large, longitudinal studies across generations are required, often only feasible in the laboratory on a subset of amenable species. Moreover, if such benefits accrue via mechanisms of post-copulatory sexual selection such as sperm competition (Chapter 10) and cryptic female choice (Chapter 11), these mechanisms, let alone their outcomes, are themselves typically difficult to study and disentangle. Indirect genetic-benefit models are also controversial because direct costs of mating multiply are predicted to outweigh indirect benefits to females and, whereas some studies in D. melanogaster have supported this prediction (Section 9.3.2), other studies in this species have found that indirect benefits outweigh direct mating costs to females (e.g. Rundle et al. 2007; Priest et al. 2008), and that indirect benefits can sometimes be even greater than direct benefits (Tuni et al. 2013). There is a substantial number of reviews on direct and indirect benefits for polyandry (e.g. Jennions and Petrie 2000; Zeh and Zeh 2001a; Simmons 2005; Puurtinen et al. 2009; Slatyer et al. 2012), so here we briefly cover only the major points.

(p.168) 9.3.4 Direct benefits

In many tettigoniids (bushcrickets), the male transfers a spermatophore containing a nutritious spermatophylax that is attached to the sperm-containing ampulla. After transfer of the spermatophore, females consume the spermatophylax while the sperm drain from the ampulla and are stored in her spematheca. In the Australian bushcricket, Kawanaphila nartee, spermatophylax feeding increases the number and weight of egg (Simmons 1990).

Such material benefits have a direct positive effect on female fitness, through a variety of fitness-related traits, including assuring fertility, increased fecundity, increased offspring number, increased longevity, predator avoidance, and paternal care. Many of these were discussed extensively by Thornhill and Alcock (1983, e.g. pp. 374–390). Recent meta-analyses have found strong support for direct benefits explaining polyandry generally, and in insects in particular (Arnqvist and Nilsson 2000; Slatyer et al. 2012). Many of these studies address the role of nuptial gifts and other forms of male donation in offspring production and female longevity. Here we shall focus on a second less-well-documented direct benefit, fertility assurance, and whether it can explain the origin of polyandry or its maintenance.

In some insect species, such as parasitoid wasps with local mate competition, males may mate with a large number of females in quick succession such that fertility assurance may explain female multiple mating (Boivin 2013). Likewise, females may remate to replenish sperm in the fire ant, Solenopsis invicta. While several papers suggest that S. invicta is predominantly monogamous, approximately 20% of queens from some areas can be polyandrous. Such polyandrous behavior is associated with whether their first mate carried a particular haplotype (Lawson et al. 2012). Males carrying a Gp-9b haplotype produce fewer sperm than males carrying the Gp-9B haplotype. Females mated to Gp-9b males remate more frequently than those mated to Gp-9B males. Thus, females may remate to replenish sperm stores—a direct benefit—although this does not entirely explain paternity skew when females mate with both male types (Lawson et al. 2012), leaving the potential for remating also to increase indirect fitness through selection against selfish genetic elements (Chapter 13).

However, in many insect species, fertility assurance is thought to be an unlikely explanation for the origin of polyandry because many researchers state that males in most (but not all) insect species transfer sufficient sperm for the female’s lifetime in one copulation. One issue is that we generally don’t know what a ‘lifetime’ is. For example, in a laboratory population of D. pseudoobscura, females that had mated just once had reduced average number of offspring per day but no differences in average number of eggs per day compared to females mated multiply with the same male, suggesting sperm limitation (Gowaty et al. 2010). However, this experiment lasted well beyond 45 days and it is unclear how long females live in the wild, or have opportunities to oviposit, as extrinsic mortality present in natural populations is not reflected in laboratory populations. If females on average do not live very long in the wild, or have few opportunities to find oviposition sites, then fertility assurance may be unlikely to explain the origin of polyandry.

Fertility assurance may, however, explain the maintenance of polyandry. While sperm are relatively inexpensive compared to eggs, the ejaculates themselves can be expensive and, to avoid sperm depletion, males of some species have evolved strategic allocation (see Chapter 10) (Wedell et al. 2002). As a consequence of strategic sperm allocation, females may not receive enough sperm per mating. Taken at face value, this then suggests that polyandry is less likely to have evolved originally in order to counteract sperm limitation (as males had plenty of sperm when females only mated once). However, once females become polyandrous, strategic allocation and other factors influencing male ejaculate (p.169) sizes may generate sperm limitation which may reinforce whatever other selection has favoured polyandry. Sperm limitation may then be a factor maintaining polyandry in a population, rather than the original cause of polyandry. Even in sex-role-reversed species where males (and their sperm) are the limiting resource, sperm limitation may still not be the evolutionary origin of polyandry. For example, in some Drosophila species, males produce exceptionally long sperm and relatively few of them, such that females tend to remate often. While this result supports the idea of multiple mating evolving in response to fertility assurance, both theory and empirical results show that long sperm likely evolved in response to the post-copulatory process of cryptic female choice (Miller and Pitnick 2002). Thus, sperm limitation in this case may not be the origin of female multiple mating, but arises as a consequence of females preferring investment in costly long sperm.

9.3.5 Indirect benefits

The functional significance of polyandry in the absence of direct benefits was not readily apparent when Thornhill and Alcock’s book was published. When direct benefits can be ruled out, only indirect (‘genetic’) benefits can offset the costs of polyandry. But whether and to what extent genetic benefits could explain polyandry was very unclear, as good-genes models of sexual selection (and indeed formal models of mate choice more generally; see Chapter 2) were still relatively new and the controversies over good-genes sexual selection only just warming up. Since then, many studies of genetic benefits, and the ongoing attendant controversies surrounding these, have been published.

Selection favours polyandry via genetic benefits when the mean offspring fitness of polyandrous females rises above the average value obtained via monandry, as a consequence of certain genes from the male (intrinsic male quality) or combinations of male and female genes (genetic compatibility). In quantitative genetics terms, these benefits are either additive, that is the effects on reproduction that are due to the male’s breeding value for fitness, or non-additive, that is the effects on reproduction that are due to the genetic interactions between the parents’ haplotypes. We have seen how additive and non-additive genetic benefits influence the evolution of mate choice in Chapter 8. In this section, we briefly discuss how these two dominant hypotheses for indirect benefits might influence the evolution of polyandry.

Polyandry incites sperm competition (Chapter 10) and facilitates cryptic female choice (Chapter 11). These post-copulatory mechanisms can bias paternity, such that males with greater breeding values for fitness are selected. Females may benefit from mating with multiple males either through ejaculate components that are correlated with intrinsic male quality (e.g. ‘good sperm’; Sivinski 1984) or success in sperm competition (‘sexy sperm’; Keller and Reeve 1995). One simple prediction is that if females gain from polyandry through either good or sexy sperm, then female mate choice should be congruent within a population; that is, females should prefer the same male phenotype. These additive good-genes benefits result in directional selection on the male trait which is expected to deplete additive genetic variation in associated fitness-related traits, implying that good-genes benefits should be small. Nonetheless, additive genetic variation for these traits may be quite high, resulting in the so-called lek paradox (see Chapter 8).

Genetic benefits through post-copulatory mechanisms can also be obtained through genetic compatibility which brings favourable parental genetic combinations together and/or avoids unfavourable gene combinations (i.e. the benefits are non-additive genetic benefits). In this case, if females gain compatible alleles from polyandry, then different females should prefer different males from the same population and no directional (p.170) selection on male phenotype occurs. However, it may be difficult to envisage how females can optimize mating for genetic compatibility as the fitness outcomes for offspring may be hard to estimate. If such benefits are associated with a few linked loci, or reflect a relatively simple mechanism to avoid inbreeding, then polyandry may be beneficial. In contrast, if extensive pleiotropy occurs, then polyandry for genetic compatibility may be more difficult to evolve (Puurtinen et al. 2009). Few studies on the genetic architecture of polyandry exist, but of those we have, pleiotropy appears to be common (Section 9.4.3).

Because of these predictions, good-gene and compatible-gene models for the evolution of polyandry are often treated as mutually exclusive. However, the same alleles may contribute to variation in both additive and non-additive indirect benefits (Puurtinen et al. 2009). As such, females should make their mating decisions taking both into account. Depending on the extent to which intrinsic genetic quality versus genetic compatibility dominate the fitness consequences, we should expect some agreement among females as to the best male partners (perhaps all avoiding avowedly poor or sick males), but there should also be some disagreement among females associated with genetic compatibility. It will be an empirical challenge to tease these effects apart, especially given the sampling error that inevitably arises in behavioural studies and the large sample sizes needed to disentangle additive and non-additive effects in the required quantitative genetics framework. This line of enquiry is still too new for any studies to have accumulated.

9.4 Experimental approaches to understanding polyandry

9.4.1 Measuring the economics of polyandry

Much of the debate over the evolution of polyandry hinges on properly estimating the economics of polyandry. The relative magnitude of direct and indirect costs and benefits is likely not only to be taxon specific but also environmentally (context) specific, and thus could be highly labile (Section 9.5). This means that polyandry may vary across populations of the same species, or that benefits to polyandry may be found in some populations but not others (including populations living in the laboratory), rendering conclusions about the functional significance of polyandry challenging. Additionally, measuring the fitness consequences of multiple mating should include multiple measures of offspring fitness. For example, in C. maculatus, a quantitative genetics approach found variation in the contribution of additive and non-additive genetic variance depending on what fitness-related trait was measured (Bilde et al. 2008). Moreover, a recent study suggests that indirect benefits may be transient across multiple generations; D. melanogaster females exposed to high levels of sexual conflict produced sons with increased fitness but grandsons with decreased fitness assayed as competitive fertilization success (Brommer et al. 2012). These studies indicate that, while it may be easier to pick one reasonable fitness proxy and constrain measurements to the F1 generation, these choices will impact our understanding of the evolutionary significance of polyandry.

Another concern with current studies is that most do not consider whether multiple factors could contribute to the evolution of polyandry. For example, when direct benefits are found, few studies examine whether indirect benefits may also be acting, likely because the gains through indirect benefits generally are thought to be relatively weak. One recent study, however, found that females in the nuptial-gift-giving spider, Pisaura mirabilis, benefited marginally from mating multiply via earlier reproduction—a direct benefit—but also acquired substantial indirect benefits through increased egg-hatching success (Tuni et al. 2013). Such indirect benefits may represent the origin of polyandry for (p.171) this species or could be derived from male manipulation of the female motivation to forage (see Tuni et al. 2013). Likewise, both good-genes and compatible-genes benefits may be occurring simultaneously and their relative magnitudes have rarely been examined in insects (but see Section 9.4.3).

Some powerful techniques have been employed since Thornhill and Alcock that have helped to uncover the functional significance of polyandry. These include experimental protocols to clarify differences in mating number versus polyandry per se, genetic approaches including quantitative genetic models to partition additive and non-additive variance, and genomic techniques to identify genes and gene networks associated with mating.

9.4.2 Controlling for multiple mating

One issue with earlier studies on the evolution of polyandry was that experiments did not control for differences in multiple mating per se and mating with multiple partners (polyandry). The first experimental approach to combat this potentially confounding issue was that of Tregenza and Wedell (1998) who allowed females to mate multiple times with the same male or multiple times with different males. Using the cricket Gryllus bimaculatus, they found that polyandrous females had increased offspring production compared to females who mated multiply with the same male. Subsequent experiments suggested that the benefits of polyandry were associated with inbreeding avoidance (Tregenza and Wedell 2002). This protocol, along with including singly mated female controls (Ivy and Sakaluk 2005), is now widely adopted for discriminating the potential effects of polyandry on female fitness.

9.4.3 Quantitative genetics

Given that distinguishing the effects of intrinsic male quality and genetic compatibility hinge on whether such benefits are additive or non-additive (Puurtinen et al. 2009), quantitative genetics will play a significant role in our understanding of the evolution of polyandry. Yet only a few studies have examined the genetic architecture of the fitness benefits of polyandry in insects.

In the decorated cricket, Gryllodes sigillatus, diallel crosses in which a set of genotypes is crossed in all possible combinations revealed that whether intrinsic male quality or genetic compatibility explained polyandry depended on the fitness trait being measured (Ivy 2007). For development time and female adult offspring mass, non-additive effects strongly influenced phenotypic variance, whereas offspring survival to adulthood was primarily influenced by additive effects (Figure 9.4). A large proportion of phenotypic variance in hatching success was influenced by sire effects (Ivy 2007). Whether these were direct or indirect paternal effects was not determined. In the Australian field cricket, Teleogryllus oceanicus, a quantitative genetics design indicated additive benefits to polyandry arising from sires via increased survival of embryos (García-González and Simmons 2005a). This benefit could be mediated through good genes per se—that is, increased viability by direct transmission of paternal genes—or through indirect paternal effects mediated by ejaculate components, which themselves are heritable. Subsequent work supported the role of ejaculate components promoting embryo survival (García-González and Simmons 2007a).

The evolution of polyandry

Figure 9.4 The relative impact of additive and non-additive genetic effects on phenotypic variation of six fitness-related traits in Gryllodes sigillatus. Six variance components, representing additive (nuclear genes and paternal effects) and non-additive (dominance/epistasis and extranuclear interactions), were estimated using a diallel cross design. From Ivy (2007). Reproduced with permission from Wiley-Blackwell.

Combined additive and non-additive genetic benefits may make the evolution of polyandry more likely in the face of direct costs to female multiple mating. However, a major (p.172) caveat is that net selection for female choice is critical in determining the potential for good genes and genetic compatibility to drive polyandry. As previously mentioned, negative genetic correlations between son and daughter performance may occur (Figure 9.3), limiting indirect benefits even if both additive and non-additive benefits act. Additionally, While quantitative genetic designs are powerful, they cannot show that paternity is biased towards either intrinsically superior genotypes or genetically compatible genotypes. As Simmons (2005) points out, this is because paternity is assigned by the investigator at a relatively late stage of development, such that any earlier effects of genetic incompatibility or intrinsic sire effects on embryo mortality are unknown. Thus, early development paternal or sire-by-dam effects may confound estimates of the genetic effects of polyandry. Maternal effects via differential allocation may similarly confound genetic estimates. Hence, studies that employ rather sophisticated quantitative genetics designs are necessary to delineate additive, non-additive, and maternal effects while controlling for mate number and variation in early embryo mortality. The sample sizes required for such studies are large, and laboratory-friendly species are needed, potentially biasing the insects we choose for such studies.

9.4.4 Experimental evolution

Early artificial selection experiments contributed to our understanding of the functional significance of polyandry by showing that male and female remating rate were not genetically correlated (see Section 9.3.2). A relatively recent technique is experimental evolution in which the mating system, rather than a target trait, is manipulated over an evolutionary time frame to uncover (for instance) the effects of removing and/or elevating the strength of sexual selection and sexual conflict (Kawecki et al. 2012). In insects, such (p.173) experimental sexual selection has shown that a variety of pre- and post-copulatory traits in males evolves under different mating systems (for example see Chapter 10). With relevance to polyandry, two of these experimental evolution approaches have revealed the adaptive significance of polyandry.

Selfish genetic elements in the form of meiotic drivers may result in sperm limitation since 50% of sperm—in sex-ratio distorters those bearing either the X or the Y chromozomes—are rendered sterile as a consequence of the driving element. Many natural populations harbour such drivers, and Chapter 13 discusses how in D. pseudoobscura, after just 10 generations of evolution, females in female-biased populations harbouring the X driver had evolved greater remating rates and a greater likelihood of remating on the first opportunity (i.e. decreased female choosiness) compared with populations lacking the X driver (Price et al. 2008b). Thus, selfish genetic elements may promote polyandry by allowing females to mitigate against mating only with males with depleted sperm and/or with sperm carrying a sex-ratio distorter. Interestingly, monogamous populations carrying the sex-ratio distortion gene were also more likely to go extinct, suggesting that polyandry can protect against population extinct risk (Price et al. 2010a).

Experimental evolution has also identified the role of inbreeding avoidance in generating polyandry. If polyandry is beneficial to avoid genetic incompatibilities arising from inbreeding, then such behaviour should be found in spatio-temporal conditions subject to high inbreeding risk, such as species in which females have to leave crowded resource patches, colonize an empty patch, and then whose offspring will be restricted to siblings as potential mates until further colonization. Theoretical modelling examining whether such conditions could explain polyandry via indirect benefits indicates that this depends on whether inbreeding depression occurs through deleterious recessive alleles or overdominance (Cornell and Tregenza 2007). While the latter can maintain inbreeding depression, thus fostering continual potential indirect genetic benefits for polyandry, deleterious recessive alleles would be purged from the system and benefits from polyandry would be small and of transient importance.

An experimental evolution study in which populations of the red flour beetle, Tribolium castaneum, a species that meets the spatio-temporal conditions of the Cornell and Tregenza (2007) model, has been used to show that inbreeding can promote the evolution of polyandry (Michalczyk et al. 2011). In this experiment, T. castaneum populations were subjected to eight generations of genetic bottlenecking via sib–sib matings and the fitness consequences for inbred and outbred populations were compared under evolutionary scenarios of either monogamy or polyandry. Monogamous inbred females had reduced composite fitness (offspring number and offspring survival) compared with non-inbred controls, but polyandrous inbred females regained this loss; these results indicate that inbreeding depression occurs and that polyandry may combat inbreeding depression (Figure 9.5a). Because polyandry benefitted inbred populations, female remating behaviour was compared with non-inbred control lines 15 generations after genetic bottlenecking. Previously inbred females had higher levels of polyandry than non-inbred controls (Figure 9.5b, c; Michalczyk et al. 2011). These results suggest that past inbreeding may have current effects on mating system structure. However, whether inbreeding depression in T. castaneum is mediated via deleterious recessive alleles or overdominance is currently unknown.

The evolution of polyandry

Figure 9.5 Female fitness and mating behaviour response to experimental inbreeding in T. castaneum. Inbred females regain fitness to control, non-inbred, levels if they mate polyandrously compared to a single mating (upper panel). Females from the previously inbred lines mate quicker (middle panel) and have more matings (lower panel) than control, non-inbred females. Modified from Michalczyk et al. (2011). Reproduced with permission from The American Association for the Advancement of Science.

Thus, experimental evolution has been used to great effect to test aspects of sexual selection and sexual conflict in general and polyandry in particular. This technique is particularly well suited to insects as they tend to have relatively large effective population sizes, short generation times, are relatively easy to rear in the laboratory, and thus relatively easy (p.174) to replicate. This approach will remain informative, especially in combination with quantitative genetics and genomics techniques (see Section 9.4.5). However, such laboratory studies must always be tempered with the realization that they serve as hypotheses to test in wild populations (see Section 9.3.1).

9.4.5 Genomics

One considerable difference between the current emphasis on understanding polyandry compared to when Thornhill and Alcock’s book was published are the genetic techniques that have not only demonstrated that polyandry is pervasive (Section 9.2), but also now allow the study of the genes themselves that are associated with polyandry (see also Chapter 4). Microarray and next-generation sequencing permit the identification of genes that control polyandry. While the use of these techniques to study polyandry is still in its infancy, such studies represent a fast-growing area for understanding evolutionary consequences of polyandry at the genetic level.

Drosophila melanogaster has served as a model laboratory system in which to examine how specific genes influence the outcome of mating (see Chapter 4). Perhaps the best understood of these genes is Sex Peptide (SP), a seminal fluid protein that alters female behaviour and physiology and mediates sexual conflict. Relevant for studies of polyandry, SP reduces female receptivity and decreases female survival. Using microarrays, one study has focused on how the sole transfer of SP influenced female mRNA expression (Gioti et al. 2012). A (p.175) large number of genes altered expression in response to the receipt of SP, indicating that it has extensive pleiotropic effects and that SP may be a ‘global regulator’ of female reproductive processes. As a consequence of large-scale pleiotropy, females may be constrained in evolving any response that mediates the toxic effect of receiving SP upon mating, rendering the mitigation of the negative fitness consequences of multiple mating difficult.

Another study of D. melanogaster used artificial selection to alter mating speed, also referred to as copulation latency, defined as the time it takes for males and females once together to begin copulation. Artificial selection was used to effectively increase (‘Fast’) or reduce (‘Slow’) female receptivity (Mackay et al. 2005), essentially serving as a proxy for populations to be either more polygamous or less polygamous respectively. After 29 generations of selection, microarrays were used to assess whole genome transcriptional response to changes in the mating system. Twenty-one per cent of the transcriptome changed as a consequence of selection. Of the genes that were differentially expressed between selection lines, approximately 38% were altered uniquely (i.e. up- or downregulated in only one line) with 68% of these upregulated in Fast females. A variety of different types of genes was altered in response to selection for mating system, suggesting extensive pleiotropy. Intriguingly, some genes with altered expression are involved in the basal sex determination pathway (doublesex, transformer, transformer 2, and fruitless), with Fast females exhibiting upregulation of female sex determination genes and Slow females exhibiting upregulation of male sex determination genes (Mackay et al. 2005). Perhaps the mating system either influences or is influenced by genes associated with sex determination and downstream effects on sex-specific patterning of neural architecture, and potentially impacting the extent of maleness and femaleness.

Genomics studies on the evolution of mating systems are still in their infancy (Chapter 4). Two studies in D. melanogaster suggest pervasive and pleiotropic effects on the genome in response to sexual conflict and changes in female mating behaviour (Mackay et al. 2005; Gioti et al. 2012). The extent to which this constrains or facilitates responses to sexual conflict, and affects direct and indirect benefits to polyandry, has not yet been examined. Future work might capitalize on the decreasing costs and increasing ease of genomic interrogation in non-model organisms. When combined with experimental evolution, such studies have the potential to substantially increase our understanding of the evolution and maintenance of polyandry. Although powerful, these studies can only suggest candidate genes that may be important in mediating functional responses to polyandry. Subsequent work is necessary to determine the evolutionary relevance of such changes regarding the economics of polyandry.

9.5 The ecology of polyandry

9.5.1 Mating system plasticity and ecology

An explosion of data on the evolutionary significance of polyandry has accumulated since Thornhill and Alcock. There is now a large and ever-expanding body of literature on the consequences of polyandry for specific species; somewhat of a directory of costs and benefits for each species. Likewise, as the chapters in this volume attest, there is a growing appreciation for the extensive effects of polyandry on other evolutionary processes. Just within the past few years, researchers have found new evolutionary factors that promote polyandry, such as selfish genetic elements, population extinction, and inbreeding.

Yet, despite the tremendous advance, explanations for the underlying interspecific patterns that generate variation in remating and the economics of polyandry, particularly (p.176) when material benefits are absent, are lacking. Individual studies provide the building blocks for supporting or rejecting particular hypotheses regarding the evolution of polyandry that can then be tested generally using meta-analyses, which have found substantial and uncontroversial evidence for material benefits but little evidence for genetic benefits, despite some taxa clearly benefitting indirectly from polyandry (Arnqvist and Nilsson 2000; Slatyer et al. 2012). How can we reconcile the individual cases with the overall pattern? What are the abiotic and biotic conditions, if any, that generally promote genetic benefits arising from polyandry? What ecological conditions limit such benefits?

For more than 30 years, behavioural ecologists have recognized that mating systems are influenced by ecology; indeed mating system theory is predicated on ecology (e.g. Emlen and Oring 1977). That is, mating systems are context dependent. For example, in the burying beetle genus Nicrophorus, individuals may experience monogamy, polyandry, polygyny, or polygynandry depending on the number of males and females that show up at the carcass, which is at least partially determined by carcass size (Muller et al. 2006). In bushcrickets, in which males produce expensive spermatophores that directly benefit females, the availability of resources in the environment determines the limiting sex (Gwynne and Simmons 1990; for review, see Lehmann 2012). It is therefore odd that, despite this strong history of recognizing the role of ecology in mating systems, this appreciation has not informed our understanding of the economics of polyandry to the extent that it perhaps should have done. If we assume that polyandry is a variable life-history strategy, then studying parameters that influence such life-history decisions may provide a more general understanding of what types of benefits females may obtain from mating and whether these benefits outweigh the costs of multiple mating.

The burying beetle and bushcricket examples are clear cases in which females benefit directly, and changes in ecology have clear implications for the extent to which females gain from multiple mating. Not all examples, even for direct benefits, are so clear cut. For example, in the butterfly, Pieris napi, females should be polyandrous as males transfer a large, nutritious, ejaculate (Figure 9.6a) which increases female fitness with little cost, as polyandrous females live longer than monandrous females (Wiklund et al. 1993). Yet, while polyandry is under genetic control (Wedell et al. 2002), populations vary in the proportion of females that are polyandrous, with more northern populations being more monandrous than southern populations (Figure 9.6b; Välimäki and Kaitala 2010). The association between latitude and female mating pattern suggests that the costs and benefits of multiple mating vary based on environmental and associated life-history parameters that are linked to latitude. In particular, northern populations will experience a truncated summer and lower average temperatures throughout the year, limiting reproduction and offspring development. Hence, populations may vary in the number of generations per year which could change the relative costs and benefits of polyandry within the same species. Intriguingly, whereas northern females have lower overall fecundity under standard conditions, they have greater early fecundity because they avoid the timing costs of additional matings (Välimäki et al. 2008), suggesting that in a shortened reproductive season northern females would not benefit from polyandry despite receiving direct benefits from multiple males. In the southern multivoltine populations, females from directly developing generations exhibit higher levels of polyandry than females from the pupal diapausing generation, suggesting that something about the economics of polyandry differs even in a species with material benefits. What this difference may be is currently unknown (Larsdotter Mellström and Wiklund 2010).

The evolution of polyandry

Figure 9.6 Variance in mating frequency between and within populations of Pieris napi. (a) Copulating P. napi. Reproduced with permission from P. Välimäki. (b) Females from southern populations are more promiscuous than females from northern populations. Modified from Välimäki and Kaitala (2010). Reproduced with permission from Elsevier. (c) Females from the directly developing generation are more promiscuous than diapausing females. Numbers on the open bars are lifetime number of matings. From Larsdotter Mellström and Wiklund (2010). Reproduced with permission from Elsevier.

If different ecological conditions and life histories select for different levels of polyandry within a species, this may have downstream effects on other aspects of the mating system. (p.177) For example, in P. napi, since females from the directly developing generation exhibit higher levels of polyandry, perhaps males may be sperm-depleted, which may impact on other aspects of sexual selection and the costs and benefits to polyandry (for review, see Kvarnemo and Simmons 2013). Thus, determining the causes of interspecific variation in the functional significance of polyandry, taking into account abiotic and biotic processes within a species that facilitate and limit benefits to female multiple mating, is challenging. It will require integrating studies of intraspecific plasticity of mating systems driven by identified ecological variables, with comparative studies across species incorporating those variables.

9.5.2 Quantitative genetics and mating system plasticity

In a quantitative genetics framework, the context dependency of mating systems is reflected by genotype × environment interactions, that is, the relative phenotypic performance of different genotypes is influenced by the environment in which genotypes are expressed. In quantitative genetic studies of indirect benefits to polyandry, sources of phenotypic variance may be significantly affected by the environment (Bilde et al. 2008; but see Watson and Simmons 2011). The effect of environmental heterogeneity on the economics of polyandry can arise from abiotic factors (e.g. latitude, P. napi; see Section 9.4.1) or from biotic factors (indirect genetic effects; Wolf et al. 1998) such as those arising between the interacting male(s) and female. A study of D. melanogaster which altered social heterogeneity by changing group composition found that female reproductive behaviour changed (p.178) in a single generation with an increase in offspring diversity through alterations in either female preference or remating frequency (Billeter et al. 2012). While the study did not assess female costs to altering mating behaviour, or directly quantify these in a quantitative genetics framework, the results clearly argue for such studies. A caveat is that a quantitative genetics framework will be limited to species that lend themselves to such studies (see Section 9.4.3). Nevertheless, such integration will facilitate an understanding of the association between relevant abiotic and biotic parameters and how changes in these might influence the costs and benefits (both direct and indirect) of polyandry.

9.5.3 Reproductive mode and indirect benefits

Other researchers have proposed that focusing on life-history trait differences between species may help elucidate the evolutionary significance of polyandry (e.g. Hosken and Stockley 2003, Zeh and Zeh 2001a). For example, Zeh and Zeh (2001a) proposed that reproductive mode can influence the adaptive significance of polyandry. Reproductive mode is a catch-all phrase, used to describe asexual versus sexual reproduction, whether reproduction is terrestrial or aquatic, whether fertilization is internal or external, or—as considered by the Zehs—whether females are viviparous or ovi/ovoviviparous (give birth to live young or lay eggs). The hypothesis that reproductive mode influences the evolution of polyandry arises from a distinction between whether genetic benefits gained from polyandry are additive or non-additive, and the effect of physiology, immunology, and evolutionary conflict on maternal/zygote interactions during embryonic development. As argued by Zeh and Zeh (2001a), if species are viviparous, then maternal–zygote interactions during development are critical for female reproductive success and any evolutionary genetic conflict with the paternal genetic contribution may have a detrimental fitness effect. In this case, polyandry should be associated with benefits arising from genetic compatibility. In oviparous taxa, because the zygote is not reliant on continuous input of maternal nutrient contributions during development, polyandry should perhaps be associated with benefits arising from intrinsic male quality.

Even though all insects produce eggs and there is no placenta, some insects are nonetheless pseudoplacental, giving birth to live offspring and providing maternal nutrients during development (Chapter 12). Conflict over such investment may occur between the female and her mate(s), the female and developing embryos, and between developing siblings which may lead to spontaneous abortions (Zeh and Zeh 2001b). Whether such conflict is manifested in insects has not been determined. Viviparity in dipterans may have evolved independently more than 60 times (Meier et al. 1999). Given the prediction about reproductive mode and benefits to polyandry, comparative analyses of oviparous/ovoviviparous with viviparous dipterans would predict that the former should exhibit intrinsic benefits and the later non-additive benefits, all other things being equal. Such studies have not been done, likely because the benefits of polyandry have not been well studied in viviparous systems.

Whereas reproductive mode is one specific life-history trait, the general point is that very little progress has been made moving from the catalogue of intraspecific studies to understanding the broader interspecific patterns. As argued in Sections 9.5.1 and 9.5.2, such integration is important in helping to understanding polyandry and its fundamental consequences for evolutionary change.

9.4.5 The potential applied relevance of polyandry

Female multiple mating behaviour may be influenced by anthropogenic environmental modifications, such as climate change, habitat fragmentation, pollution and selective (p.179) harvesting, which changes the species’ ecology and will affect the underlying genetic variation for sexually selected traits (Lane et al. 2011). For example, one of the first biological responses to climate change is an alteration in the temporal distribution of life-history traits (Parmesan 2006). Since variation in the life histories of males and females determines the relative costs and benefits of multiple mating, any phenological response to climate change also has the potential to change the frequency and functional significance of polyandry. Likewise, if habitat fragmentation results in the loss of lekking or nesting sites, then smaller areas will support fewer males and females. How is the mating system affected? Sexual selection could be intensified on males to compete for fewer females that are patchily distributed, or the system may move to a more monogamous one. If females still mate multiply, then whether they do so to gain primarily additive (good- or sexy-sons benefits) or non-additive benefits could depend on the extent to which genetic factors are altered as a consequence of habitat fragmentation. Does habitat fragmentation increase the chances of inbreeding and subsequent inbreeding depression? If so, then females may mate multiply to avoid costs of inbreeding, assuming that inbreeding depression is mediated through overdominance. Selective harvesting can eliminate dominant males from the population and have a variety of genetic consequences on the population but can also change life-history traits and demographics (for review, see Coltman 2008), which may influence the mating system. Understanding whether anthropogenic influences on the environment alters the mating system and, if so, what the likely consequences are for the evolutionary trajectories of affected populations has only recently been considered as an important research paradigm in behavioural ecology, and thus few studies have been published. However, given the effect of ecology on mating systems and the rapidity of anthropogenic changes in ecology, our relative ignorance of the long-term consequences of such changes should make such studies highly relevant.

9.6 Conclusion

In the 30 years since Thornhill and Alcock’s book, our understanding of the putative drivers of polyandry has increased substantially. The study of indirect benefits, in its infancy at that time, has now moved into a quantitative genetics framework. Perhaps the largest change in emphasis has been on the direct costs of polyandry as a consequence of sexual conflict. Although individual empirical studies have demonstrated indirect benefits to female multiple mating, meta-analyses have shown overall weak support for indirect benefits. The use of experimental evolution has documented the effect of polyandry on a variety of putatively sexually selected traits, has demonstrated the action of sexual conflict, and has identified the benefits of polyandry in different evolutionary contexts. Further development of genomics techniques has identified pervasive and pleiotropic genetic effects of polyandry. And the consequences of polyandry on a variety of different evolutionary phenomena has been discovered, many of which are discussed throughout this volume.

Yet much work remains for the new student of polyandry. A continuing key challenge is to determine the functional significance of polyandry, particularly in species with no direct benefits. Many researchers suggest, and find, that direct mating costs, mediated primarily through sexual conflict, are too large relative to the predicted weak genetic benefits. However, this is not always true; recent work has found evidence for indirect benefits that can even be larger than both direct benefits and direct costs. Assuming that genetic models of sexual selection hold true, what is the relative contribution of additive and non-additive effects? Quantitative genetics studies can help pinpoint this, but such efforts (p.180) must include multiple fitness-related traits and integrate between laboratory and natural populations. Integration is complicated by the fact that mating systems are still typically studied as a static phenomenon, despite the recognition that the economy of polyandry is context dependent, influenced by both abiotic and biotic interactions, and potentially across more than one generation. Environmental influences may change the economics of polyandry but empirical studies, particularly in wild populations, remain scarce. To understand the origin and maintenance of polyandry, researchers should move beyond examining individual static case studies to investigating how spatiotemporal variability in ecology and life history within a species changes the costs and benefits of female multiple mating. Such studies should be done in controlled laboratory settings but also attempted in the wild. We currently have a catalogue of intraspecific studies that have helped tremendously in our understanding of polyandry, but we remain largely ignorant of how we can scale up our understanding of what drives adaptive significance of polyandry between species—that is, how we might predict the cost/benefit ratio for any given system. The increasing affordability of genomic data can be used to understand how polyandry influences and is influenced by genetic architecture, and what the evolutionary consequences of such interactions may be. Rapid ecological change and the context-dependent nature of the economics of polyandry strongly argue for studies that link these two to foster an understanding of how anthropological effects may influence mating systems.

Over the past 30 years, studies of the impact of polyandry have demonstrated its evolutionary power, such that polyandry recently has been argued to be the most fundamental agent of evolutionary change (Kvarnemo and Simmons 2013). Given the realization that net selection operating on males and females is more nuanced than previously appreciated, the study of polyandry remains at least as vibrant and as important as it was 30 years ago. There is no indication that we are at the limit of our understanding and new students of polyandry have much to offer.