By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Leigh W. Simmons, Centre for Evolutionary Biology, School of Animal Biology (M092), The University of Western Australia, Crawley 6009, Australia. Tel.: +61 8 6488 2221; fax: +61 8 6488 1029; e-mail: firstname.lastname@example.org
Life history theory provides a powerful tool to study an organism's biology within an evolutionary framework. The notion that males face a longevity cost of competing for and displaying to females lies at the core of sexual selection theory. Likewise, recent game theory models of the evolution of ejaculation strategies assume that males face a trade-off between expenditure on the ejaculate and expenditure on gaining additional matings. Males of the dung beetle Onthophagus binodis adopt alternative reproductive tactics in which major males fight for and help provision females, and minor males sneak copulations with females that are guarded by major males. Minor males are always subject to sperm competition, and consistent with theoretical expectation, minor males have a greater expenditure on their ejaculate than major males. We used this model system to seek evidence that mating comes at a cost for future fertility and/or male expenditure on courtship and attractiveness, and to establish whether these traits vary between alternative mating tactics. We monitored the lifespan of males exposed to females and nonmating populations, and sampled males throughout their lives to assess their fertility and courtship behaviour. We found a significant longevity cost of reproduction, but no fertility cost. On average, males from mating populations had a lower courtship rate than those from nonmating populations. This small effect, although statistically nonsignificant, was associated with significant increases in the time males required to achieve mating. Minor males had lower courtship rates than major males, and took longer to achieve mating. Although we did not measure ejaculate expenditure in this study, the correlation between lower courtship rate and longer mating speed of minor males documented here with their greater expenditure on the ejaculate found in previous studies, is consistent with game theory models of ejaculate expenditure which assume that males trade expenditure on gaining matings for expenditure on gaining fertilizations.
Sperm competition, the competition between sperm from two or more males for the fertilization of ova (Parker, 1970, 1998), is now widely accepted as a potent force of evolutionary change, favouring adaptations in male behaviour, morphology and physiology that contribute to competitive fertilization success (Smith, 1984; Birkhead & Parker, 1997; Simmons, 2001). Game theory has provided a theoretical framework within which to assess the evolutionary consequences of sperm competition for male expenditure on the ejaculate (Parker, 1998). At the core of sperm competition game theory lies the assumption that males face a life history trade-off between expenditure on the ejaculate and expenditure on gaining matings. Although there are a wealth of studies to suggest that males exercise prudence in ejaculate expenditure (Wedell et al., 2002), implying that ejaculate production is costly, sperm production has rarely been studied within a life history context. Mating is known to reduce male lifespan in a number of species (Kotiaho & Simons, 2003; Fedorka et al., 2004; Martin & Hosken, 2004; Simmons & Kvarnemo, 2006), but whether this cost is associated with sperm production remains unclear. Research on the nematode Caenorhabditis elegans concluded that sperm production decreased lifespan in males, based on the observation that mutant spermless males had longer lifespans than normal males (Van Voorhies, 1992). It is now clear that it is the presence of an active germline, rather than the somatic gonad, that incurs a longevity cost of reproduction for males in this species (Hsin & Kenyon, 1999; Barnes & Partridge, 2003). Although recent work with Drosophila melanogaster found no, or only a slight extension in longevity in males with germline ablation (Barnes et al., 2006), previous work suggests that fertility costs of reproduction for male Drosophila may be of greater evolutionary significance than longevity costs. Thus Prowse & Partridge (1997) found that mating male D. melanogaster incurred irreversible reductions in sperm production and fertility at a time when most of their nonmating cohort were still alive.
Here we use the dung beetle Onthophagus binodis as an empirical model to examine the costs of reproduction for males. Typical of many onthophagines (Lee & Peng, 1981; Sowig, 1996; Emlen, 1997; Hunt & Simmons, 2000; Moczek & Emlen, 2000; Hunt & Simmons, 2002), O. binodis exhibit a suite of morphological and behavioural traits that characterize alternative male reproductive tactics. Major male O. binodis develop an enlarged pronotal horn that is used in competition for access to female breeding chambers where they assist in the production of brood masses. In contrast, minor males develop only rudimentary horns and sneak copulations with females guarded by major males (Cook, 1987, 1988, 1990). By the nature of their alternative tactic, minor males are always subject to sperm competition from major males, while major males will be subject to a lower risk of sperm competition, dependent on the probability of a sneak mating occurring. Sperm competition game theory predicts that this asymmetry in sperm competition risk should favour differences in male expenditure on the ejaculate; major males should expend less on their ejaculate than minor males if they are subject to a consistently lower risk of sperm competition (Parker, 1990a,b). Consistent with theory, major male O. binodis have smaller testes and transfer smaller ejaculates at copulation than minor males (Simmons et al., 1999). These alternative reproductive tactics make the species ideal for examining empirically the life history trade-off assumed by sperm competition game models, that males trade investment in gaining fertilizations for investment in obtaining additional mates (Parker, 1998). Previously, we found that although males incurred a general longevity cost of mating, elevated expenditure on the ejaculate by minor male O. binodis was not associated with a relatively shorter lifespan (Kotiaho & Simons, 2003). The principle aim of this study was to look for evidence that mating comes at a cost for future fertility and/or male expenditure on courtship and attractiveness, and to establish whether these traits vary between alternative mating tactics.
Animals used in this study were F1 offspring derived from parents collected from Walpole, Western Australia. Beetles were collected from dairy pastures and maintained in large mixed sex cultures. Each week for a period of 2 weeks, 200 females were isolated and placed into individual breeding chambers (PVC piping, 30 cm in length and 9 cm in diameter, three quarters filled with moist sand, and topped with 250 mL of cow dung). Breeding chambers were sieved after 10 days and batches of ca 50 brood masses were buried in moist sand in 10-L plastic containers. Broods were incubated under constant conditions of 28 ± 2 °C and a 12 : 12 h light : dark photoperiod. Containers were checked daily for emerging beetles. Newly emerged beetles were measured (pronotum width for both sexes and horn length of males) and housed individually in 50-mL plastic containers provided with moist sand and fresh dung.
Within 10 days of their emergence, F1 beetles were used to establish six populations, three mating populations and three nonmating populations. Mating populations consisted of 50 females, 25 major males and 25 minor males. Male morph was distinguished from the scaling relationship between horn size and body size following the method of Kotiaho & Tomkins (2001); we selected minor males with horns < 0.5 mm, and major males with horns > 1.0 mm in length (mean ± SE horn lengths: minors 0.39 ± 0.02 mm; majors 1.24 ± 0.02 mm). Each population was housed in a 15-L bucket filled to 10 L with moist sand and topped with 2 L of cow dung. Nonmating populations consisted of 25 major males and 25 minor males. Equal population density between treatments was achieved by housing nonmating populations in identical buckets, filled to 5 L with moist sand and topped with 1 L of cow dung. Thus, the density of both mating and nonmating populations was 10 beetles/L sand/L dung. We did not provide enough dung to allow beetles to produce broods, because we wanted to exclude any costs of reproduction that might be associated with paternal provisioning. Every 10 days, sand and dung were changed and the survival of beetles recorded.
Male courtship success and fertility
We assessed male courtship success and fertility 10 days following the establishment of populations (early life), at 40 days (mid life), and at 60 days (late life). We selected five major and five minor males from each of the mating and nonmating populations. Males were placed into observation chambers (1.3 cm × 3.6 cm × 6 cm) that had a 1-cm thick plaster-of-Paris base, moistened with water and fresh dung. A virgin female was introduced to each observation chamber and the pair observed under red light until they had copulated. Males court females by climbing onto their backs and tapping the female with their head and forelegs in bouts lasting several seconds. Courtship rate was determined by scan sampling at a rate of six scans per minute. Thus, on each scan the male was noted as courting or not. Male courtship rate was calculated as the frequency of courtship observations divided by the total time from first courtship to mating (Kotiaho et al., 2001; Kotiaho, 2002). We also recorded mating speed, as the time from introduction of the female until the pair copulated (Manning, 1961; Parsons, 1974). Mating speed provides a measure of male attractiveness (Moore & Moore, 1999; Shackleton et al., 2005), and has been shown to be a good predictor of lifetime mating success (Simmons, 1988; Shackleton et al., 2005). Thus, males with high mating speeds take a long time to achieve copulation, are less attractive to females, and have a lower lifetime mating success.
After mating, males were removed from observation chambers and marked by making a small scratch on the pronotum with a sharpened dissection pin. Thus marked, we were able to avoid using them again on subsequent screening days. Males were returned to their appropriate populations. Once females had mated they were removed from observation chambers and placed into individual breeding chambers to construct broods. A single egg is laid into each brood mass which constitutes all of the resources available for growth and development from hatching to adulthood. Each week following mating, breeding chambers were sieved, any broods buried in moist sand in 1-L plastic containers, and the females re-established with fresh sand and dung. Females were thus monitored until death. On the week following sieving, broods were opened to determine if they contained a hatched larvae. Broods were thus scored as fertile or infertile.
We looked for a longevity cost of reproduction for males by comparing patterns of survival across the six populations. Parametric survival analysis was conducted using a log-normal distribution and a model that included treatment (mating/nonmating), morph (major/minor), and their interaction. We also looked for differences between populations nested within treatments. The interaction term between treatment and morph was not significant (P = 0.469) and was removed from the final model which explained a significant proportion of the variance ( =68.24, P < 0.001). Populations varied within treatments ( = 40.55, P < 0.001) but after accounting for this variation, there was a significant effect of treatment ( = 28.74, P < 0.001) but no significant effect of morph ( = 2.73, P = 0.10). Thus, exposure to females significantly reduced male lifespan (Fig. 1).
Courtship and mating
Courtship rate and mating speed data were log transformed to achieve normality. We analysed variation in these variables using a General Linear Modelling (GLM) approach. Our full model included treatment, morph, life stage (10, 40 or 60 days) and all possible interaction terms. We also accounted for variation between populations nested within treatments.
For male courtship rate, neither the population within treatment (P = 0.171), or any of the interaction terms were significant (P > 0.560) in the full model. Neither were there significant main effects of treatment (F1,153 =0.68, P = 0.409) or life stage (F2,153 = 0.71, P = 0.493) (Fig. 2). Stepwise deletion of insignificant terms yielded a reduced model that explained only 2% of the variance in courtship rate, with a trend for major males to have a higher courtship rate than minor males (F1,167 = 3.68, P = 0.057) (untransformed means: majors, 3.0 ±0.9 bouts min−1; minors 0.6 ± 0.9 bouts min−1). Given the borderline statistical significance of the effect of male morph on courtship rate, the effect size and its 95% confidence intervals offers a more robust assessment of biological significance (Colegrave & Ruxton, 2003). The effect size, Cohen's d (95% CI), was 0.300 (0.002, 0.596). Thus we can reject an effect smaller than 0.002 and larger than 0.596 at P = 0.025. We also note that on average males from mating populations courted at half the rate (1.44 ± 0.92 bouts min−1) of males from nonmating populations (2.33 ± 0.92 bouts min−1). The effect size for treatment was 0.125 (−0.171, 0.421). The 95% CI for this effect size suggest that a small positive biological effect is more likely than no effect (Johnson, 1999), and that our sample size was insufficient to demonstrate a statistical difference between treatments. Inspection of Fig. 2 also suggests that any effect is probably most attributable to a greater courtship rate of nonmating major males.
In our analysis of mating speed, the population within treatment term was not significant (P = 0.10), and neither were any of the interaction terms (P > 0.30). Stepwise deletion of insignificant terms yielded a reduced model that explained 20% of the variance in mating speed (F4,167 = 10.57, P < 0.001) (Fig. 3). Consistent with the patterns observed in male courtship rate, males from mating populations took longer to achieve copulation than males from nonmating populations (F1,167 =4.08, P < 0.05) (untransformed mean ± SE: mating, 28.3 ± 2.1 min; nonmating, 22.9 ± 2.2 min), and major males achieved copulation sooner than minor males (F1,167 = 7.69, P < 0.01) (untransformed mean values: majors, 21.8 ± 2.1; minors, 29.7 ± 2.2). Males also took longer to achieve copulation when young (F2,167 =14.78, P < 0.001) (untransformed mean values: 10 days, 38.6 ± 2.6 min; 40 days, 18.7 ± 2.7 min; 60 days, 19.4 ± 2.6 min). Courtship rate was negatively associated with mating speed, both when data were pooled across all individuals tested (r2 = 0.132, F1,167 = 25.41, P < 0.001; effect estimate −0.36 ± 0.06), and when controlling for effects of treatment, morph, and life stage (F1,163 = 25.40, P < 0.001; effect estimate −0.29 ± 0.06).
We used the same GLM approach in our analysis of fertility. We first examined the total number of broods each female produced during her lifetime in relation to whether her mate came from the mating or nonmating populations, the morph of her mate, and the life stage of her mate when they copulated. We also looked for variation between male populations nested within treatments. None of these terms or their interactions explained any of the variation in the numbers of broods produced (full model, F15,160 = 1.07, P = 0.384). Females produced on average 18.4 ± 1.1 (range 0–63) broods during their lifespans.
Fertility was assessed as the proportion of broods in which the egg hatched (proportions were arcsine square root transformed for statistical analysis). In our full model there was no significant variation among populations nested within treatments (P = 0.379). With the exception of the interaction between treatment and life stage, no other interactions were significant (P > 0.30). Neither morph (P = 0.872) or life stage (P = 0.376) were significant. The reduced model explained 9% of the variance in fertility (F5,159 = 3.28, P < 0.01). There was a significant main effect of treatment (F1,159 = 4.14, P < 0.05), no significant effect of life stage (F2,159 =0.99, P = 0.372), and a significant treatment by life stage interaction (F2,159 = 4.92, P < 0.01). At mid life (40 days of age) males from the mating populations had a higher proportion of fertile broods than those from the nonmating populations (Fig. 4).
We found a significant longevity cost of exposure to females for male O. binodis. In a previous study of this species we also found that males housed with females had a reduced lifespan compared with males from nonmating populations (Kotiaho & Simons, 2003). However, in that study population density was allowed to vary across mating and nonmating populations, and although population density had been shown elsewhere not to influence lifespan (Ridsdill-Smith et al., 1982), we could not be certain that reduced lifespan of males in our mating treatment resulted from a cost of reproduction alone. Here we controlled for population density across treatments and found the same reduction in lifespan for mating males compared with males from nonmating populations. This longevity cost of reproduction may arise from costs associated with courtship, contest competition, mating, or sperm production. As in our previous study, we failed to find differences between alternative reproductive tactics in the longevity costs of reproduction. Given that minor males have a greater expenditure on ejaculate production than major males (Simmons et al., 1999), the primary goal of the current study was to determine whether minor males trade their greater ejaculate expenditure for some other aspect of their life history, such as expenditure on gaining additional reproductive opportunities.
We found that mating males were less attractive to females, taking longer to achieve mating compared with males from nonmating populations. These data suggest that increased mating speed represents a cost of reproduction for males. Previous work with three species of onthophagines (O. binodis, O. taurus and O. australis) has found that courtship rate is a condition-dependent male trait. Limiting the availability of nutrient resources has the effect of reducing a male's courtship rate indicating that it is an energetically costly activity (Kotiaho, 2002), and in O. taurus courtship rate is also a heritable trait that is genetically correlated with condition (Kotiaho et al., 2001). Moreover, females exhibit linear preference functions for courtship rate, such that males with low courtship rates are rejected as mating partners (Kotiaho, 2002). The patterns of mating speed variation found in our study could be explained by males suffering a reduction in energetically expensive courtship rate as a consequence of reproduction, so that they become less attractive to females.
There was no statistically significant effect of reproduction on courtship rate in our data. However, males from nonmating populations courted at twice the rate of males from mating populations, and the confidence intervals of the observed effect size caution against accepting the null hypothesis of no effect. Rather, the confidence intervals suggest that there may be a small effect of biological relevance (Johnson, 1999). There are undoubtedly aspects of male courtship, such as the number of leg and/or head taps performed in a bout of courtship or the force of leg/head tapping, that we are unable to capture with our crude metric of bouts per minute. A more accurate metric of courtship may have yielded a larger effect size. Nevertheless, courtship rate explained 13% of the variation in mating speed (males with high courtship rates mated sooner) so that the small biological effect of reproduction on courtship rate was associated with a statistically significant effect of reproduction on male mating speed.
Other unmeasured aspects of male courtship might also contribute to mating speed. For example, exocrine glands in the forelegs of onthophagines are thought to be involved in pheromone production and mate attraction (Houston, 1986). Pheromone production has been shown to represent a significant cost of reproduction in Drosophila (Johansson et al., 2005), and reductions in the production of pheromones may be an additional cost of reproduction not examined in our study, that could account for the increased mating speed of males from mating populations relative to males from nonmating populations. We also found that young males took longer to mate, although they did not differ in their courtship rates. This finding could be explained by a female preference based on male age (Kokko & Lindström, 1996; Kokko, 1998; Jones et al., 2000).
We found consistent effects of male morph on both courtship rate and mating speed suggesting that reduced courtship rate and attractiveness is an intrinsic cost associated with the minor male reproductive tactic. Previous studies of O. binodis have also reported significantly lower courtship rates in minor males (Cook, 1990; Kotiaho, 2002), and it is possible that minor males trade energetically expensive courtship for their greater ejaculate expenditure (Simmons et al., 1999). If true, we would expect differences in courtship rates between morphs to be confined to species in which morphs differ in their ejaculate expenditure. The limited evidence suggests this may be the case; major and minor male O. taurus do not differ significantly in their expenditure on the ejaculate (Simmons et al., 1999), and neither do they differ in their courtship rates (Kotiaho, 2002).
Males have been shown to become sperm depleted with successive copulations in a number of taxa (Smith et al., 1990; Danielsson, 2000; Preston et al., 2001), and life history studies of Drosophila suggest that reduced fertility is a more important cost of reproduction for males than reductions in lifespan (Prowse & Partridge, 1997). Our data suggest that reduced fertility is not a significant cost of reproduction for male O. binodis. Fertility did not decline through a male's life span, and mating males did not have lower fertility than males from nonmating populations. Male fertility was influenced by an interaction between treatment and life stage such that at mid life, mating males had higher fertility than males from nonmating populations. Changes in male fertility through life have also been reported in hide beetles Dermestes maculatus (Jones & Elgar, 2004) and sand flies Lutzomyia longipalpis (Jones et al., 2000). The pattern we found for O. binodis is remarkably similar to that found in hide beetles, where intermediate aged males had the highest fertility. However, in contrast to our finding, mating and nonmating hide beetles had equal fertility in mid life (Jones & Elgar, 2004). Jones & Elgar (2004) suggest that the reduced fertility of young and old males can be explained in part by the fact that these males transfer fewer sperm. Our data for O. binodis might suggest that sperm ageing also influences fertility. Mating males would have transferred sperm that had been kept in storage for less time than those transferred by nonmating males. The viability of sperm can deteriorate with time in storage (Pellestor et al., 1994; Siva-Jothy, 2000) so that we might expect nonmating males in the older cohorts to have lower fertility than mating males. Sperm storage effects alone would predict an effect of mating on male fertility in both our mid- and late-life samples, which was not the case. Additional effects of accumulating germline mutations (Crow, 1997; Kidd et al., 2001) might offer an explanation for the lower fertility in late-life of both mating and nonmating males.
Following Prowse & Partridge (1997), we measured male fertility in a noncompetitive context. Reductions in the competitive ability of ejaculates may represent a cost of mating that was not measured in our study (Danielsson, 2000; Preston et al., 2001). Although Tomkins & Simmons (2000) failed to find a difference between the competitive abilities of ejaculates produced by previously unmated major and minor males, it remains possible that the male morphs may differ in their ability to sustain sperm-competitive ability in the face of continued mating activity.
In conclusion, we found a significant longevity cost of reproduction for male O. binodis, but no fertility costs of reproduction. Reproduction also increased the time required for males to achieve mating and this effect might be explained by a decline in courtship rate. Minor males had lower courtship rates than major males and were less attractive to females, requiring longer periods of courtship to achieve copulation. Minor males have been shown to have a greater expenditure on the ejaculate (Simmons et al., 1999). Although we did not measure ejaculate expenditure in this study, our results are consistent with the assumption of game theory models of ejaculate expenditure, that minor males trade decreased expenditure on gaining matings for increased expenditure on gaining fertilizations (Parker, 1998).
This work was supported by the Australian Research Council (L.W.S.) and The Academy of Finland (J.S.K.). We thank Dan Kamien for his assistance with the experiments.