Introduction
Female mating with multiple males within a single reproductive event is a widespread form of polyandry that has profound evolutionary consequences (Pizzari & Wedell, Reference Pizzari and Wedell2013). The resulting ‘polyandry revolution’ (Pizzari & Wedell, Reference Pizzari and Wedell2013) has led to renewed interest in the evolution of polyandry, in particular why females mate multiply and what the ecological and evolutionary consequences of polyandry might be (Boulton & Shuker, Reference Boulton and Shuker2015). However, despite many studies of polyandry in recent years, the costs and benefits of multiple mating for females are still not fully understood (e.g., Arnqvist & Nilsson, Reference Arnqvist and Nilsson2000; Jennions & Petrie, Reference Jennions and Petrie2000; Schwartz & Peterson, Reference Schwartz and Peterson2006). Studies suggest various mechanisms by which females may benefit from multiple mating. Most simply, multiple mating reduces the chance of infertility by restocking depleted sperm supplies (Drnevich et al., Reference Drnevich, Papke, Rauser and Rutowski2001; Osikowski & Rafinski, Reference Osikowski and Rafinski2001; Campbell, Reference Campbell2005), replacing degraded sperm (Wang & Davis, Reference Wang and Davis2006), or offsetting unsuccessful sperm transfer (Pai et al., Reference Pai, Bennett and Yan2005; Hasson & Stone, Reference Hasson and Stone2009). Multiple mating may also directly enhance female fecundity (Arnqvist & Nilsson, Reference Arnqvist and Nilsson2000), a benefit often derived from the sperm or from the chemical substances produced by the male reproductive tract secretory tissues-accessory glands, seminal vesicles, ejaculatory duct, ejaculatory bulb and testes and transferred during copulation (Gillot, Reference Gillot2003; Schwartz & Peterson, Reference Schwartz and Peterson2006; Avila et al., Reference Avila, Sirot, LaFlamme, Rubenstein and Wolfner2011; South & Lewis, Reference South and Lewis2011). The chemical substances induce numerous physiological and behavioral post-mating changes in females. These changes include decreasing receptivity to re-mating, affecting sperm storage parameters, increasing egg production, modulating sperm competition, feeding behaviors, and mating plug formation (Wiklund et al., Reference Wiklund, Kaitala, Lindfors and Abenius1993; Jennions & Petrie, Reference Jennions and Petrie2000; Wagner et al., Reference Wagner, Kelley, Tucker and Harper2001; Gillot, Reference Gillot2003; Avila et al., Reference Avila, Sirot, LaFlamme, Rubenstein and Wolfner2011; South & Lewis, Reference South and Lewis2011). In addition, the chemical substances also have anti-microbial functions and induce expression of anti-microbial peptides in at least some insects (Avila et al., Reference Avila, Sirot, LaFlamme, Rubenstein and Wolfner2011). All these materials may stimulate an increase in the number and rate of development of eggs and modulate ovulation and/or oviposition (Gillot, Reference Gillot2003). All of these material benefits may also enhance female fitness by increasing female longevity, especially in species with nuptial feeding (Arnqvist & Nilsson, Reference Arnqvist and Nilsson2000).
Another interpretation is that polyandry can also enhance a female fitness through indirect genetic benefits (Yasui, Reference Yasui1998). By mating with different males, females have the potential to utilize post-copulatory mechanisms (i.e. sperm competition and/or cryptic female choice) to acquire good genes for their offspring (Fisher et al., Reference Fisher, Double, Blomberg, Jennions and Cockburn2006; Slatyer et al., Reference Slatyer, Mautz, Backwell and Jennions2012), or to reduce the risk of syngamy with sperm carrying incompatible genes (Zeh & Zeh, Reference Zeh and Zeh1996, Reference Zeh and Zeh1997), or damaged genes (Radwan, Reference Radwan2003; Velando et al., Reference Velando, Torres and Alonso-Alvarez2008), or the risk of infertility (Simmons, Reference Simmons2005). The indirect genetic benefit also gains from the ‘sexy sperm’ suggest that males with the best or ‘sexiest’ sperm will attain the highest insemination success (Parker, Reference Parker1970; Simmons, Reference Simmons2001; Simmons & Kotiaho, Reference Simmons and Kotiaho2002). If sperm quality is heritable to some extents a polyandrous female's son will inherit ‘sexy sperm’ from their father (Fisher, Reference Fisher1930). Moreover, such genetic benefits also include offspring with increased genetic diversity (Xu & Wang, Reference Xu and Wang2009), avoidance of inbreeding depression (Nakamura, Reference Nakamura1996; Tregenza & Wedell, Reference Tregenza and Wedell2002; Cornell & Tregenza, Reference Cornell and Tregenza2007), and parasite resistance in offspring (Baer & Schmid-Hempel, Reference Baer and Schmid-Hempel1999). One alternative explanation is that female polyandry is thought to act as a genetic bet-hedging mechanism, by which females can reduce the assessment error in regard to mates’ genetic quality when only uncertain information is available (Yasui, Reference Yasui2001). These benefits, either separately or combined, ultimately increase female fitness, driving females to mate multiple times.
Although females can derive both direct and indirect fitness gains from multiple mating, excessive multiple mating may at the same time bring substantial costs. These include loss of time to feed or oviposit (Keller & Reeve, Reference Keller and Reeve1995), loss of energy (Watson et al., Reference Watson, Arnqvist and Stallman1998), increased risk of sexually transmitted diseases (Arnqvist, Reference Arnqvist1989; Chapman et al., Reference Chapman, Miyatake, Smith and Partridge1998; Watson et al., Reference Watson, Arnqvist and Stallman1998), or increased risk of predation because of reduced mobility or increased visibility (Arnqvist, Reference Arnqvist1989; Chapman et al., Reference Chapman, Miyatake, Smith and Partridge1998). Other costs include damage from deleterious chemicals produced in the male's sperm or accessory glands (Arnqvist & Nilsson, Reference Arnqvist and Nilsson2000; Wigby & Chapman, Reference Wigby and Chapman2005) or increased risk of external (Michiels & Newman, Reference Michiels and Newman1998) or internal (Blanckenhorn et al., Reference Blanckenhorn, Hosken, Martin, Reim, Teuschl and Ward2002) injury of the female reproductive organs from male activity. Consequently, costs of mating can reduce lifetime reproductive fitness through longevity costs and reduced resources available for reproduction.
It has been argued that the balance between the fitness costs and benefits of polyandry should limit the potential for the evolution of extrema levels of polyandry (Arnqvist & Nilsson, Reference Arnqvist and Nilsson2000). Empirical support for this argument comes from studies showing that females of some insects have maximal fitness at intermediate levels of polyandry (Arnqvist et al., Reference Arnqvist, Nilsson and Katvala2004). In this context, the balance between the costs and benefits of polyandry results in most polyandrous females mating with optimal mating rates globally, although there is substantial variation among species and mating systems (Jones et al., Reference Jones, Walker and Avise2001). Nevertheless, examples of extreme levels of polyandry suggest that the occurrence of polyandry might be explained by the existence of minimal costs in combination with substantial benefits (Kraus et al., Reference Kraus, Neumann, Van Praagh and Moritz2004). However, little is known about the balance between costs and benefits in systems in which females are extremely polyandrous. Here, we investigate such a system using Brontispa longissima.
The coconut hispine beetle B. longissima (Gestro) (Coleoptera: Chrysomelidae) is one of the most serious pests of the coconut palm Cocos nucifera L. (Arecaceae), which has multiple uses as the source of economically and industrially important crops such as copra, coconut oil, and coconut shell charcoal (Liebregts & Chapman, Reference Liebregts and Chapman2004; Nakamura et al., Reference Nakamura, Konishi, Takasu, Ku and Chiang2006). B. longissima is thought to have originated in Indonesia or Papua New Guinea, or both, and is currently distributed in Australia and many of the Pacific Islands (Rethinam & Singh, Reference Rethinam, Singh, Appanah, Sim and Sankaran2007). Since the occurrence of a heavy infestation in Vietnam in 2003, the beetle has spread to other mainland Asian countries, including Thailand and China, as well as to the Philippines (Liebregts & Chapman, Reference Liebregts and Chapman2004; Nakamura et al., Reference Nakamura, Konishi, Takasu, Ku and Chiang2006; Lu et al., Reference Lu, Tang, Peng, Salle and Wan2008; Takasu et al., Reference Takasu, Takano, Konishi and Nakamura2010).
B. longissima adults engage in multiple mating throughout the 24-h cycle (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012). Considering that copulation might be random, and that females seldom refuse to mate, even immature females with ovaries not yet fully developed will accept mating, the question is: What do females gain from such profuse multiple mating? Adult females of B. longissima continue to lay small numbers of eggs nearly every day until they die (Yamauchi, Reference Yamauchi1985); it is therefore possible that females copulate multiple times to ensure an adequate supply of sperm over their lifetimes. If so, females that copulate only once may be sperm limited and thus have reduced total egg numbers and reduced egg viability. To evaluate whether the extreme level of polyandry in female B. longissima is associated with substantial benefits and minimal costs, we assessed whether females obtain direct benefits from polyandry and whether highly polyandrous females have shorter lives than monandrous females. Moreover, there have been some comparisons between females mated multiple times with the same male versus a diverse array of males. For instance, as previously reported for polyandry and repeated treatment, the number of mates did not affect the number of eggs laid, but positively affect the hatching success of eggs (Tregenza & Wedell, Reference Tregenza and Wedell1998). We performed this comparison to assess the effects of genetically diverse versus uniform mating partners; as a control representing the possibility of sperm limitation we used females mated only once.
Materials and methods
Study species
B. longissima colonies were originally collected from a Satakentia liukiuensis (Hatusima) plant on Ishigaki Island in 2010 (Okinawa Prefecture, Japan) and reared at the National Institute for Agro-Environmental Sciences laboratory in Tsukuba, Japan. Because it is difficult to obtain fresh leaves of coconut plants in Japan, we reared the beetle colony on fresh leaves of the narrowleaf cattail, Typha domingensis (Pers.), an alternative food source (Winotai et al., Reference Winotai, Sindhusake, Morakote and Arancon2007; Yamashita et al., Reference Yamashita, Winotai and Takasu2008). The preoviposition period and number of eggs laid when the beetle is reared on the leaves of Typha sp. do not differ substantially from those when the beetle is reared on the leaves of C. nucifera (Sugeno et al., Reference Sugeno, Kawazu, Takano, Nakamura and Mochizuki2011). Beetles growth periods on T. domingensis from larval hatching to adult emergence are 26.7 days (Yamashita et al., Reference Yamashita, Winotai and Takasu2008), and female longevity is approximately 65 days in our experiments. Larvae and adults were maintained separately on fresh leaves in 15.5 × 11.5 × 5.0 cm3 (width × length × height) plastic containers with a 6.3 × 3.8 cm2 (width × length) mesh window in the lid. About 50 individuals were maintained together in a container. After emergence, female and male adults were kept separately at 25°C and 65% relative humidity under a 12:12-h (light:dark) photoperiod.
Females have no oocytes in their ovarioles at emergence, and vitellogenesis has begun at 1 week after emergence. Oogenesis has begun at 2 weeks after emergence, and thereafter substantial ovarian development, including the presence of mature oocytes, occurs 3 weeks after emergence (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012). When mating in these experiments, all females used had immature ovaries. Females with undeveloped ovaries can copulate immediately after emergence. Whereas males begin copulating 3 weeks after emergence, and can copulate with females regardless of ovarian developmental stage (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012). Development of the female ovaries began regardless of whether copulation occurred, because even virgin females lay eggs approximately 3 weeks after emergence (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012). Differences in the numbers of eggs, the viability of the eggs, the length of the preoviposition period, and longevity were not detected between females that had copulated immediately after emergence (females with no oocytes) and females that had copulated 3 weeks after emergence (females with mature ovaries) (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012; unpublished data). Because the timing of copulation by females does not affect reproduction, copulation during the preoviposition period might not influence female reproduction (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012). Newly emerged virgin females (0-week-old females) were allowed to mate individually with one 3-week-old virgin male. When using females with no oocytes, we could rule out the possibility that oviposition during and/or prior to the mating treatments influenced the results. The duration of a single copulation lasts approximately half hour. B. longissima is multivoltine, and life table of B. longissima is available in the reference described as (Yamauchi, Reference Yamauchi1985).
Observation of copulation
All observations of copulations were performed as described by Kawazu et al. (Reference Kawazu, Ichiki, Dang and Nakamura2011). In all cases, one 3-week-old virgin male and one 0-week-old initially virgin female were placed on a filter paper sheet (76 × 26 mm2) (width × length) surrounded by glass (2 mm thick, placed on top). Copulation of the beetles, defined as insertion of the male genitalia into the female's, was observed for 2 h day−1 under a 12:12-h light:dark photoperiod at 25°C.
Effects of multiple mating on reproduction
To evaluate the effects of multiple mating on reproductive parameters, we used three treatments: (1) single mating, females mated once with a single male; (2) repetition mating, females mated 10 times with the same male; and (3) polyandrous mating, females mated once with each of 10 different males. Females assigned to (3) the polyandry treatment were divided into 3 groups of 10. The same 10 males were used within each group, such that each female mated first with a virgin male, second with a male who had mated once previously and so on. This method meant those females in (2) the multiple mating and (3) polyandry treatments had the same mating experience with regards to the mating order and mating history of the males. In theory, we designed the test so that sperm quantity and/or quality become identical. Each treatment had 30 replications. In the preliminary test, the males mate equally willing with virgin female versus mated female. If males constantly harass females, resisting mating can be even more costly than mating itself (Rowe et al., Reference Rowe, Arnqvist, Sih and Krupa1994; Watson et al., Reference Watson, Arnqvist and Stallman1998). We eliminated (or at least reduced) the opportunity for male harassment, because we never exposed females to more than one male at a time. All pairs were observed for 2 h day−1 in individual experimental arenas between 8 and 10 h after lights-on, as mating activity in B. longissima is highest at the onset of the photophase and 8–12 h after the lights-on (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012). After a single copulation the pair was separated, thus allowing a maximum of one copulation per day for each female and male. All pairs were kept in the experimental arena only for copulation. The single treatment was therefore performed on only 1 day; the other two treatments were performed over 10 days. The single treatment was mated on day 10 of the mating treatment. The last mating day for females was the same in the single, repetition, and polyandry treatments. After separation from the male, the female was reared on T. domingensis leaves under the same conditions as stated above. B. longissima females were provided with enough food and water prior to and during egg lying, although females were deprived of food and water only during the mating treatment process (2 h day−1). In the case of B. longissima, regardless of multiple mating, food- and water-deprived beetles were weaker, resulting in death within several days (unpublished data). The duration of preoviposition is approximately 3 weeks. No females started laying until after the completion of the mating treatments. Once a female began laying eggs, the eggs (fecundity) were counted, collected, and placed in a Petri dish with a moist paper towel lining the bottom until hatching; hatching success (fertility) was recorded. In addition, we noted the length of the preoviposition period, the period from emergence to termination of oviposition, the ovipositing period, and the longevity of each female.
Statistical analysis
We used analysis of variance in JMP (version 5.0.1 J for windows, SAS Institute Inc., Cary, NC, USA) to compare the preoviposition period, the period from emergence to termination of oviposition (length of time between emergence and termination of oviposition), and the ovipositing period among treatments and in SPSS statistics (version 22 for windows, IBM Corporation, USA) to compare temporal patterns of the numbers of eggs laid by females and proportion of eggs laid that hatched. To evaluate the effect of treatments, data for the preoviposition periods, the period from emergence to termination of oviposition, and the ovipositing period were tested for significance by one-factor ANOVA, and thereafter analyzed by use of Tukey's honestly significant differences test. Female longevity data were analogously analyzed by using parametric survival analysis to assess whether multiple mating (with the same male or different males) increased or decreased longevity in JMP. Pairwise comparisons of the longevity of females among treatments were made by using the log-rank test of the Kaplan-Meier log-rank test (Sokal & Rohlf, Reference Sokal and Rohlf1995). The temporal patterns of the numbers of eggs laid by females between females in each treatment were analyzed using a generalized linear mixed model (GLMM) with a poisson error structure and log link function, in which the individual was incorporated as a random effect with Holm adjustment. This analysis using a GLMM was conducted separately for each of three specified periods; (1) 21–40 days after emergence, (2) 41–60 days after emergence, and (3) 61–70 days after emergence. The explanatory variables were mating treatment and time (day), and the interaction between mating treatment and the time. Both mating treatment and time were regarded as a fixed effect, although mating treatment was the categorical variable and time was the continuous variable. The egg hatchability between females in each treatment was analyzed using a generalized linear model (GLM) with a binomial error structure and logit link function with Holm adjustment. In analysis of the egg hatchability, the mating treatment was regarded as a fixed effect; the proportion of eggs laid that hatched was regarded as a dependent variable.
Results
Overall, egg production in three treatments decreased gradually; in particular, it deteriorated quickly 61–70 days after emergence (fig. 1). The temporal pattern of egg production in each treatment was analyzed using a GLMM separately for each the three specified periods; (1) 21–40 days after emergence, (2) 41–60 days after emergence, and (3) 61–70 days after emergence. At 21–40 days after adult emergence, the GLMM showed that the treatments and the days differed significantly (GLMM: Treatment: F = 105.646, df = 2, P = 0.0001; Time: F = 5.257, df = 19, P = 0.0001) (fig. 1). A significant interaction between treatments and days was detected (GLMM: F = 3.186, df = 38, P = 0.0001). Significant differences in egg production were detected between the polyandry treatment and the repetition treatment (GLMM: F = 10.636, df = 1, P = 0.002), between the polyandry treatment and the single treatment (GLMM: F = 224.334, df = 1, P = 0.0001), and between the single treatment and the repetition treatment (GLMM: F = 126.395, df = 1, P = 0.0001). Significant interactions between treatments and days were detected between the polyandry treatment and the repetition treatment (GLMM: F = 3.382, df = 87, P = 0.001) and between the polyandry treatment and the single treatment (GLMM: F = 4.398, df = 87, P = 0.001). No interaction between treatment and days was detected between the single treatment and the repetition treatment (GLMM: F = 0.948, df = 87, P = 0.522).
Daily mean fecundity of females mated a single time, repetitively with the same male, or with multiple males. N = 30 for each treatment.
At 41–60 days after adult emergence, the GLMM showed that the treatments and days differed significantly (GLMM: Treatment: F = 36.683, df = 2, P = 0.0001; Time: F = 4.619, df = 19, P = 0.0001). No interaction between treatments and days was detected (GLMM: F = 0.825, df = 38, P = 0.767). Significant differences in egg production were detected between the polyandry treatment and the repetition treatment (GLMM: F = 7.932, df = 1, P = 0.007), between the polyandry treatment and the single treatment (GLMM: F = 113.629, df = 1, P = 0.0001), and between the single treatment and the repetition treatment (GLMM: F = 27.034, df = 1, P = 0.0001).
At 61–70 days after adult emergence, the GLMM showed that the treatments and days differed significantly (GLMM: Treatment: F = 7.601, df = 2, P = 0.001; Time: F = 6.892, df = 10, P = 0.0001). No interaction between treatments and days was detected (GLMM: F = 1.530, df = 18, P = 0.076). Significant differences in egg production were detected between the polyandry treatment and the single treatment (GLMM: F = 11.546, df = 1, P = 0.001) and between the single treatment and the repetition treatment (GLMM: F = 5.954, df = 1, P = 0.019). No significant differences in egg production were detected between the polyandry treatment and the repetition treatment (GLMM: F = 1.585, df = 1, P = 0.213). Mating treatment affected the total number of eggs laid by females as well as the total number of hatched eggs (figs 1 and 2). Both multiple mating treatments increased egg production, with the polyandry treatment resulting in the highest total egg production.
Total numbers of eggs laid by females mated a single time, repetitively with the same male, or with multiple males, and total numbers of hatched eggs.
The whole-model effect on the proportion of hatched eggs by the treatments was significant (χ2 = 18.381, df = 2, P = 0.0001) according to the GLM. The proportion of hatching eggs in the polyandry treatment (91.05 ± 0.59%) was significantly higher than that in the single treatment (83.34 ± 1.45%), with the proportion of hatching eggs in the repetition treatment (88.62 ± 0.69%) being intermediate between the polyandry treatment and the single treatment. Significant differences in the proportion of hatched eggs were detected between the polyandry treatment and the repetition treatment (GLM: χ2 = 5.967, df = 1, P = 0.015), between the polyandry treatment and the single treatment (GLM: χ2 = 18.035, df = 1, P = 0.0001), between the single treatment and the repetition treatment (GLM: χ2 = 5.090, df = 1, P = 0.024).
Female longevity did not differ among the three mating treatments (polyandry treatment: 69.33 ± 2.2 days; repetition treatment: 65.51 ± 2.3 days; single treatment: 68.26 ± 3.1 days; parametric survival analysis: χ2 = 1.59, df = 2, P = 0.3601) (fig. 3). In addition, the length of the preoviposition period (polyandry treatment: 21.39 ± 0.37 days; repetition treatment: 20.78 ± 0.33 days; single treatment: 20.32 ± 0.38 days) was not significantly influenced by the mating treatments (ANOVA: F = 2.07, df = 2, P = 0.1315) (fig. 4). The period from emergence to termination of oviposition (polyandry treatment: 65.33 ± 1.80 days; repetition treatment: 61.64 ± 1.87 days; single treatment: 61.14 ± 1.67 days) was not significantly influenced by the mating treatments (ANOVA: F = 0.16606, df = 2, P = 0.1958). Furthermore, the length of the ovipositing period (polyandry treatment: 44.10 ± 1.78 days; repetition treatment: 43.14 ± 1.85 days; single treatment: 41.77 ± 1.65 days) was also not significantly influenced by the mating treatments (ANOVA: F = 0.4672, df = 2, P = 0.6283).
Survival curves of females mated a single time, repetitively with the same male, or with multiple males. N = 30 for each treatment.
Mean length of the preoviposition period, mean length of the period from emergence to termination of oviposition, and mean length of the ovipositing period of females mated with a single time, mated repetitively with the same male, or mated with multiple males. N = 30 for each treatment.
Discussion
Our results show that multiple mating in the extremely polyandrous female B. longissima substantially increases the number of eggs, the number of hatched eggs and the proportion of eggs that successfully hatched (hatching success) (figs 1 and 2). Notably, a significant increase in fecundity and fertility occurred when females copulated with 10 different males as opposed to 10 times with the same male. Thus, mate diversity augmented the direct and indirect benefits that females gained from multiple mating. In contrast, the length of the preoviposition period, the length of the period from emergence to termination of oviposition, the length of the ovipositing period, and female longevity did not differ among treatments (figs 3 and 4). Thus multiply mating females could achieve an increase in daily egg production and hatching success without paying a cost in terms of longevity; B. longissima females can derive substantial direct and indirect fitness benefits from their extreme level of polyandry.
Although our experiment has not yet elucidated the mechanism by which polyandry confers such strong benefits on B. longissima females, our temporal analyses provide relevant evidence to help narrow down the possibilities. Only for 21–40 days after adult emergence was a significant interaction between treatment and days detected between the polyandry treatment and the repetition treatment, and between the polyandry treatment and the single treatment (fig 1), indicating that the differences in daily egg production among treatments were not constant. In other words, the temporal pattern of polyandry treatment was significantly different from those of the other two treatments. Thus, egg production declined significantly faster in the polyandry treatment than in the other two treatments, over 21–40 days after adult emergence. This result suggests that females run out of sperm at different rates in the polyandry treatment. Thus, that egg production declined at a variable rate across females depending on treatments raise the possibility that sperm limitation underlie the benefit of multiply mating in B. longissima (e.g., Fjerdingstad & Boomsma, Reference Fjerdingstad and Boomsma1998; Mac-Diarmid & Butler, Reference Mac-Diarmid and Butler1999; Kraus et al., Reference Kraus, Neumann, Van Praagh and Moritz2004). However, we cannot currently exclude the other possibility that other components of the ejaculate, such as nutrient-rich ejaculate, oviposition stimulant, and/or oocyte maturation stimulant, confer a benefit to polyandrous females (e.g., Boggs & Gilbert, Reference Boggs and Gilbert1979; Markow & Ankney, Reference Markow and Ankney1984; Butlin et al., Reference Butlin, Woodhatch and Hewitt1987; Burpee & Sakaluk, Reference Burpee and Sakaluk1993; Fox, Reference Fox1993; Wiklund et al., Reference Wiklund, Kaitala, Lindfors and Abenius1993; Wagner et al., Reference Wagner, Kelley, Tucker and Harper2001). These possibilities also need to be further explored.
A greater proportion of hatching success when females mate with 10 different males (versus 1 male 10 times) may be attributed to both indirect benefits and direct benefits. High sperm diversity in the spermathecae of polyandrous B. longissima may lead, by any postcopulatory mechanism (e.g., sperm competition or cryptic female choice), to acquire good genes for their offspring (Fisher et al., Reference Fisher, Double, Blomberg, Jennions and Cockburn2006; Slatyer et al., Reference Slatyer, Mautz, Backwell and Jennions2012) or to reduce the risk of syngamy with sperm carrying incompatible genes (Zeh & Zeh, Reference Zeh and Zeh1996, Reference Zeh and Zeh1997) or damaged genes (Radwan, Reference Radwan2003; Velando et al., Reference Velando, Torres and Alonso-Alvarez2008) or avoidance of the risk of infertility (Simmons, Reference Simmons2005). Therefore, the increase in the proportion of eggs that successfully hatched of B. longissima females mated with 10 different partners compared with those of females mated 10 times with one partner, suggesting that indirect genetic benefits (increased offspring fitness, the avoidance of syngamy with sperm carrying incompatible genes, increased genetic diversity, etc.) are acting in combination with direct material benefits, which result in the increased fecundity of multiply mated females (sperm limitation, replenished sperm supply, a nutritional contribution made by the males to the females or zygotes, enhanced paternal care, etc.). Further experiments are required to clarify the mechanism underlying the increase in the proportion of eggs that successfully hatched of polyandrous B. longissima females.
Our experiment revealed no apparent costs of polyandry to females in terms of longevity (fig 3). Because there were no significant lifespan differences between the single and multiple treatments, polyandry per se appears to levy no additional costs over the costs to females of lifetime exposure to males. It has been reported that polyandrous females showed reduced longevity compared with females that had repeated matings in Orthoptera, Lepidoptera, Heteroptera, Diptera, and Coleoptera (South & Lewis, Reference South and Lewis2011). Nevertheless, the lack of lifespan differences between the repetition and polyandry treatments indicates that longevity in B. longissima is unaffected by the genetic diversity of a female's mating partners and/or the interactions among multiple males ejaculates.
Our examination in the presence of sufficient water and food revealed that the direct and indirect benefit that females gain from polyandry, coupled with a lack of longevity costs, accrue. Female Callosobruchus maculatus (Coleoptera: Bruchidae) are polyandrous despite apparent costs to remating, including male harassment and genital tract injury (Rönn et al., Reference Rönn, Katvala and Arnqvist2007), which decrease female life span (Eady et al., Reference Eady, Hamilton and Lyons2007). The direct benefit of polyandry C. maculatus only accrues when females are deprived of food and water, because of the water that is available in the ejaculate (Edvardsson, Reference Edvardsson2007). The most notable effect of costs of mating is shown when females are hydrated, because females do not need water in the ejaculate; in this case polyandry actually reduces their fitness (Edvardsson, Reference Edvardsson2007; Ursprung et al., Reference Ursprung, Hollander and Gwynne2009). Additional work is needed to clarify whether the costs of mating might become apparent if female B. longissima were kept in stressed conditions with limited food and water rather than in relatively comfortable conditions with sufficient food and water.
Adult female B. longissima continued to lay a small number of eggs nearly every day throughout their lives (fig 1). However, fecundity decreased over time, in particular, the egg-laying rate in the three mating treatments deteriorated quickly at 61–70 days after emergence. The observed reductions in the number of eggs laid daily may have been due to physiological changes associated with age, independent of the number of matings. Although B. longissima females were provided with enough food and water during egg lying, females have fewer resources to invest in reproduction as they age, resulting in lower female reproduction. Alternatively, this result also suggested that the benefit conferred by multiple matings is lost at this stage (the end of the reproductive period). In our previous study, differences in the numbers of eggs, the viability of eggs, and the preoviposition period are not detected between females that had copulated immediately after emergence (females with no oocytes) with one 4-week-old virgin male and females that had copulated 3 weeks after emergence (females with mature ovaries) with one 4-week-old virgin male (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012). Previous research has also shown that a temporal difference of 3 weeks does not affect the quality of sperm available to the mated females. It is likely that sperm aging in B. longissima may be low at a young sperm age but may accelerate for higher sperm ages (approximately 60–70 days). This pattern may possibly be observed in the long-term sperm storage of insects (Collins & Donoghue, Reference Collins and Donoghue1999; Reinhart, Reference Reinhart2007). Further experiments are required to clarify this drastic reduction in egg production.
Our results showed that the length of the preoviposition period was not significantly influenced by the mating treatments (fig 4). Previous studies have shown that copulation during the preoviposition period might not influence the length of the preoviposition period (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012). Furthermore, development of the female ovaries began regardless of whether copulation occurred because even virgin females laid eggs approximately 3 weeks after emergence (Kawazu et al., Reference Kawazu, Sugeno, Mochizuki, Takano, Murata and Nakamura2012). Given these previous studies, the present result indicated that female ovary maturation, but not the date of first or latest mating, may determine the onset of oviposition.
We conclude that B. longissima females receive direct and indirect fitness benefits from polyandrous mating without any associated longevity costs; such benefits likely facilitate and maintain the extreme polyandry of this species. Our results should support research into, and programs on, the suppression of B. longissima populations. As the population density of B. longissima increases, adults will engage in increasing numbers of multiple matings, resulting in increased reproductive rates. Therefore, to control populations we need to decrease the number of matings and thus reduce population density. If the sex pheromone of B. longissima is identified, mating disruption may be used as a control technique. Generally, mating disruption using sex pheromone not only inhibits mating behavior but also reduces the number of matings of females. Our results suggested that, if mating disruption is used as a control technique for B. longissima, reduced egg production and reduced hatchability will occur if the number of matings of females is reduced and if the mating is inhibited. As a result, mating disruption may reduce the B. longissima population density.
Acknowledgements
The authors are grateful to Professor Wolf Blanckenhorn of the University of Zurich – Irchel for critical review and for valuable suggestions on their manuscript.
