It is well established that females of many species exhibit polyandry. Although such behavior often increases female fitness by augmenting fecundity or enhancing the genetic diversity and vigor of their offspring, it often reduces female longevity. It has been argued that trade-offs between these costs and benefits should limit the degree to which females remate. However, the existence of highly polyandrous species suggests substantial polyandry benefits and/or minimal costs in some systems. Females of the leaf beetle, Chrysochus cobaltinus, are extremely polyandrous, providing an opportunity to examine the factors influencing the evolution of such behaviors. We compared the fecundity and longevity of singly mated females, females that mated multiple times with the same male, and females that mated multiple times with different males. Compared with females in the single mating treatment, females in both multiple mating treatments exhibited a significant reduction in latency to oviposition and, due to an increase in daily egg production, significant increases in lifetime fecundity. This difference diminished as the time since last mating increased. There were no differences in fecundity between the 2 multiple mating treatments, indicating that mate identity does not influence the material benefits of multiple mating. Surprisingly, female longevity did not differ among treatments. The pronounced fecundity benefits that females gain from multiple mating, coupled with a lack of longevity costs, apparently explains the extreme polyandry in this species. In addition, the existence of material fitness benefits via conspecific matings raises the intriguing possibility that in a C. cobaltinus–Chrysochus auratus hybrid zone, heterospecific matings may confer similar benefits to Chrysochus females.
For most of the last century, the prevailing view in sexual selection theory was that female reproductive success could be maximized with one or at most a few copulations (Simmons 2005). However, the discovery that females of many species, spanning a broad array of taxa, are polyandrous and benefit from mating with multiple males (Ridley 1988; Arnqvist and Nilsson 2000; Jennions and Petrie 2000) has forced a reexamination of the central tenets of sexual selection theory (Arnqvist and Nilsson 2000; Kokko and Jennions 2003; Zeh JA and Zeh DW 2003; Simmons 2005).
The growing evidence that polyandry is widespread (Ridley 1988; Arnqvist and Nilsson 2000; Jones et al. 2001; Torres-Vila et al. 2004) and even occurs in species that have long been assumed to be monandrous (Gray 1997; Bennett and Owens 2002; Griffith et al. 2002) has prompted a flurry of research on the costs and benefits of multiple mating for females (Arnqvist and Nilsson 2000; Jennions and Petrie 2000; Fox and Rauter 2003; Simmons 2005). These studies have shown a variety of mechanisms by which females may benefit from multiple mating. On one hand, this behavior may directly enhance female fecundity, either by augmenting sperm stores (Simmons 1988; Drnevich et al. 2001; Osikowski and Rafinski 2001) or by providing females with material benefits via oviposition stimulants and/or nutrients that can be used to increase the rate at which offspring are produced (Boggs and Gilbert 1979; Markow and Ankney 1984; Butlin et al. 1987; Burpee and Sakaluk 1993; Fox 1993; Wiklund et al. 1993; Wagner et al. 2001). Such material benefits may also enhance female fitness by increasing their longevity, particularly in species with nuptial feeding (Arnqvist and Nilsson 2000). On the other hand, females may enjoy indirect fitness benefits from multiple mating. Specifically, their offspring may be more genetically diverse or they may tend to be sired by the more vigorous and/or genetically compatible of the males with which they mate (Ridley 1988; Watson 1991; Reynolds 1996; Tregenza and Wedell 1998; Yasui 1998; Arnqvist and Nilsson 2000; Jennions and Petrie 2000; Zeh JA and Zeh DW 2001; Fox and Rauter 2003; Ivy and Sakaluk 2005).
Although females may gain both direct and indirect fitness gains from multiple mating, this behavior can also confer substantial costs on females. Specifically, mating with multiple males requires an investment of time and energy and exposes females to greater risks of diseases and predators, as well as to an increased likelihood of injury during copulation or exposure to toxic compounds in seminal fluids (Daly 1978; Thornhill and Alcock 1983; Wing 1988; Arnqvist 1989; Fowler and Partridge 1989; Hurst et al. 1995; Rice 1996; Chapman et al. 1998; Siva-Jothy et al. 1998). Thus, female longevity may be reduced as a result of multiple mating, particularly in species that lack nuptial feeding (Arnqvist and Nilsson 2000). It has been argued that the balance between the fitness costs and benefits of polyandry should limit the potential for the evolution of extreme levels of polyandry (Arnqvist and Nilsson 2000). Empirical support for this argument comes from studies showing that females of some insects have maximal fitness at intermediate levels of polyandry (e.g., Arnqvist et al. 2004). Ultimately, the balance between direct and indirect fitness gains and decreased longevity may explain why females of most polyandrous species mate with a modest number of males (Drummond 1984; Jones et al. 2001; Strassman 2001). Nonetheless, examples of extreme levels of polyandry (Dickinson 1997; Kraus et al. 2004; Rheindt et al. 2004) suggest that the costs of multiple mating are minimal in some systems and/or that the benefits of this behavior can be rather substantial. However, few studies on the costs and benefits have examined systems in which females are extremely polyandrous (Kraus et al. 2004; Kronauer et al. 2004; Rheindt et al. 2004) leaving us with little insight regarding the evolution of such extreme levels of polyandry.
In this paper, we describe research on the costs and benefits of extreme polyandry to females of the chrysomelid beetle, Chrysochus cobaltinus LeConte (Eumolpinae). Chrysochus cobaltinus has host associations with dogbane (Apocynum) and milkweed (Asclepias) throughout its range in western North America (Sady 1994; Dickinson 1995; Dobler and Farrell 1999). The adults of C. cobaltinus spend the majority of their lives on their host plants, which serve as their food source, mating location, and egg-laying site. Adults eat leaves of the host plant, whereas larvae develop on roots of the same plants (Dickinson 1995).
These beetles have mating behaviors that are remarkably easy to study in the field and manipulate in the laboratory (Peterson, Honchak, et al. 2005). Chrysochus cobaltinus adults are highly polygamous, in that individuals average approximately one mating per day, and males engage in extended periods of postcopulatory mate guarding (Dickinson 1995). Polygynous males and polyandrous females each copulate with an average of 12–13 different mates in the field and mate approximately 26 times on average during their 6- to 8-week adult life span (Dickinson 1995). At the extreme, some females engage in 50 or more copulations (Dickinson 1997), constituting one of the highest reported levels of polyandry to date (c.f., Moritz et al. 1995; Sakaluk et al. 2002; Kraus et al. 2004; Kronauer et al. 2004; Paar et al. 2004; Rheindt et al. 2004). In short, because of the extreme nature of polyandry in Chrysochus beetles and the ease with which their mating can be studied, this system presents an excellent opportunity in which to examine the costs and benefits of extreme levels of polyandry.
To examine whether the extreme level of polyandry in Chrysochus is associated with strong benefits and minimal costs, we assessed whether C. cobaltinus females obtain direct benefits from polyandry and whether highly polyandrous females suffer reduced longevity compared with monandrous females. Furthermore, we determined if the benefits of multiple matings depend on whether females mate multiple times with the same male (repetitive mating) or mate with a diverse array of males (polyandry). Given that the costs and benefits of polyandry have been examined in few species in which polyandry is extreme (Arnqvist and Nilsson 2000), this research improves our understanding of the conditions that favor the evolution of highly polyandrous mating strategies.
Collection and maintenance
We collected all beetles during June 2002 from a C. cobaltinus population occupying a patch of dogbane (Apocynum cannabinum) approximately 4 km northwest of Yakima, Washington, USA. (120.58°N, 46.7°W). This Chrysochus population is outside of the Chrysochus hybrid zone (Peterson et al. 2001; Peterson, Monsen, et al. 2005). To ensure the virgin status of females, freshly emerged adult females were collected by excavation from their underground pupation chambers, in which they remain isolated until they burrow to the surface. Males were a mix of virgin beetles that were collected from their underground pupation chambers and presumably nonvirgin beetles that were collected from the same population but from aboveground dogbane plants. Nonvirgin male beetles were no more than a week old when collected and most likely did not have many mating opportunities within the first week since their emergence because during the first week of emergence, population densities and mating frequencies are quite low (MA Peterson, personal observation). In the field, beetles were placed individually in clear plastic containers (44 ml) and were transferred back to the laboratory. In the laboratory, all beetles were kept in these containers in an incubator (21 °C day, 18 °C night) on a 15:9 h light:dark cycle and given a supply of fresh dogbane cuttings, following methods of Peterson, Monsen, et al. (2005). Apocynum cannabinum cuttings were kept fresh by placing the stem cuttings into a hole in the cap of a 0.6-ml microfuge tube that was filled with water. In order to decrease the accumulation of moisture, a small piece of paper towel was added to each 44-ml container and replaced on a regular basis. In the laboratory, all fully sclerotized adults were sexed following standard methods for this system (Peterson, Honchak, et al. 2005).
To determine if C. cobaltinus females benefit from polyandry, we performed mating experiments in the laboratory approximately 2 weeks after collection. For this experiment, we assigned C. cobaltinus females (96 initially) at random to 1 of 3 treatments (32 females per treatment): 1) treatment S (single), females mated with a single male for a single mating; 2) treatment R (repetitive), females mated 10 times with the same male; and 3) treatment P (polyandry), females mated once with each of 10 different males (10 matings total). Males were randomly drawn from the pool of mixed (virgin and nonvirgin) males (169 total), for pairing with females. All beetle pairs were observed for 3 h per day in individual clear plastic containers (44 ml) at 28 °C in an environment-controlled room. A daily mating period of 3 h was used since on the first day of mating, all females successfully mated within the first 3 h. Furthermore, experiments have shown no difference in mate choice between encounters lasting a half hour and those lasting 3 h (Peterson, Honchak, et al. 2005). After a single copulation, the pair was separated, thus allowing a maximum of one mating per day for each female, the typical mating rate for this species (Dickinson 1995). Because C. cobaltinus males and females have a similar appearance, we marked males with a single dot of nail polish on the elytra to simplify identification of individuals in each pair after copulation. If a pair did not copulate, that specific male was not assigned to that same female again. After all mating attempts, males were placed back in the pool of potential males from which to be drawn for subsequent matings. On average, males participated in 3.2 (±0.2) (mean ± standard error [SE]) successful matings. Mating treatments began on 2 July 2002 and were finished on 19 July 2002. After the completion of their mating treatment, each female was maintained individually until death, under the same conditions as previously stated. The amount of time needed to complete the required matings for individual females ranged from 1 to 18 days (2.9 ± 0.58, 11.8 ± 0.45, and 13.1 ± 0.58 days [mean ± SE] for the single, repetitive, and polyandry treatments, respectively). Once a female began laying eggs, we collected her egg masses (a single egg mass is a clutch of eggs which is covered by a dome-shaped fecal shield) every 2 days over her life span, with some females surviving until September 2002. Subsequently, we determined the number of egg masses, eggs per mass, and total eggs for each female. Total egg production was used as a measure of lifetime fecundity for each female. In addition, we noted the longevity of each female and calculated the average daily production of egg masses and eggs for each female.
We used analysis of variance with repeated measures (ANOVAR) to compare temporal patterns of daily egg production between females in each treatment (α = 0.05), using Tukey's honestly significant difference (HSD) test for the 3 pairwise treatment comparisons (S vs. R, S vs. P, and R vs. P). For this analysis, we compared daily egg production rates across treatments during successive 10-day periods after the initial oviposition by each female. Evidence for material benefits from multiple matings would be an increase in daily egg production with an increase in the number of matings. Furthermore, if those benefits are enhanced as a result of mating with many different males, fecundity should be greater for the polyandry treatment (P) than for the repetitive treatment (R). To assess whether multiple mating (with the same male or different males) increases or reduces longevity, we used analysis of variance (ANOVA) (α = 0.05), using Tukey's HSD test for all pairwise comparisons. Similarly, we used ANOVA to assess the influence of mating treatment on female lifetime fecundity, a response variable that reveals the combined effects of daily egg production with longevity and most accurately reflects the fitness consequences of the different mating regimes. To determine if any differences in daily egg production between treatments were due to differences in either the rate of production of egg masses or the number of eggs laid per mass, we used ANOVA (α = 0.05). In addition, partial correlation analysis enabled us to identify whether lifetime egg production was influenced by the rate of egg mass production and/or the number of eggs produced per mass.
Only those C. cobaltinus females that were alive on 19 July 2002 (the final day of mating treatments) were included in the data analyses. Not all females that began each treatment successfully completed the requisite number of matings. Many females within each treatment completed only a few matings and then died. By including only the females that met the 19 July 2002 cutoff, these less vigorous females were excluded for the analyses for all treatments. By excluding these females, we removed a potential bias that would have underestimated the relative reproductive success of singly mated females had all females that completed their prescribed mating regime been included. In addition, the temporal patterns of daily egg production were assessed only for females that survived at least 30 days after their first oviposition. Too few females survived long enough to allow us to examine temporal patterns in daily egg production beyond that point.
Fifty-one females (17 in each treatment) survived to the 19 July 2002 cutoff date and were included in the full set of analyses. An individual female in the single (S) treatment was determined to be an outlier using Dixon's test statistic for outliers (Dixon 1950; Sokal and Rohlf 1981) in 4 of 6 categories (lifetime fecundity, eggs per day, egg masses, and egg masses per day, but not eggs per mass or longevity) and thus was removed from the analyses. However, inclusion of this female produced qualitatively similar results. The majority of the remaining 50 females survived at least 30 days after their initial oviposition (single mating treatment [S] = 14, repetitive treatment [R] and polyandry treatment [P] = 15 each) and was included in the analyses of daily egg production rates.
Mating treatment had a significant influence on daily egg production rates (ANOVAR: F2,41 = 19.77, P < 0.001), with both multiple mating treatments resulting in substantially higher egg production rates compared with the single mating treatment (Figure 1). Notably, egg production did not differ for females in the 2 multiple mating treatments. For all 3 treatments, time since first oviposition had a significant negative effect on daily egg production rates (ANOVAR: F1,41 = 52.95, P < 0.001; Figure 1). Furthermore, the difference between multiply and singly mated females was most pronounced at the beginning of this period, as revealed by the significant interaction between time and treatment on daily egg production rates (ANOVAR: F2,41 = 4.53, P = 0.017; Figure 1). This interaction is attributable to a more rapid decline in daily egg production in multiply mated females compared with singly mated females. Nonetheless, there was still a significant benefit of multiple mating even 21–30 days after initial oviposition.
The increased daily egg production corresponded with an elevated number of egg masses produced per day by multiply mated females (ANOVA: F2,47 = 12.26, P < 0.001; Figure 2), which, compared with the single mating treatment, increased by 51.7% for the repetitive treatment (R) and 63.2% for the polyandry treatment (P). Multiply mated females also produced a greater number of eggs per mass (ANOVA: F2,47 = 8.27, P = 0.001; Figure 3), with an increase of 22.3% for the repetitive treatment (R) and 31% for the polyandry treatment (P), compared with singly mated females (S). In addition to their elevated rate of egg and egg mass production, multiply mated females had a reduced latency to oviposition after their first mating, compared with singly mated females (ANOVA: F2,47 = 6.069, P = 0.005; Figure 4).
Surprisingly, in spite of the influence of mating treatment on daily egg production, longevity did not differ for females in the 3 treatments (ANOVA: F2,47 = 0.532, P > 0.05; Figure 5). As would be expected given the results for daily egg production, latency to oviposition, and longevity, lifetime fecundity (log transformed for analysis to remedy nonnormality) was significantly greater for females that mated multiply (R, P) relative to that of singly mated females (S) (ANOVA: F2,47 = 10.29, P < 0.001; Figure 6). Indeed, compared with the single mating treatment (S), multiple mating increased lifetime fecundity by 76.3% for the repetitive treatment (R) and 122.8% for the polyandry treatment (P). Partial correlation analyses quantified formally the relationship between lifetime fecundity and both the rate of egg mass production and the number of eggs laid per mass. Specifically, across all treatments, there was a positive correlation between lifetime fecundity and daily egg mass production, when holding eggs per mass constant (r = 0.98, degrees of freedom [df] = 47, P < 0.001). There was also a positive correlation between lifetime fecundity and the number of eggs per mass, when holding the rate of egg mass production constant (r = 0.86, df = 47, P < 0.001). A similar pattern was produced from partial correlation analyses within each treatment (Table 1).
|Egg masses||Eggs per mass|
|Singly mated females (S)|
|Repetitively mated females (R)|
|Polyandrous females (P)|
|Egg masses||Eggs per mass|
|Singly mated females (S)|
|Repetitively mated females (R)|
|Polyandrous females (P)|
For treatment (S), df = 13, P < 0.001; for treatments (R) and (P), df = 14, P < 0.001.
We found evidence that multiple mating by females of the extremely polyandrous leaf beetle, C. cobaltinus, substantially increases their lifetime fecundity. This effect was due to an increase in daily egg production, which in turn was driven by increases in both the number of egg masses produced per day and the number of eggs laid per mass. In addition, multiply mating females began oviposition sooner after their initial mating. Notably, there were no significant differences between the repetitive and polyandry treatments for any of the above comparisons. Thus, there is no evidence at present that mate identity influences the benefits females gain from multiple mating. However, it remains possible that subtle differences among the 2 multiple mating treatments existed but were masked by within-treatment variation in daily egg production stemming from variation in female size. Because we did not quantify female size, we could not include this factor in our analyses. It is important to note that differences in female size are almost certainly not the explanation for the striking daily egg production differences between singly and multiply mating females as females were assigned at random to the various treatments.
Interestingly, in contrast to several recent examples (Fox 1993; Hou and Sheng 1999; Arnqvist and Nilsson 2000), we found that multiply mating females achieved elevated daily egg production rates with no concomitant reductions in longevity, in spite of the fact that this species does not engage in nuptial feeding. Although we did not assess the fertility of the eggs laid by females in the different treatments in the present study, results from related ongoing experiments have shown that the proportion of eggs hatching does not differ for C. cobaltinus females mating once versus 10 times (MA Peterson, M Brassil, E Larson, K Buckingham, and K Monsen, unpublished data). Thus, it appears that C. cobaltinus females derive substantial direct fitness benefits from their extreme level of polyandry. Indeed, the increase in fecundity achieved by the polyandrous females in our experiment exceeded that of 92% of the examples included in Arnqvist and Nilsson's (2000) review of the benefits of multiple mating for female insects.
Although we have not yet established the mechanism by which polyandry confers such strong benefits to C. cobaltinus females, our temporal analyses provide relevant evidence for narrowing down the possibilities. Specifically, the benefits of multiple mating were stronger in the initial 10 days after the onset of oviposition than over the next 20 days. All of the matings for each of the multiply mating females (R and P) were completed within the first 10 days of the onset of their oviposition, with the exception of an individual female in the repetitive (R) treatment that completed her matings 11 days after her initial oviposition. This result indicates that the benefits of multiple mating are greatest during the period of time in which females are engaged in multiple mating. Such a result suggests that the increased daily egg production of multiply mated females is probably due to either a nutrient-rich ejaculate or an oviposition stimulant (e.g., Boggs and Gilbert 1979; Markow and Ankney 1984; Butlin et al. 1987; Fox 1993; Burpee and Sakaluk 1993; Wiklund et al. 1993; Wagner et al. 2001). In contrast, the benefits of multiple mating in this species are apparently not due to increasing limited sperm stores (e.g., Fjerdingstad and Boomsma 1998), for if that were the case, the differences between singly and multiply mated females should increase as the time since the last copulation increased, as the relatively limited sperm stores of singly mated females became depleted. Resolving whether the profound benefits of multiple mating for C. cobaltinus females are due to oviposition stimulants or nutritional ejaculates will require further experimentation. Interestingly, the fact that the benefits of multiple mating decline with time may explain why C. cobaltinus females continue mating throughout their lives. Thus, the evolutionary maintenance of extreme polyandry in C. cobaltinus can be understood in terms of the pronounced direct fitness benefits females receive via multiple mating along with an absence of correlated longevity costs.
It is known that males of many species may be infertile or gametically incompatible with particular females (Ridley 1988; Tregenza and Wedell 1998; Zeh JA and Zeh DW 2001; García-González 2004) and that such male-specific effects may influence the results of mating experiments (García-González 2004). Thus, the relatively poor performance of singly mated females in our experiment may reflect male infertility or genetic incompatibility with the females with which they were paired. However, if this were the case, we would expect female fecundity in the single (S) and repetitive (R) treatments to be identical because in both treatments a single male was randomly paired with a single female and only the number of matings were different (i.e., once vs. 10 times). As the single (S) treatment did differ from the repetitive (R) treatment, it appears unlikely that the reduced fecundity of singly mated females was due to male infertility and/or male–female incompatibility.
Implications for reinforcement in the Chrysochus hybrid zone
Recently, it has been suggested that benefits associated with polyandry may have implications for evolutionary dynamics in hybrid zones (Marshall et al. 2002). Specifically, in many hybrid zones, there is evidence of selection against hybridization, as a result of low hybrid fitness (Howard 1993). However, if polyandrous females gain material benefits from multiple matings with conspecific males, it is possible that they may also realize similar material benefits from heterospecific matings (Marshall et al. 2002). For example, nutrients or oviposition stimulants in the ejaculate of heterospecific males, as with that of conspecific males, may enhance female fecundity. Such benefits could theoretically reduce the cost of mating mistakes for females (Marshall et al. 2002). Furthermore, if females are capable of employing conspecific sperm precedence (Howard 1999), they may be able to mate with many males (regardless of whether they are conspecific or heterospecific), gain the material benefits from those copulations, and fertilize their eggs primarily with conspecific sperm. As a result, they may produce more conspecific offspring than do choosy females that only mate with conspecific males but have fewer lifetime matings. Under these conditions, Marshall et al. (2002) noted that heterospecific matings might be advantageous for females, but disadvantageous for males, creating a sexual conflict with regard to the evolution of premating barriers.
Chrysochus cobaltinus forms a hybrid zone with its sister species, Chrysochus auratus, in south-central Washington, USA. (Peterson et al. 2001; Peterson, Monsen, et al. 2005). Low fitness of hybrids in both the laboratory and the field suggests that this system is one in which selection against hybridization may be operating (Peterson, Monsen, et al. 2005). Additionally, reproductive character displacement in male mate preference supports the notion of selection against hybridization in this system (Peterson, Honchak, et al. 2005). Furthermore, evidence for incomplete conspecific sperm precedence in this system (MA Peterson, K Buckingham, M Brassil, E Larson, M White, and K Monsen, unpublished data) suggests that if females do gain material benefits from multiple matings, this might be a system in which there is a conflict between the sexes with regard to the evolution of premating barriers, as envisioned by Marshall et al. (2002). Thus, our analysis of the costs and benefits of polyandry in C. cobaltinus not only clarifies the selective basis for the extreme level of polyandry in C. cobaltinus but also begins to elucidate the potential for benefits from polyandry to influence the evolution of reproductive barriers in the Chrysochus hybrid zone. To establish whether such conditions promoting sexual conflict in speciation are indeed realized in the Chrysochus hybrid zone, it will be important to determine if females of either species can gain direct benefits from heterospecific matings and if those benefits outweigh the cost of producing hybrid offspring.
We are grateful to J. Bearden, T. Beeman, M. Brassil, K. Buckingham, K. Helm, B. Honchak, D. Juárez, E. Larson, S. Locke, and K. Monsen for their assistance with beetle collection and husbandry. In addition, we thank A. Acevedo-Gutiérrez, R. Anderson, W. Wagner, A. Bourke, and 2 anonymous reviewers for thoughtful comments that improved an earlier version of this manuscript. Funding for this research was provided by National Science Foundation grant DEB-0212652 awarded to M.A.P.