Abstract

The adaptive value of polyandry in the absence of direct benefits is often assumed to lie in the production of more viable or more attractive offspring, mediated by additive genetic effects. Alternative models propose nonadditive effects through the selective matching of compatible genomes. If genetic incompatibility, for example, through hybridization, inbreeding, or selfish genetic elements, reduces viability of offspring, selection should favor pre- or postcopulatory mechanisms of inbreeding avoidance. Postcopulatory inbreeding avoidance might be achieved by polyandry in combination with cryptic female choice. Because female spiders have paired and independent sperm storage organs that are only filled one at a time, they have been suggested to be ideal organisms to investigate cryptic female choice. Here we used orb-web spiders of the Mediterranean species Argiope lobata to investigate whether females treat ejaculates from siblings or nonsiblings differently. In double-mating trials using sibling and nonsibling males in all possible combinations, we experimentally manipulated which male mated into which sperm storage organ and subsequently counted spermatozoa in these storage organs. This experimental design allowed us to unambiguously assign ejaculates to individual males. We found no differential storage of sperm from first males but a significantly reduced amount of stored sperm from the second male if he was a sibling. Our results suggest that females cryptically chose sperm to trade up to more compatible males through storing different quantities.

Polyandry as an active female mating strategy should only evolve if the benefits outweigh the costs of multiple mating (Arnqvist and Nilsson 2000). Such costs include the time and effort spent searching for additional mating partners (Parker 1970); an increased risk of infection with sexually transmitted diseases (Thrall et al. 2000; Knell and Webberley 2004); possible injuries resulting from male genital penetrations (Simmons and Siva-Jothy 1998; Crudgington and Siva-Jothy 2000; Blanckenhorn et al. 2002); and accessory substances in the male's ejaculate that decrease female life expectancy (Chapman et al. 1995). Multiple mating also reduces foraging opportunities, for example, for female orb-web spiders, where courting males reduce the prey capture rate and increase the predation risk of females (Herberstein et al. 2002).

Benefits of polyandry can be divided into direct material as well as indirect genetic benefits. Direct benefits of polyandry arise when the fitness of a female is increased because males provide better resources for them and their offspring (Kempenaers 2007), for example, the access to additional breeding territories and additional paternal care (Davies and Hatchwell 1992), protection from male harassment and infanticide (Mesnick and Leboeuf 1991; Zinner and Deschner 2000), additional nutrient donations which increase female fecundity (Zeh and Smith 1985), and the acquisition of sufficient sperm for complete fertilization (Ridley 1988).

In combination with a postcopulatory choice mechanism, polyandry can also provide indirect genetic benefits, either through additive or nonadditive processes (Jennions and Petrie 2000). Additive genetic benefits arise if a male passes on intrinsically superior alleles to the offspring (Kokko et al. 2002). Nonadditive genetic benefits arise if a male's alleles are particularly compatible with the female's own alleles, leading to favorable allele combinations in their offspring. This means that particular males are suitable mates for some females but not for others (Zeh JA and Zeh DW 2001).

Failure to choose a compatible male can lead to reduced reproductive success, which in extreme cases can mean total failure to produce viable offspring. These effects may be mediated by cellular endosymbionts (Zeh JA and Zeh DW 1996), transposable elements (Hurst et al. 1992), segregation distorter alleles (Jaenike 1996), maternal effect lethals (Thomson and Beeman 1999), and genomically imprinted genes (Moller and Thornhill 1998), or they can be a result of extreme outbreeding (e.g., in interspecies crosses). One of the most common sources of incompatibility are incestuous matings, which often lower offspring viability through heterozygosity reduction and the increased expression of deleterious recessive alleles (Charlesworth D and Charlesworth B 1987; Tregenza and Wedell 2000). All those incompatibilities should be avoided through pre- or postcopulatory mate choice (Jennions and Petrie 2000). Sex-specific dispersal or kin recognition reduces the risk of inbreeding. However, if a reliable precopulatory recognition of incompatible mating partners is lacking, females should benefit from using postcopulatory choice mechanisms based on the differential treatment of ejaculates from different mating partners (i.e., cryptic female choice; Thornhill and Alcock 1983; Eberhard 1996; Zeh JA and Zeh DW 1997).

There is evidence of cryptic female choice for indirect benefits, for example, females of the black field cricket Teleogryllus commodus discriminate postcopulatorily against unattractive males (Bussiere et al. 2006) and cryptic discrimination against sibling males has been shown for the field cricket Gryllus bimaculatus (Tregenza and Wedell 2002) and has been confirmed by a marker-based paternity analysis (Bretman et al. 2004). Furthermore, cryptic female choice against sibling males has been shown in house mice (Firman and Simmons 2008). However, the results for related species are ambiguous (Jennions et al. 2004; Simmons et al. 2006).

Simultaneous storage of sperm from different sires may enable the female to select the best possible father for her offspring. Postcopulatory choice of a partner may be favored by selection in situations of sequential mate choice and unpredictable mating rates (Kokko and Ots 2006). This is the case in many polyandrous spiders where females occupy webs and are visited by roving males. A major selective advantage could be gained if cryptic choice enables females to avoid costs of inbreeding. Mating with related males is known to reduce fecundity, egg-hatching success, and offspring viability in numerous taxa, including spiders (Bilde et al. 2007). Although long-distance dispersal is known for spiders, philopatry does occur as well (Walter et al. 2005). Hence, in the absence of precopulatory inbreeding barriers, which are unknown for spiders, siblings may frequently encounter each other as potential mates (Bilde et al. 2005). Polyandry, however, does not only open opportunities for sexual selection through female choice but also to sperm competition, which is defined as the competition between ejaculates of different males for the fertilization of a given set of ova (Parker 1998). These 2 processes are often difficult to distinguish. Both are possible in entelegyne spiders that possess a complex genital morphology with 2 independent sperm storage organs (spermathecae) that are filled during 2 separate mating events either by the same male or by different males. The spermathecae in entelegyne spiders are connected separately with the uterus via 2 independent fertilization ducts. Males of entelegyne spiders inseminate females by using their paired secondary mating organs, the pedipalps, which are filled with sperm immediately after the final molt. In some spiders, including our study species, male insertion patterns are fixed such that a male can only use his right pedipalp to fill the right spermatheca and vice versa (Hellriegel and Ward 1998).

Here we investigated experimentally whether females of the orb-web spider Argiope lobata avoid inbreeding by means of cryptic female choice. To this end, we conducted double-mating trials in which each female copulated with 2 nonsibling males, with 2 sibling males, or a combination of both. Because the paired spermathecae of females are filled during independent copulations, and according to a fixed insertion pattern, it was possible to assign individual males to mate into 1 specific spermatheca by selectively inducing ectomy of 1 of their pedipalps.

We estimated male reproductive success by counting sperm stored inside the spermathecae and by assigning relative paternity using the sterile male technique (Parker 1970).

MATERIALS AND METHODS

Study animals

For our experiments, we used the sexually cannibalistic orb-web spider A. lobata (Pallas 1772). Less than 50% of 117 virgin males mated to virgin females fell victim to sexual cannibalism after their first copulation. As in all Argiope species studied, males always die during or shortly after their second copulation (Foellmer and Fairbairn 2003; Gaskett et al. 2004).

Experiments were conducted in a laboratory at the University of Hamburg from March 2006 until June 2006. All spiders used for this study were F1 descendents derived from 9 mated females collected in June 2005 near the national park of Cabo de Gata, Spain. The females were collected along a dry riverbed within a distance of 500 m. All 9 females laid up to 5 egg sacs in the laboratory. The first egg sacs contained 220 ± 16 eggs each.

Spiderlings that hatched from those egg sacs were initially kept in groups in plastic containers (15 × 12 × 7 cm). They were fed ad libitum with Drosophila species twice a week and sprayed with water on 5 days of the week. After the spiderlings were large enough to build their own webs, they were separated and housed individually in 250 ml upturned plastic cups.

Once the still immature spiders could be sexed (males have bulbous pedipalps and females are larger), females were moved into larger, 330 ml, cups. Depending on their body size, spiders were fed with either Drosophila species ad libitum or with 3 Calliphora flies twice a week. They were sprayed with water every second day and checked for molts daily. The adult stage can easily be recognized because of the differentiated secondary mating organs of males (the paired pedipalps) and the sclerotized epigyne in females.

After their final molt, females were transferred into perspex frames (36 × 36 × 6 cm), where they built webs in which the mating trials took place. Males stayed in plastic cups until used for mating experiments. All animals were weighed within 1 day after their final molt. Females were weighed again immediately after copulation trials. They were not weighed before in order to avoid damage to the capture web. Males were weighed before copulation trials because they often lost legs or even died during copulation.

Mating procedure

All animals used for this study were unmated, and the family origin of each individual was known (they all derived from 9 families). Each male was used only once and was assigned after a preset schedule to an irradiation treatment and a mating treatment. Females were assigned to 1 of the 4 double-mating treatments in which they were mated to 2 nonsibling (NN), 2 sibling (SS) males, or 1 nonsibling and 1 sibling male in random order (NS and SN). (Because the mating history of their wild-caught mothers was unknown, siblings may have been half siblings or full siblings).

Males use only 1 pedipalp per copulation bout, and copulation follows a fixed ipsilateral insertion pattern meaning that the right pedipalp is always used to inseminate the right spermatheca and the left pedipalp inseminates the left spermatheca. This enabled us to predetermine which male inserted into which of the 2 spermathecae of a female. Thus, we were able to exclude direct sperm mixing by ensuring that each male inseminated into an unused spermatheca. We achieved this by selectively amputating 1 of the 2 pedipalps, so that males used in a double-mating trial had complementary pedipalps. The pedipalp to be amputated was pressed at its femur with tweezers until it was ectomized by the male. This was done at least 1 day before the experimental trial. As in previous studies on other spiders (Rovner 1967; Nessler et al. 2007), we noticed no obvious effect of pedipalp removal on male behavior.

Based on their weight, age (measured in days since maturation), and family identity, we assigned males and females to the different mating treatments. We tried to match males with similar weight and age between treatment groups to reduce possible confounding effects of these parameters.

Females in the different treatments did not differ in weight after final molt, weight after copulation, and age at copulation date (weight after final molt: Kruskal–Wallis test: χ2 = 2.79, P = 0.43, N = 79; weight after copulation: Kruskal–Wallis test: χ2 = 3.08, P = 0.38, N = 86; age at copulation date: one-way analysis of variance [ANOVA]: F3,81 = 1.70, P = 0.17).

Males in the different treatment groups did not differ in weight after final molt, weight before copulation, and age at copulation (first males—weight after final molt: Kruskal–Wallis test, χ2 = 2.47, P = 0.48, N = 85; weight before copulation: one-way ANOVA: F3,82 = 1.3, P = 0.28; copulation age: Kruskal–Wallis test: χ2 = 2.99, P = 0.39, N = 85; second males: weight after final molt: Kruskal–Wallis test, χ2 = 4.1, P = 0.25, N = 85; weight before copulation: one-way ANOVA: F3,81 = 2.07, P = 0.11; age at copulation: one-way ANOVA: F3,81 = 0.65, P = 0.59).

Subsets were used for different analyses (sperm counts and paternity, see below).

All copulations took place in perspex frames in which virgin females had previously built their webs. The virgin males were carefully introduced into one of the upper corners of the web with a soft paintbrush. Males were given a time limit of 1 h to make a copulation attempt; otherwise, they were replaced.

We mated 118 females to 236 males and carefully documented their behavior. Copulation duration was measured using a stopwatch, and we noted whether a male was killed by the female. Cannibalization of a male could have a direct benefit for the female by gaining additional nutrient donations. Therefore, we removed the male from the web before he was devoured by the female.

The initial intention was to use 50% of the females for measuring fecundity, but egg laying failed in many cases due to the undesired introduction of parasitic flies into our Drosophila cultures. Females were infected by those parasites after mating, and these flies laid eggs into the egg sacs and into the epigynal cavities of the females causing almost complete breeding failure. Therefore, sample sizes vary greatly between the behavioral and reproductive data sets. Only uninfected females were used for sperm counts and paternity analysis.

Sperm counts

A subset of 37 females was used for counting the number of sperm stored in their 2 spermathecae. Sperm counts were conducted from females derived from the SN and NS treatments (NS: N = 10, SN: N = 8) and from 19 females from the NN (N = 9) and SS treatments (N = 10). The subsets were compared again for differences in weight and age. Females did not differ in weight and age between the treatments (weight after final molt: Kruskal–Wallis test: χ2 = 3.94, P = 0.27, N = 34; weight after copulation: one-way ANOVA: F3,33 = 1.12, P = 0.35; age at copulation date: one-way ANOVA: F3,33 = 1.79, P = 0.17). First males showed no significant difference in weight and age between treatments. Second males were not significantly different in age but although we tried to match males with similar weights between treatment groups, sibling and nonsibling males differed in weight after final molt and weight before copulation. Hence, we included these variables into the analysis.

Mated females were killed by hypothermia and fixed in 70% ethanol after death or after they had laid their fourth egg sac. We assumed that sperm numbers stored in the spermathecae are not influenced by the production of eggs as was shown for another orb-web spider, Nephila plumipes (Schneider and Elgar 2001). After dissecting out the spermathecae, under a stereomicroscope, they were separately placed in Eppendorf low bind tubes with 80 μl Casytone solution (Casytone, Schärfe System, Reutlingen, Germany). Using a small mortar, each spermatheca was crushed inside the Eppendorf tube, and spermatozoa were dissolved in the Casytone solution. Spermatozoa were prevented from agglutination by an ultrasonic treatment (Bandelin UW 2070, Bandelin Electronic GmbH, Berlin, Germany). Therefore, it was necessary to concentrate spermatozoa on the ground of the tube by centrifuging them for 60 s with 5000 g in an Eppendorf MiniSpin. Ultrasonication was applied 5 times for 60 s with 60% power and with a 20 s break between bouts. After the last ultrasonic bout, samples were mixed for 60 s on a vortex test tube mixer to obtain an even distribution of spermatozoa in the solution, and the samples were vortexed again for 15 s immediately before counting the samples on a 0.1 mm deep hemocytometer (improved Neubauer). Spermatozoa were counted with a 400× magnification with photographs taken under a light microscope using a Leica DFC 320 digital camera and the software Leica IM 4.0.

Paternity analysis

We used the sterile male method to determine patterns of paternity in double-mating trials (see above). Males were irradiated with a dose rate of 0.8 Gy/min for 50 min, amounting to a dosage of 40 Gray ± 5% (Gulmay X-Strahl RS225, Labor für Strahlenbiologie und Radioonkologie, Universitätsklinikum Eppendorf). Males were divided into weight classes and then randomly allocated to either the normal or the irradiated group.

In order to control for sterilization success, we used double matings with 2 irradiated males. No spiders hatched out of eggs laid by females mated with 2 sterilized males (N = 7), confirming that sterilization was successful. Additionally, we conducted double matings in which females were mated to 2 fertile nonsibling males in order to estimate a correction factor for unhatched eggs in a natural clutch (N = 13). However, only 2 of those females produced viable egg sacs (varying from 19% to 76% unhatched eggs), and we were not able to calculate a correction factor. Although we cannot determine sperm precedence patterns of first and second males, controls are not necessary to assess the relative paternity success because the reciprocal treatment effectively corrects for a systematic bias due to the sterilization method.

After mating, most females laid eggs within 4 weeks. Perspex frames were checked daily for egg sacs. Egg sacs were carefully removed from the frames and placed in small gauze-covered plastic vials. If egg sacs and females were unharmed by the parasites, they were fixed in 70% ethanol after a developmental period of 4 weeks in order to count eggs as well as developed spiders. Even if not all developed spiders hatched out of the eggs, it was visible under a stereomicroscope if an embryo had developed. Developed eggs were counted and assigned to the fertile male; undeveloped eggs were assigned to the sterile male. The data are given as the proportion of eggs sired by the second male to mate with a female (P2).

The frequency of cannibalism (here defined as the killing, not the consumption of a male) did not differ between sterile and fertile males.

Statistical analysis

The data were analyzed using linear mixed models (LMMs) with restricted maximum likelihood (REML) approximation for sperm counts of the second male as well as an LMM with REML approximation for P2 values. All spiders derived from 9 female families, and family ID was included as random factor. Sperm counts of the first male had to be log transformed prior to the analysis to reach normal distribution. P2 values were arcsine transformed to reach normal distribution.

Variables that entered the models are specified in the Results. Wilcoxon tests presented in the Results are Wilcoxon rank-sum tests for nonparametric independent samples. All statistical tests were done in JMP 7.1.

RESULTS

Behavioral data

We observed no reluctance of males or females to mate with their siblings. We measured the copulation duration and the occurrence of sexual cannibalism in 118 double matings. The mean duration of a copulation was 49.15 ± 8.5 s for the first copulation (N = 112, 5 outliers were excluded, 1 data point was missing) and 58.95 ± 10.69 s for the second copulation (N = 108, 8 outliers were excluded, 2 data points were missing). Outliers were defined as extreme long copulations up to 120 min. Those males were not attacked by females but died nevertheless, eventually falling off the female's opisthosoma when she moved in the web. Sexual cannibalism occurred in 47% of first matings (55 of 117, 1 data point missing) and in 43% of second matings (50 of 117, 1 data point missing) and did not differ between siblings and nonsiblings (first mating: among males cannibalized, 22 were siblings and 26 nonsiblings: χ2 = 0.18, P = 0.67, N = 102, 1 data point missing; second mating: among males cannibalized, 22 were siblings and 22 nonsiblings: χ2 = 0.01, P = 0.93, N = 103). Copulation durations of siblings and nonsiblings did not differ, neither in first matings (Wilcoxon: Z = 1.5, P = 0.13, N = 98, 4 outliers excluded, 1 data point missing) nor in second matings (Wilcoxon: Z = −0.17, P = 0.87, N = 97, 3 outliers excluded, 3 data points missing).

Sperm counts

Stored sperm numbers from first males did not differ depending on the males’ sibship status (t-test 37 = −0.71, P = 0.48, Figure 1). However, second males were associated with higher sperm numbers if they were a nonsibling rather than a sibling (Wilcoxon: Z = 2.23, P = 0.03, N = 37; Figure 1).

Figure 1

Sperm numbers stored in female spermathecae inseminated by sibling and nonsibling males separated by copulation order. Sperm numbers did not differ between siblings and nonsiblings in first copulations (P = 0.48), whereas females stored significantly more sperm of nonsibling males in the second copulation (P = 0.03). The results for the LMM are presented in Table1.

Figure 1

Sperm numbers stored in female spermathecae inseminated by sibling and nonsibling males separated by copulation order. Sperm numbers did not differ between siblings and nonsiblings in first copulations (P = 0.48), whereas females stored significantly more sperm of nonsibling males in the second copulation (P = 0.03). The results for the LMM are presented in Table1.

Table 1

LMM (REML) of number of sperm stored from a second male depending on sperm stored from the previous male and the relatedness of the male to the female

Parameter F P 
First ♂ sibling or nonsibling 0.002 0.97 
Second ♂ sibling or nonsibling 7.51 0.01 
Log sperm numbers from first ♂ 8.32 0.01 
First ♂ sibling or nonsibling × second ♂ sibling or nonsibling 1.65 0.21 
Parameter F P 
First ♂ sibling or nonsibling 0.002 0.97 
Second ♂ sibling or nonsibling 7.51 0.01 
Log sperm numbers from first ♂ 8.32 0.01 
First ♂ sibling or nonsibling × second ♂ sibling or nonsibling 1.65 0.21 

The model explained 58% of the variation. The random factor female family was not significant.

Sperm numbers varied enormously, and the means were strongly influenced by individual differences between females. There was a significant and positive correlation between sperm stored in a female's 2 spermathecae (pairwise correlation: r = 0.63, P < 0.0001, N = 37), indicating that sperm storage differed consistently between females. To account for this variation between females, we computed an LMM with REML that initially included a list of factors that have been shown to influence sperm counts and paternity in other studies. We initially included body weight of females and of first and second males, age of females and of first and second males, copulation duration of both males, cannibalism of both males, and whether a female had produced an egg sac. Because none of these variables significantly improved model fit, they were subsequently removed from the model. Removal resulted in a combined reduction of 6% of the explained variation.

The final, reduced model contained female family as random factor and sperm count (log) from the first male and first and second males’ sibship status as fixed factors. It was irrelevant whether the first male was related to her or not. The relatedness to the second male, however, explained a significant part of the variation (Table 1). The inclusion of female family as a random factor made no difference to the results. This model contained all 4 treatments—leaving out the control treatments would result in a very small sample size.

We found no correlation between sperm numbers and copulation duration (copulation 1: Pearson product–moment correlation: r = −0.11, degrees of freedom (df) = 16, P = 0.67; copulation 2: r = −0.04, df = 17, P = 0.86).

Paternity analysis

We assigned P2 values to doubly mated females that had copulated with sibling and nonsibling males in all possible combinations. We found a difference between the 4 treatments (Kruskal–Wallis test; P = 0.048, N = 38, χ2 = 7.93). P2 values of nonsibling males were higher than those of sibling males in the mixed treatments (sibling–nonsibling: 0.75 ± 0.12, N = 10; nonsibling–sibling: 0.48 ± 0.13, N = 12; Figure 2), but, perhaps because of the limited sample size (see parasite-induced deaths described in Materials and methods), the difference was not significant (Wilcoxon: Z = 1.45, Nsib = 12, Nnonsib = 10, P = 0.14). P2 values were 0.89 ± 0.11 (N = 9) in the nonsibling (NN) control group and 0.58 ± 0.17, N = 7, in the sibling (SS) control group (Figure 2). We ran an LMM with REML approximation on the arcsine transformed P2 values with female family as random factor and sibship status of first and second male and their interaction as fixed factors, which included all 4 treatments. Inclusion of female family as random factor had no influence on the results. The model showed a significant influence of sibship status of the second male on the P2 values (Table 2).

Table 2

LMM (REML) of the arcsine transformed P2 values depending on sibship of the first and second males

Parameter F P 
First ♂ sibling or nonsibling 0.06 0.81 
Second ♂ sibling or nonsibling 4.33 0.046 
First ♂ sibling or nonsibling × second ♂ sibling or nonsibling 0.43 0.52 
Parameter F P 
First ♂ sibling or nonsibling 0.06 0.81 
Second ♂ sibling or nonsibling 4.33 0.046 
First ♂ sibling or nonsibling × second ♂ sibling or nonsibling 0.43 0.52 

The model explained 29% of the variation. The random factor female family was not significant.

Figure 2

P2 values of the first egg sacs of the 4 double-mating treatments. Variation of the P2 values is explained by sibship status of the second male. Nonsibling (N) males sire more offspring than sibling males (S). The results for the LMM are presented in Table 2.

Figure 2

P2 values of the first egg sacs of the 4 double-mating treatments. Variation of the P2 values is explained by sibship status of the second male. Nonsibling (N) males sire more offspring than sibling males (S). The results for the LMM are presented in Table 2.

DISCUSSION

Studies of postcopulatory female choice for compatible males are scarce, and the outcomes vary between closely related species even if they show similar degrees of inbreeding depression (Tregenza and Wedell 2002; Jennions et al. 2004). We used the entelegyne orb-web spider A. lobata to test whether females possess the ability to cryptically favor sperm of unrelated and thus more compatible males. Our results suggest that this is the case. Despite a limited sample size and large variation in sperm numbers stored by individual females, we found a lower storage of sperm from related males. Differences in paternity of second males according to their relatedness show that these differences in sperm storage translate into reduced paternity of siblings although significance was marginal.

The detected reduction in sperm numbers and in paternity of siblings is highly variable and is obscured by large individual variation in females’ sperm storage. There may be several reasons why the detected pattern was not more pronounced; first, inbreeding depression may be moderate in F1 matings, resulting in some degree of inbreeding tolerance in a sequential mate choice scenario (Kokko and Ots 2006). Furthermore, because males may differ in the number of spermatozoa they carry in their pedipalps (Schneider et al. 2006), females may have limited control over the amount of sperm transferred. For example, if the second male has fewer sperm available than the number already stored from the first male, then the female cannot favor the second male by storing a higher number of his sperm.

In fact, this source of uncertainty may explain the observation that, in many polyandrous spiders, females are observed to attempt premature termination of their first copulation (Foellmer and Fairbairn 2004; Snow and Andrade 2004). The pinnacle of this may be sexual cannibalism, resulting in males limited to only a single copulation. This happens to more than 40% of the males in A. lobata and to 80% of the males in the congener Argiope bruennichi (Schneider et al. 2006). Interestingly, mating plugs through parts of male genitalia have been found in several of these cannibalistic species (Foellmer 2008). Depending on the efficiency of these plugs, females may be selected to retain a virgin spermatheca for a second male, thus securing the opportunity for polyandry rather than being monopolized by a single male.

Variation in sperm transfer unlikely results from male strategic decisions because in Argiope, males use each of their pedipalps only once (Herberstein et al. 2005). Hence, males gain nothing by saving sperm for future copulations (Eberhard 2004). Additionally, spider males can otherwise only avoid inbreeding through precopulatory mechanisms, which were not observed in our study species and are not reported from other spiders.

Our study design excluded the possibility of sperm mixing during storage; ejaculates were stored in different spermathecae. However, there was a possibility of sperm mixing at a later stage because ejaculates from different males may come into contact in the uterus during the fertilization process (Morishita et al. 2003).

The proximate processes through which females achieve a bias in sperm numbers stored or used from particular males are unknown for most spiders including our study species. In the congener A. bruennichi and possibly in Argiope keyserlingi as well, sperm transfer is an approximately linear function of copulation duration, and relative sperm numbers directly translate into relative paternity (Schneider et al. 2006). By selectively terminating copulation, for example, in the form of sexual cannibalism, females in these species can exert control over sperm transfer (Andrade 1996; Snow and Andrade 2004; Herberstein et al. 2005). However, a relationship between copulation duration and sperm transfer has not been found in several other spider species, and copulation duration may be associated with functions other than sperm transfer, such as the manipulation of female genitalia or physiology (Bukowski and Christenson 1997; Bukowski et al. 2001; Snow and Andrade 2004). Unlike in the congeners, but in accordance with other species, we did not detect a correlation between sperm numbers and copulation duration in A. lobata. This suggests that the simple option of selectively terminating copulation is not available to females and that they manipulate sperm numbers in other ways. Sperm dumping is another option reported for a number of species (Snook and Hosken 2004; Cordoba-Aguilar 2006) and recently for haplogyne spiders, for example, Silhouettella loricatula biases paternity by selectively dumping sperm (Burger 2007), and a similar postcopulatory mate choice mechanism has been proposed for Antrodiaetus unicolor (Michalik et al. 2005). Haplogynes do not possess 2 separate spermathecae, but they store sperm in a bursa where ejection is easier to achieve than in entelegyne spiders. Sperm digestion, as reported in flatworms (Vreys et al. 1997), mollusks (Haase and Baur 1995), and mammals (Birkhead et al. 1993), may be another possibility. Sperm selection requires the ability to recognize individual sperm. Such mechanisms have not (yet) been discovered in spiders. In mammals, however, antisperm leucocytes and antibodies have been described that impair fertilization success of incompatible spermatozoa (Zeh JA and Zeh DW 1997), and in house flies, sperm are disabled by accessory gland secretions (Holman and Snook 2006). The morphology of entelegynes with separate storage sites for different ejaculates opens the interesting option of selective activation of ejaculates. Spider males transfer coated spermatozoa that need to be activated to become mobile and swim to the place of fertilization (Brown 1985). However, the mixed paternity patterns found in the present study are not consistent with selective activation of sperm from only one or the other spermatheca, which should have led to a complete paternity bias toward the unrelated male.

The difference we have found in stored sperm numbers from sibling and nonsibling males was only present in second copulations. Such a mating order effect is consistent with the “trade-up” hypothesis (Halliday 1983), which suggests that females choose between sequential mating partners (Gabor and Halliday 1997). This could be explained by the fact that A. lobata females may live with a certain degree of uncertainty as to whether they will be encountered by more than 1 male. Hence, it may be adaptive if they initially accept the sperm they can get and regulate relative fertilization success only after polyandry is ensured. During their second copulation, they may then decide to up- or downregulate sperm uptake depending on the relative compatibility or quality of the second male. This has been shown for number of species, for example, flour beetles (Fedina and Lewis 2007), guppies (Pitcher et al. 2003), three-spined sticklebacks (Bakker and Milinski 1991), and smooth newts (Gabor and Halliday 1997).

In conclusion, our study provides suggestive evidence that female A. lobata can reduce inbreeding by cryptically favoring the sperm of a more compatible male under polyandry. Our data add to the slowly increasing body of literature reporting similar phenomena. However, several assumptions and particularly the mechanisms involved require clarification in future studies. In-depth morphological investigations of the genital organs of the female, and studies of the fertilization process within females, are possible next steps in this endeavor. Multigenerational studies are necessary to assess the extent of inbreeding depression and hence to quantify selection on cryptic female choice.

FUNDING

Deutsche Forschungsgemeinschaft (SCHN 561/5-2 to J.M.S.).

We thank Dr Ingo Brammer at the Labor für Strahlenbiologie und Experimentelle Radioonkologie at the Universitätsklinikum Hamburg-Eppendorf for his help with the sterilization; Dr Lutz Fromhage for his help with the manuscript; and Dr Peter Michalik, Dr Jordi Moya-Larano, Dr Gabriele Uhl, Stefan Nessler, Tomma Dirks, and Angelika Taebel-Hellwig for their valuable support.

References

Andrade
MCB
Sexual selection for male sacrifice in the Australian redback spider
Science
 , 
1996
, vol. 
271
 (pg. 
70
-
72
)
Arnqvist
G
Nilsson
T
The evolution of polyandry: multiple mating and female fitness in insects
Anim Behav
 , 
2000
, vol. 
60
 (pg. 
145
-
164
)
Bakker
TCM
Milinski
M
Sequential female choice and the previous male effect in sticklebacks
Behav Ecol Sociobiol
 , 
1991
, vol. 
29
 (pg. 
205
-
210
)
Bilde
T
Lubin
Y
Smith
D
Schneider
JM
Maklakov
AA
The transition to social inbred mating systems in spiders: role of inbreeding tolerance in a subsocial predecessor
Evolution
 , 
2005
, vol. 
59
 (pg. 
160
-
174
)
Bilde
T
Maklakov
AA
Schilling
N
Inbreeding avoidance in spiders: evidence for rescue effect in fecundity of female spiders with outbreeding opportunity
J Evol Biol
 , 
2007
, vol. 
20
 (pg. 
1237
-
1242
)
Birkhead
TR
Moller
AP
Sutherland
WJ
Why do females make it so difficult for males to fertilize their eggs
J Theor Biol
 , 
1993
, vol. 
161
 (pg. 
51
-
60
)
Blanckenhorn
WU
Hosken
DJ
Martin
OY
Reim
C
Teuschl
Y
Ward
PI
The costs of copulating in the dung fly Sepsis cynipsea
Behav Ecol
 , 
2002
, vol. 
13
 (pg. 
353
-
358
)
Bretman
A
Wedell
N
Tregenza
T
Molecular evidence of post-copulatory inbreeding avoidance in the field cricket Gryllus bimaculatus
Proc R Soc Lond Ser B Biol Sci
 , 
2004
, vol. 
271
 (pg. 
159
-
164
)
Brown
S
Mating-behavior of the golden-orb-weaving spider Nephila clavipes: II. sperm capacitation, sperm competition, and fecundity
J Comp Psychol
 , 
1985
, vol. 
99
 (pg. 
167
-
175
)
Bukowski
T
Christenson
T
Determinants of sperm release and storage in a spiny orbweaving spider
Anim Behav
 , 
1997
, vol. 
53
 (pg. 
381
-
395
)
Bukowski
TC
Linn
CD
Christenson
TE
Copulation and sperm release in Gasteracantha cancriformis (Araneae: Araneidae): differential male behaviour based on female mating history
Anim Behav
 , 
2001
, vol. 
62
 (pg. 
887
-
895
)
Burger
M
Sperm dumping in a haplogyne spider
J Zool
 , 
2007
, vol. 
273
 (pg. 
74
-
81
)
Bussiere
LF
Hunt
J
Jennions
MD
Brooks
R
Sexual conflict and cryptic female choice in the black field cricket, Teleogryllus commodus
Evolution
 , 
2006
, vol. 
60
 (pg. 
792
-
800
)
Chapman
T
Liddle
LF
Kalb
JM
Wolfner
MF
Partridge
L
Cost of mating in Drosophila melanogaster females is mediated by male accessory-gland products
Nature
 , 
1995
, vol. 
373
 (pg. 
241
-
244
)
Charlesworth
D
Charlesworth
B
Inbreeding depression and its evolutionary consequences
Annu Rev Ecol Syst
 , 
1987
, vol. 
18
 (pg. 
237
-
268
)
Cordoba-Aguilar
A
Sperm ejection as a possible cryptic female choice mechanism in Odonata (Insecta)
Physiol Entomol
 , 
2006
, vol. 
31
 (pg. 
146
-
153
)
Crudgington
HS
Siva-Jothy
MT
Genital damage, kicking and early death: The battle of the sexes takes a sinister turn in the bean weevil
Nature
 , 
2000
, vol. 
407
 (pg. 
855
-
856
)
Davies
NB
Hatchwell
BJ
The value of male parental care and its influence on reproductive allocation by male and female dunnocks
J Anim Ecol
 , 
1992
, vol. 
61
 (pg. 
259
-
272
)
Eberhard
WG
Female control: sexual selection by cryptic female choice
 , 
1996
Princeton (NJ)
Princeton University Press
Eberhard
WG
Why study spider sex: special traits of spiders facilitate studies of sperm competition and cryptic female choice
J Arachnol
 , 
2004
, vol. 
32
 (pg. 
545
-
556
)
Fedina
TY
Lewis
SM
Female mate choice across mating stages and between sequential mates in flour beetles
J Evol Biol
 , 
2007
, vol. 
20
 (pg. 
2138
-
2143
)
Firman
RC
Simmons
LW
Polyandry facilitates postcopulatory inbreeding avoidance in house mice
Evolution
 , 
2008
, vol. 
62
 (pg. 
603
-
611
)
Foellmer
MW
Broken genitals function as mating plugs and affect sex ratios in the orb-web spider Argiope aurantia
Evol Ecol Res
 , 
2008
, vol. 
10
 (pg. 
449
-
462
)
Foellmer
MW
Fairbairn
DJ
Spontaneous male death during copulation in an orb-weaving spider
Proc R Soc Lond Ser B Biol Sci
 , 
2003
, vol. 
270
 
2 Suppl
(pg. 
183
-
185
)
Foellmer
MW
Fairbairn
DJ
Males under attack: sexual cannibalism and its consequences for male morphology and behaviour in an orb-weaving spider
Evol Ecol Res
 , 
2004
, vol. 
6
 (pg. 
163
-
181
)
Gabor
CR
Halliday
TR
Sequential mate choice by multiply mating smooth newts: females become more choosy
Behav Ecol
 , 
1997
, vol. 
8
 (pg. 
162
-
166
)
Gaskett
AC
Herberstein
ME
Downes
BJ
Elgar
MA
Changes in male mate choice in a sexually cannibalistic orb-web spider (Araneae: Araneidae)
Behaviour
 , 
2004
, vol. 
141
 (pg. 
1197
-
1210
)
Haase
M
Baur
B
Variation in spermathecal morphology and storage of spermatozoa in the simultaneously hermaphroditic land snail Arianta arbustorum (Gastropoda, Pulmonata, Stylommatophora)
Invertebr Reprod Dev
 , 
1995
, vol. 
28
 (pg. 
33
-
41
)
Halliday
TR
Bateson
P
The study of mate choice
Mate choice
 , 
1983
Cambridge (UK)
Cambridge Press
(pg. 
3
-
32
)
Hellriegel
B
Ward
PI
Complex female reproductive tract morphology: its possible use in postcopulatory female choice
J Theor Biol
 , 
1998
, vol. 
190
 (pg. 
179
-
186
)
Herberstein
ME
Gaskett
AC
Schneider
JM
Vella
NGF
Elgar
MA
Limits to male copulation frequency: sexual cannibalism and sterility in St Andrew's cross spiders (Araneae, Araneidae)
Ethology
 , 
2005
, vol. 
111
 (pg. 
1050
-
1061
)
Herberstein
ME
Schneider
JM
Elgar
MA
Costs of courtship and mating in a sexually cannibalistic orb-web spider: female mating strategies and their consequences for males
Behav Ecol Sociobiol
 , 
2002
, vol. 
51
 (pg. 
440
-
446
)
Holman
L
Snook
RR
Spermicide, cryptic female choice and the evolution of sperm form and function
J Evol Biol
 , 
2006
, vol. 
19
 (pg. 
1660
-
1670
)
Hurst
GDD
Hurst
LD
Johnstone
R
Intranuclear conflict and its role in evolution
Trends Ecol Evol
 , 
1992
, vol. 
7
 (pg. 
373
-
378
)
Jaenike
J
Sex-ratio meiotic drive in the Drosophila quinaria group
Am Nat
 , 
1996
, vol. 
148
 (pg. 
237
-
254
)
Jennions
MD
Hunt
J
Graham
R
Brooks
R
No evidence for inbreeding avoidance through postcopulatory mechanisms in the black field cricket, Teleogryllus commodus
Evolution
 , 
2004
, vol. 
58
 (pg. 
2472
-
2477
)
Jennions
MD
Petrie
M
Why do females mate multiply? A review of the genetic benefits
Biol Rev Camb Philos Soc
 , 
2000
, vol. 
75
 (pg. 
21
-
64
)
Kempenaers
B
Mate choice and genetic quality: a review of the heterozygosity theory
In: Advances in the Study of Behavior, Vol. 37
 , 
2007
San Diego (CA)
Elsevier Academic Press Inc
(pg. 
189
-
278
)
Knell
RJ
Webberley
KM
Sexually transmitted diseases of insects: distribution, evolution, ecology and host behaviour
Biol Rev Camb Philos Soc
 , 
2004
, vol. 
79
 (pg. 
557
-
581
)
Kokko
H
Brooks
R
McNamara
JM
Houston
AI
The sexual selection continuum
Proc R Soc Lond Ser B Biol Sci
 , 
2002
, vol. 
269
 (pg. 
1331
-
1340
)
Kokko
H
Ots
I
When not to avoid inbreeding?
Evolution
 , 
2006
, vol. 
60
 (pg. 
467
-
475
)
Mesnick
SL
Leboeuf
BJ
Sexual-behavior of male northern elephant seals: Female response to potentially injurious encounters
Behaviour
 , 
1991
, vol. 
117
 (pg. 
262
-
280
)
Michalik
P
Reiher
W
Suhm-Tintelnot
M
Coyle
F
Alberti
G
The female genital system of the folding-trapdoor spider Antrodiaetus unicolor (Hentz, 1842) (Antrodiaetidae, Araneae): An ultrastructural study of form and function with notes on reproductive biology of spiders
J Morphol
 , 
2005
, vol. 
263
 (pg. 
284
-
309
)
Moller
AP
Thornhill
R
Male parental care, differential parental investment by females and sexual selection
Anim Behav
 , 
1998
, vol. 
55
 (pg. 
1507
-
1515
)
Morishita
R
Ferreira
SA
Santiago
A
Faraco
CD
Studies on oogenesis and oviposition in the brown spider Loxosceles intermedia (Araneae: Sicariidae)
Anat Rec Part A Discov Mol Cell Evol Biol
 , 
2003
, vol. 
273A
 (pg. 
575
-
582
)
Nessler
SH
Uhl
G
Schneider
JM
Genital damage in the orb-web spider Argiope bruennichi (Araneae: Araneidae) increases paternity success
Behav Ecol
 , 
2007
, vol. 
18
 (pg. 
174
-
181
)
Pallas
PS
Spicilegia zoologica. Tomus 1. Continens quadrupedium, avium, amphibiorum, piscium, insectorum, molluscorum aliorumque marinorum fasciculos decem
Berolini
 , 
1772
, vol. 
1
 
9
(pg. 
44
-
50
)
Parker
GA
Sperm competition and its evolutionary consequences in the insects
Biol Rev
 , 
1970
, vol. 
45
 (pg. 
525
-
567
)
Parker
GA
Birkhead
TR
Moller
AP
Sperm competition and the evolution of ejaculates: towards a theory base
Sperm competition and sexual selection
 , 
1998
San Diego (CA)
Academic Press
(pg. 
3
-
54
)
Pitcher
TE
Neff
BD
Rodd
FH
Rowe
L
Multiple mating and sequential mate choice in guppies: females trade up
Proc R Soc Lond Ser B Biol Sci
 , 
2003
, vol. 
270
 (pg. 
1623
-
1629
)
Ridley
M
Mating frequency and fecundity in insects
Biol Rev Camb Philos Soc
 , 
1988
, vol. 
63
 (pg. 
509
-
549
)
Rovner
JS
Copulation and sperm induction by normal and palpless male linyphiid spiders
Science
 , 
1967
, vol. 
157
 pg. 
835
 
Schneider
JM
Elgar
MA
Sexual cannibalism and sperm competition in the golden orb-web spider Nephila plumipes (Araneoidea): female and male perspectives
Behav Ecol
 , 
2001
, vol. 
12
 (pg. 
547
-
552
)
Schneider
JM
Gilberg
S
Fromhage
L
Uhl
G
Sexual conflict over copulation duration in a cannibalistic spider
Anim Behav
 , 
2006
, vol. 
71
 (pg. 
781
-
788
)
Simmons
LW
Beveridge
M
Wedell
N
Ttregenza
T
Postcopulatory inbreeding avoidance by female crickets only revealed by molecular markers
Mol Ecol
 , 
2006
, vol. 
15
 (pg. 
3817
-
3824
)
Simmons
LW
Siva-Jothy
MT
Birkhead
TR
Moller
AP
Sperm competition in insects: mechanisms and the potential for selection
Sperm competition and sexual selection
 , 
1998
London
Academic Press
(pg. 
341
-
432
)
Snook
RR
Hosken
DJ
Sperm death and dumping in Drosophila
Nature
 , 
2004
, vol. 
428
 (pg. 
939
-
941
)
Snow
LSE
Andrade
MCB
Pattern of sperm transfer in redback spiders: implications for sperm competition and male sacrifice
Behav Ecol
 , 
2004
, vol. 
15
 (pg. 
785
-
792
)
Thomson
MS
Beeman
RW
Assisted suicide of a selfish gene
J Hered
 , 
1999
, vol. 
90
 (pg. 
191
-
194
)
Thornhill
R
Alcock
J
The evolution of insect mating systems
 , 
1983
Cambridge (MA)
Harvard University Press
Thrall
PH
Antonovics
J
Dobson
AP
Sexually transmitted diseases in polygynous mating systems: prevalence and impact on reproductive success
Proc R Soc Lond Ser B Biol Sci
 , 
2000
, vol. 
267
 (pg. 
1555
-
1563
)
Tregenza
T
Wedell
N
Genetic compatibility, mate choice and patterns of parentage: invited review
Mol Ecol
 , 
2000
, vol. 
9
 (pg. 
1013
-
1027
)
Tregenza
T
Wedell
N
Polyandrous females avoid costs of inbreeding
Nature
 , 
2002
, vol. 
415
 (pg. 
71
-
73
)
Vreys
C
Steffanie
N
Gevaerts
H
Digestion of spermatophore contents in the female genital system of the hermaphroditic flatworm Dugesia gonocephala (Tricladida, Paludicola)
Invertebr Biol
 , 
1997
, vol. 
116
 (pg. 
286
-
293
)
Walter
A
Bliss
P
Moritz
RFA
The wasp spider Argiope bruennichi (Arachnida, Araneidae): ballooning is not an obligate life history phase
J Arachnol
 , 
2005
, vol. 
33
 (pg. 
516
-
522
)
Zeh
DW
Smith
RL
Paternal investment by terrestrial arthropods
Am Zool
 , 
1985
, vol. 
25
 (pg. 
785
-
805
)
Zeh
JA
Zeh
DW
The evolution of polyandry I: intragenomic conflict and genetic incompatibility
Proc R Soc Lond Ser B Biol Sci
 , 
1996
, vol. 
263
 (pg. 
1711
-
1717
)
Zeh
JA
Zeh
DW
The evolution of polyandry II: post-copulatory defences against genetic incompatibility
Proc R Soc Lond Ser B Biol Sci
 , 
1997
, vol. 
264
 (pg. 
69
-
75
)
Zeh
JA
Zeh
DW
Reproductive mode and the genetic benefits of polyandry
Anim Behav
 , 
2001
, vol. 
61
 (pg. 
1051
-
1063
)
Zinner
D
Deschner
T
Sexual swellings in female hamadryas baboons after male take-overs: “deceptive” swellings as a possible female counter-strategy against infanticide
Am J Primatol
 , 
2000
, vol. 
52
 (pg. 
157
-
168
)