Mating plugs have been described in many species, and their presence often implies a function in protecting a male’s ejaculate. Yet, explicit functions are not always tested. In this study, we test whether fragments of male genitalia lodged in the female genital opening of the St Andrew’s Cross spider (Argiope keyserlingi) are mating plugs and prevent female remating. Further, we test whether copulation duration, cannibalism, and male or female size affect the lodgement and persistence of these genital fragments. We show that males always break off a genital fragment, which when lodged in the female genital opening, can successfully prevent female remating. However, the lodgement of a genital fragment is not always successful and it may not persist for a prolonged period. Whether a genital fragment is successfully retained is influenced by female control over copulation duration. We have previously shown that females can terminate copulation duration by attacking the male, which may or may not lead to cannibalism. If females terminate copulations early, genital fragments are either not lodged or do not persist. Male size can offset female control with larger males lodging more persistent fragments. Contrary to predictions, sexual cannibalism was not related to how long the fragment persisted within the female. We demonstrate the existence of mating plugs in St Andrew’s Cross spiders and document considerable variation in the formation and persistence of mating plugs that is likely to reflect male and female conflict over mate plugging.
Mating (also sperm, genital, or copulatory) plugs are solid substances, which are left within the female genital tract after mating. These plugs have been described in many different taxa including nematodes (Hodgkin and Doniach 1997), insects (Alcock 1994; Chapman 2001; Simmons 2001), spiders (Eberhard and Huber 2010; Uhl et al. 2010), reptiles (Shine et al. 2000), and mammals (Ramm et al. 2005). The plug material itself may consist of coagulated sperm (Dixson and Anderson 2002), accessory gland products (Duvoisin et al. 1999; Chapman 2001; Wedell 2005), parts of male genitalia (Miller 2007; Foellmer 2008; Uhl et al. 2010), or even the entire dead male (Foellmer and Fairbairn 2003). The term mating plug assumes a function in paternity protection, which often remains untested. An alternative terminology is “mating sign” (Uhl and Busch 2009) implying a form ofcommunication with other males.
Mating plugs are frequently evoked within the context of sexual conflict, where male mating plugs influence female remating frequency (Hosken et al. 2009). Imposing monandry may increase male fitness at a cost to the fitness of the female. Mating plugs can reduce female remating frequency in several ways. Most obviously, they may block access to the female genital opening and thus physically prevent other males from mating with the female. For instance, male Drosophila hibisci deposit a large, firm gelatinous plug that fills up the entire female uterus and remains there for several days (Polak et al. 1998). Particularly, large and externally visible plugs may additionally signal female mating status and thus deter subsequent males from even approaching the female (Orr and Rutowski 1991). Finally, the mating plug may contain chemicals that influence female receptivity. For example, linoleic acid found in bumble bee mating plugs reduces the rate of female remating, with no evidence that males can identify previously mated females (Baer et al. 2001). The above examples illustrate convincingly that mating plugs can impose monandry under a sexual conflict scenario, but alternative functions for mating plugs are less frequently tested. For example, mating plugs might prevent sperm leakage or genital infections, which may be in the interest of the female (Polak et al. 1998; Eberhard and Huber 2010). Alternatively, in some spiders, fragments of male genitalia are lodged in the female genital opening but do not prevent polyandry (Schneider et al. 2001). Instead, genital damage is the result of conflict over copulation duration, as the female forcefully removes the male, thereby breaking the genitalia rather than a mechanism to protect the male’s ejaculate (Schneider et al. 2001).
In some systems, females are actively involved in either producing the plug or in its persistence. In the tetragnathid spider Leucauge marina, plug formation is most successful when the female co-operates, which is dependent on male courtship performance (Aisenberg and Eberhard 2009). Similarly, in several species of comb-footed spiders (Theridiidae, Theridion), the secretory plug appears to be formed from both male and female secretions, suggesting direct female influence over the protection of male ejaculates (Uhl et al. 2010). Females may also influence the persistence of mating plugs by active removal, such as in squirrels, where the plug is removed by the females almost immediately following mating (Koprowski 1992). Even without obvious female contributions, not all males may successfully transfer a mating plug. In garter snakes, for instance, mating plugs are very effective in reducing female remating, but not all males transfer a mating plug (Shine et al. 2000). A similar case is described for the spider Argiope lobata in which plugging is not obligatory and mostly occurs in concert with sexual cannibalism (Nessler et al. 2009).
This study investigates the occurrence of lodged pieces of broken genitalia in St Andrew’s cross spiders (Argiope keyserlingi). We hypothesize that these fragments function as mating plugs, predicting that their presence prevents other males from inserting into the female copulatory opening. Our second hypothesis is that if these fragments act as plugs, there is a conflict over plug formation between males and females. As the larger females control copulation duration through the onset of an aggressive attack (Elgar et al. 2000), we predict that shorter copulations and larger female size prevent plug formation and persistence. Conversely, larger males are predicted to be more successful at producing persistent plugs as larger males maybe better able to resist being removed by the female during copulation.
Sexual cannibalism by females in A. keyserlingi may further affect whether or not males use a mating plug. Female A. keyserlingi always attempt to cannibalize their mates and are successful in around 40–50% of cases (Herberstein et al. 2005b). Sexual cannibalism is clearly a result of female aggression and eliminates a male’s future reproductive opportunity. Therefore, we hypothesize that if genital fragments are mating plugs, then males have a trade-off between efficient plugs and future mating opportunities. We predict that cannibalized males produce more effective plugs.
Argiope keyserlingi is a large orb-web spider, common in gardens and parks along the east coast of Australia. They preferentially build their orb-webs in Lomandra bushes (Herberstein 2000; Rao et al. 2007). Like most orb-web spiders, A. keyserlingi males have indirect paired sperm transfer appendages (pedipalps) and females have an external epigyne with paired copulatory openings, each leading into a separate sperm storage organ (Herberstein et al. 2011a). Although mating plugs in the form of broken genitalia have been described in congeners (e.g., A. lobata; Nessler et al. 2009), this is the first study to document this phenomenon in A. keyserlingi.
We collected juvenile males and females from west Pymble Park (Sydney, Australia) in December 2003, 2004, 2005, and 2010. The animals were transported to the laboratory and kept in small upturned cups (300mL) with the bottom replaced with mesh. They were fed twice a week with Drosophila sp. and houseflies (Musca sp.) and watered once a day, until they matured in the cups.
When females matured, we transferred them into large (50 × 50 × 10cm) 3D Perspex frames, where they could build a complete orb-web. For copulation trials, we gently placed the male on the edge of the female web from which he proceeded to the central hub where the female resided. He typically engaged in vibratory and tactile courtship such as tugging the web and tapping the female. After a period of courtship, he cut a hole into the web and spun a horizontal mating thread. He vibrated the mating thread, and if the female was soinclined, she moved onto the mating thread, suspended herself at a 45° angle, and exposed her epigyne. The male thenleapt onto the female and inserted one of his 2 pedipalps. Pedipalp insertion is ipsilateral and invariable; males insert their right pedipalp into the right copulatory opening of the epigyne and their left pedipalp into the left copulatory opening of the epigyne. We measured copulation duration with a stopwatch from the start of the insertion until the male broke off from the female. Males are only able to use each pedipalp once (Herberstein et al. 2005b) and never insert both pedipalps at the same time. Females always attempt to cannibalize the male. Virgin males can escape cannibalism, but mated males are always cannibalized (Gaskett et al. 2004).
Rates of pedipalp damage and lodgement of fragments
In order to estimate the rates of pedipalp damage as well as how long broken genital fragments persist in the female, we paired virgin males with virgin females (n = 36). During copulation, we monitored which pedipalp the males used but did not interfere with natural rates of cannibalism. Males (n = 36) were removed after copulation ceased, and immediately inspected under a dissection microscope to record pedipalp damage. All females (n = 36) were immediately inspected for the presence of a genital fragment under a dissection microscope after being anaesthetized with CO2. Following the initial examination, all spiders were returned to their frames, and 24h after copulation, they were euthanized and re-examined under a dissection microscope. In a second set of mating experiments, 33 virgin males were mated to virgin females and examined immediately following the copulation as described above. The females were maintained in the laboratory for another 7 days before being euthanized and re-examined. All females were watered and fed as described above. Some data were lost due to mortality during this period.
We chose 24h as the first time point to check for fragment persistence as females retain their virgin attraction for 24h after copulation and thus the chance of multiple mating is presumably high during the first 24h (Gaskett et al. 2004). After 24h, males can distinguish between a mated and a virgin female. Nonetheless, in the field, mated females still attract males (personal observation) throughout the season, and hence, we picked a second time point (7 days). We did not select a third, even later point in time, as mortality in the laboratory would have reduced our sample size substantially.
How effective are genital fragments as mating plugs?
We conducted this experiment to test if the broken male genitalia are effective mating plugs, preventing other males from inserting their pedipalps into the female copulatory opening. We mated virgin males with virgin females, preventing cannibalism by gently placing a paintbrush against the female’s chelicerae during copulation. Females do not seem to be disturbed by this and attack the paintbrush rather than the male. Nonetheless, we cannot exclude the possibility that the paintbrush affected behavior; therefore, we only used this method to create sufficient numbers of mated males and females. We noted if the male inserted into the left or right copulatory opening of the epigyne. We then assigned mated males to a different already mated female so that in the control group (n = 18) the male inserted his remaining pedipalp into an unused copulatory opening, and in the experimental group (n = 15), the male had to insert his pedipalp into a used copulatory opening. All second matings took place 24h after the female’s first mating. We measured courtship duration using a stopwatch as the time between males entering the female web and the start of copulation. We noted and compared the number of successful copulations between the control and the experimental group. Males were weighed before, and females were weighed after the experiment. We analyzed the data using SPSS version 17.0. All continuous data were normally distributed, and we used parametric 2-tailed t-tests to compare male size between males attempting to mate with either used (experimental group) or unused (control) copulatory opening. We used 2×2 chi-squared tests to compare the frequency of successful matings between males attempting to copulate into a used and an unused copulatory opening.
What factors determine if genital fragments are effective plugs?
We staged laboratory matings between virgin males and females (n = 38) and mated males and virgin females (n = 17) to identify which factors contribute to the lodgement and persistence of genital fragments in A. keyserlingi. Males were weighed prior to and females were weighed after the trial, and we measured copulation duration (stopwatch) and noted whether or not males were cannibalized. After copulation terminated, we examined the female copulatory openings to see whether or not a genital fragment was present. The females were then maintained in the laboratory under standard conditions and rechecked after 7 days. Using this information, we coded the data for 3 mating plug outcomes: 1) no fragment, if we found no fragment at the immediate inspection; 2) initial lodgement but not persistent, if the fragment was initially present, but was lost after 7 days; and 3) persistent lodgement, if the fragment persisted for 7 days. We analyzed the data with JMP (version 22.214.171.124) and R (R Development Core Team 2011, packages nnet and AICcmodavg), using stepwise multinomial logistic regression. Independent variables included in the full model were copulation duration, female weight, male weight, cannibalism, male mating status (virgin or mated), and the interaction between male mating status and male weight. We used corrected Akaike’s Information Criterion for small samples (AICc; Burnham and Anderson 1998) based on AICc weights for model selection. Using this method, models with larger AICc weights (and smaller AICc) are considered more reliable.
Rates of pedipalp damage and lodgement of fragments
Of the 36 males examined on the day of mating, all (100%) broke off part of the pedipalp (the embolus; Figure 1a,c) during copulation, and in all (n = 36), the piece was stuck in the female copulatory opening (Figure 1d). After 24h, we successfully re-examined 34 of those females and 88% (n = 30/34) still retained the genital fragment in the copulatory opening. In the second group (re-examined 7 days after mating), 33 virgin males were mated, and all males broke off the embolus during copulation, but it was only stuck immediately after copulation in the female copulatory opening in 70% of females (n = 23). After 7 days, we successfully re-examined 19 of those females, and most of them still retained the genital fragment in their copulatory opening (n = 15/19 females, 79%).
How effective are genital fragments as mating plugs?
In this experiment, we tested whether genital fragments are effective in preventing subsequent males from copulating with the female. There was no difference in the weight of males mating into an unused copulatory opening (control: mean ± standard error [SE] = 0.0155 ± 0.0017g) and males mating into a used copulatory opening (experimental: mean ± SE = 0.0118±0.0017g; 2-tailed t-test: t25 = −1.04, P = 0.31). However, the rate of successful copulations differed significantly between males inserting into a used versus unused copulatory opening. Specifically, all males (n = 18) in the control group, which inserted into an unused copulatory opening, successfully copulated. When males (n = 15) attempted to insert into a used copulatory opening, only 20% of males successfully copulated (χ22 = 19.3, P < 0.01). In the unsuccessful copulations, attempts lasted for less than a second before males separated from the female and ceased all courtship activities. In many cases, the male left the web.
What factors determine if genital fragmentsare effective plugs?
In total, we mated 38 virgin males and 17 mated males to virgin females. These matings resulted in eight cases of no lodgement of a genital fragment, 29 cases where the genital fragment persisted less than 7 days, and 15 cases where the genital fragment persisted for at least 7 days. Variables included in the final multinomial logistic regression model were copulation duration, male mating status (virgin or mated), male weight, and the interaction between male mating status and male weight. Overall, our model was significant (χ2 = 44.82, df = 8, P < 0.01) explaining around 45% (R2 = 0.44) of the observed variation in fragment lodgement and persistence. Copulation duration (χ2 = 8.74, P = 0.01; Figure 2), male weight (χ2 = 6.79, P = 0.03; Figure 3), and the interaction between male status and male weight (χ2 = 6.66, P = 0.04; Figure 3) contributed significantly to the overall model. Specifically, persistent fragments (at least 7 days) were formed following longer copulations. Mated males that formed less persistent fragments (no fragment or less than 7 days) were smaller than virgin males with less persistent fragments. However, this size difference was not apparent for virgin and mated males that formed persistent (at least 7 days) fragments. Other variables, such as whether or not males were cannibalized, or female weight, were not significant (P > 0.05).
Our study revealed that the genital fragment in A. keyserlingi is indeed a mating plug that prevents normal copulations by subsequent males. We found considerable variation in mating plug formation and persistence; mating plugs were not always formed in the female and varied in their persistence over several days. However, this variation was not due to males providing a plug in the first instance. All males examined broke off a piece of their copulatory organ, but the plug did not always successfully lodge inside the female.
The frequency with which males provide a mating plug varies in different mating systems. When the mating plug is formed by the male’s spermatophore, it is presumably transferred during every mating. However, males may be energetically limited in producing a costly spermatophore following several matings (Wedell 2005). Multiple mating in snakes reduces the size of the sperm plugs (Shine et al. 2000), and in D. hibisci, it prevents plug formation all together (Polak et al. 1998). The transfer of mating plugs involving fragments of genitals can also be quite variable.
Genital damage is common in spiders, earwigs, bees, and some ants (Hosken et al. 2009; Uhl et al. 2010). However, not every mating may result in genital damage and the lodgement of the fragment. In redback spiders, for example, around 80% of males break the tip of their genitalia during mating, most of which (>80%) may then become lodged in the female copulatory duct (Snow et al. 2006). Even in the congener, Argiope bruennichi, genital damage is not obligatory; around 15% of matings did not result in genital damage (Uhl et al. 2007). In both these spider examples, avoiding genital damage is not an adaptation to avoid sterility. Males of all Argiope species die in any case after their second copulation, and evidence is accumulating that males in several species of web-building spiders are not able to recharge their pedipalps because the testes degenerate and cease sperm production after maturation (Snow et al. 2006; Uhl et al. 2007; Michalik et al. 2010; Schneider and Michalik 2011). Whether or not males provide a mating plug may reflect species-specific costs associated with plugging, the threat of sperm competition, and future mating opportunities.
In our study, the consistency with which males damaged their genitals suggests that the resulting mating plugs are effective in blocking the female copulatory opening. Indeed, our mating experiments showed that only 20% of males were able to copulate into an already used copulatory opening. Our data on plug persistence suggest that these 20% correspond to the 12–30% of cases where we observed the transfer of a plug that did not persist for more than 24h. If this assumption is correct, it implies that mating plugs in A. keyserlingi, if they are successfully lodged and they persist, are quite effective in preventing subsequent males from ejaculating sperm into the same copulatory opening, however, we do not have data on actual paternity to confirm that males were actually prevented from ejaculating.
The efficacy of mating plugs varies in spiders and other animals. For example, in some spiders (Masumoto 1993; Schneider and Lesmono 2009) and snakes (Shine et al. 2000), the complete plug is effective in protecting a male’s ejaculate while it is in place. In other systems, such as redback spiders (Snow et al. 2006), Apis bees (Koeniger 1991), or rock lizards (Moreira and Birkhead 2004), subsequent males can access the copulatory opening and transfer sperm, even if a mating plug is present. Male A. keyserlingi could fully monopolize a female if he successfully placed a mating plug in each of her 2 copulatory openings during 2 sequential matings. Curiously, unlike other species in this genus (Nessler et al. 2007), virgin male A. keyserlingi only mate once with a female and, if they are not cannibalized, perform their final copulation with a different female (Herberstein et al. 2005a), allowing a second male to inseminate her unused copulatory opening. Nonetheless, a successful plug is likely to prevent a second male from copulating into the same copulatory opening. Storing the sperm of 2 males in the same storage organ does affect paternity patterns in spiders (Snow et al. 2006; Jones and Elgar 2008). Therefore, it might be that A. keyserlingi males maximize their fitness by protecting their ejaculate within a sperm storage organ with a mating plug in 2 different females.
In our study, not all males managed to deploy a lasting mating plug. Contrary to our predictions, however, we did not detect evidence for males trading off plug efficacy with future mating opportunity: whether or not sexual cannibalism occurred did not influence plug formation and persistence in our experiments. In accordance to our predictions, we found that persistent plugs were mostly formed following longer copulations (Uhl and Busch 2009). Copulation duration is an interesting parameter in A. keyserlingi. It directly affects sperm transfer (Herberstein et al. 2011b), and there is evidence that females control copulation duration (Elgar et al. 2000) by attacking the male during copulation and pulling him off.
Therefore, females seem to control the formation of persistent mating plugs by allowing males to mate for longer, or by terminating copulation early. If females control plug formation, why do they facilitate plugs at all? It may be that these mating plugs prevent sperm mixing, which could aid female cryptic choice by differentially activating sperm in the 2 different spermathecae (Herberstein et al. 2011a). They may also prevent sperm loss (Lachmann 2000; Polak et al. 2001), and therefore be advantageous to both males and females. In some spiders (e.g. Leucauge agyra), the plug is formed solely by female secretions and may even function to entangle males for long enough for females to successfully cannibalize the male (Aisenberg and Barrantes 2011). We found that larger males were more likely to deploy a persistent mating plug than smaller males. This is unlikely to reflect a female preference for larger males, as relatively smaller males are preferred by A. keyserlingi females and fertilize more eggs (Elgar et al. 2000). Rather, larger males may have longer pedipalps that assist in placing the mating plug deeper into the female copulatory opening. In dung beetles, for instance, elongation of male genitalia is associated with the improved deposition of the spermatophore (Takami and Sota 2007). Similarly, in redback spiders, the exact placement of the mating plug affects its efficacy (Snow et al. 2006). However, if the size of the mating plug scales with male size (Masumoto 1993), larger plugs may be more persistent irrespective of where they are placed in the female.
In summary, we found that mating plug formation and persistence in the St Andrew’s Cross spider results from an interaction of female and male traits. However, unlike other systems with mating plugs, the 2 sexes do not seem to co-operate in mating plug formation. Females control copulation duration and thereby the formation and persistence of the plug, but male size can, to some extent, offset female control.
We are grateful to Macquarie University and DAAD (German Academic Exchange) for financial support for this research. We thank Matthias Foellmer, Gill Rosenthal, and an anonymous reviewer for their helpful and constructive comments.