Abstract Males are often expected to benefit from mating with multiple females; however, in species where females are highly cannibalistic, achieving multiple matings may be a difficult task. When males employ strategies to avoid sexual cannibalism, it is presumed that there are benefits associated with survival—e.g. increased fitness associated with more mating opportunities. In the nursery web spider (Pisaurina mira), all males attempt to avoid sexual cannibalism by wrapping female’s legs in silk prior to copulation. If males are unsuccessful, however, there are fitness benefits obtained by both sexes that are associated with their consumption—heavier and longer-lived offspring. Regardless, we hypothesize that P. mira males can achieve higher fitness by avoiding sexual cannibalism. Specifically, we predict that males can and will mate with multiple females and that mating with multiple partners benefits males. To test these predictions, we conducted 1) laboratory assays to determine if males will mate with multiple females and 2) field assays to determine natural sex ratios, density, and female and male movement patterns. Finally, 3) we used our field data to construct a mathematical model that predicts natural male encounter rates with potential mates. We found that male P. mira will mate with multiple females and that increased mating numbers leads to increased offspring production. Our model suggests that under natural conditions, males have the opportunity to mate with multiple females. Overall, our findings strongly suggest that P. mira males are likely to benefit from avoiding sexual cannibalism through increased mating opportunities. INTRODUCTION Given that male reproductive success is hypothesized to depend on a male’s ability to acquire mates, transfer sperm, and successfully fertilize females’ eggs (Bateman 1948; Trivers 1972; Parker 1984; Andersson 1994), mating systems that involve males mating with multiple partners (i.e. polygyny) are predicted and often observed across many sexually reproducing animals. Thus, adaptations that allow males to not only secure a mating and increase fertilization success, but also that allow males to successfully mate with additional females, should be under positive selection. Acquiring multiple mating partners can be an especially difficult task for males of species in which females are highly cannibalistic. Sexual cannibalism can be incredibly beneficial for females. Eating a male, for example, can stop unwanted advances from a male (Eberhard 1996; Prenter et al. 2006), can provide nutrients to developing offspring (Young et al. 1988; Barry et al. 2008; Rabaneda-Bueno et al. 2008; Schwartz et al. 2016), or can function to control copulation duration and ultimately fertilization success (Elgar et al. 2000). Some, but not all, of these benefits of sexual cannibalism may be shared with males—e.g. through an increase in offspring quantity and quality (Schwartz et al. 2016). Regardless of any potential fitness benefits associated with one particular mating, however, cannibalism necessarily eliminates a male’s potential to secure future mating opportunities (Elgar and Crespi 1992; Elgar and Schneider 2004). Thus, although a small handful of species express obligate terminal investment strategies and appear to encourage female cannibalism (Foellmer and Fairbairn 2003; Schwartz et al. 2016), males of many species attempt to avoid being cannibalized through a variety of strategies, such as mating with sub-adult females (Biaggio et al. 2016) or feeding females (Fromhage and Schneider 2004), offering nuptial gifts (Toft and Albo 2016), and sedating females with pheromones (Becker et al. 2005), among others. In such species that attempt to avoid cannibalism, we hypothesize that males have the opportunity to increase their fitness through mating with additional female partners. Male nursery web spiders, Pisaurina mira (Walckenaer, 1837), employ a unique mating behavior wherein they obligately restrain females and wrap them with silk prior to and during copulation (Bruce and Carico 1988; Anderson and Hebets 2016). Previous research has demonstrated that this silk wrapping of females allows males to transfer more sperm and increase fertilization success (Anderson and Hebets 2017). Importantly, silk wrapping also reduces rates of postcopulatory sexual cannibalism of the male by the female (Anderson and Hebets 2016). This reduction in sexual cannibalism rates comes at a cost to females, since they have been shown to benefit from consuming their male mating partner by producing heavier and longer-lived offspring (Anderson and Hebets, in review). The fertilizing male necessarily shares this potential fitness benefit, yet males nonetheless attempt to avoid being cannibalized by wrapping females in silk. Given that males have evolved a strategy that facilitates their avoidance of sexual cannibalism, we hypothesize that male P. mira have opportunities for, and engage in, matings with multiple females and that there are fitness benefits associated with mating with multiple partners. In order to test this hypothesis, we conducted laboratory assays to 1) determine whether males would mate with multiple females and 2) to assess the fitness consequences of mating with multiple females, in terms of offspring number. Next, in order to next determine if multiple mating opportunities exist in nature, we conducted field surveys to collect natural history and demographic data on P. mira (e.g. density, sex ratios, and female and male movement patterns). Finally, we used our field data to create a mathematical model that explored female–male encounter rates and thus, the potential for males to mate with multiple females in our natural population. METHODS Male multiple mating and associated fitness Species collection and maintenance We collected immature female and male P. mira at night from Wilderness Park, Lancaster County, Lincoln, NE USA from 4 April 2017 to 17 April 2017. We did not collect in or near areas where our density plots were set-up (see “population density” methods). Collected individuals were transported back to the University of Nebraska-Lincoln where they were individually housed in 87.3 mm × 87.3 mm × 112.7 mm clear plastic containers (763C, AMAC Plastics, Petaluma, CA). We covered the outside of each container with opaque tape to maintain visual isolation between individuals. Spiders were maintained under a 12:12 h light:dark cycle, fed four ¼” crickets per week (Ghann’s cricket farm, GA, USA) and provided water ad libitum. We checked spiders each day for the presence of a molt and to determine the date of sexual maturity. Multiple mating assays Males and females began their first mating trials 15 ± 2 days following their maturation. To determine if males would mate with multiple females, and to assess the relationship between polygyny and offspring number, we randomly paired males (n = 60) with a total of 5 different females across a 10-day period in a mating trial to assess copulation success. Specifically, each male was paired with a new virgin female every other day. Mating arenas and procedures were similar to those carried out in previous studies (Anderson and Hebets 2016). During mating trials, we quantified copulation success. If a male did not copulate with the female within 30 min, the trial ended. Given that we were interested in the natural potential among males for obtaining multiple matings, and because all males wrap their female partners in silk, we did not manipulate males in any way. Thus, all males in our experiment engaged in wrapping the females with silk prior to and during mating. Quantifying offspring production Postmating, we maintained females in the laboratory under the same conditions and diet described above. We checked females each day for the production of an egg sac and subsequent offspring emergence. Three days after offspring emergence, we removed offspring and egg sacs from the mother’s cage. We counted the total number of offspring that emerged from each egg sac. Post offspring quantification, we maintained females under the same controlled conditions and repeated this process if the female produced a second egg sac. After females produced their second egg sac, which is typically the maximum number of successful egg sacs produced (Anderson and Hebets 2017), we released both offspring and mothers back into the field. Statistical analyses To determine if males mate with multiple females, we totaled the number of successful matings that each male achieved across the 10-day assay period. To determine the fitness consequences of males mating multiply, we constructed a generalized linear model with a negative binomial error distribution in which the total number of offspring a male fathered was modeled as a function of the number of successful matings obtained. Our choice of a negative binomial distribution was based on the presence of over dispersion in the residuals of an identical model fit using a Poisson error distribution. We confirmed that the GLM with negative binomial distribution was an appropriate model by visualizing the residuals. Statistical analyses were carried out in R version 3.3.2. Population demographics Population density and sex ratios To approximate population numbers and calculate population density, we used mark-recapture protocols similar to those described in Reed et al. (2007). Specifically, we set-up three 10 × 10 meter plots within an established field site in the wooded area of Wildness Park, Lancaster County, Lincoln, NE, USA and systematically surveyed each plot for 2 consecutive nights. Spiders were observed on the surface of dense patches of vegetation (primarily consisting of stinging nettle). Vegetation reached approximately a meter in height by the end of the late season. Pisaurina mira are easy to observe at night given that the tapetum within their eyes reflects light back from a headlamp or flashlight. On the first night (20 April 2017), we marked each observed individual found within each plot with non-toxic paint (Kids Non-Toxic Washable Watercolor Paint Set, Crayola, Pennsylvania) on the back of the prosoma. Each individual was immediately released in the exact location in which they were collected. The following night we collected all individuals within each plot and noted the number of individuals with and without markings. Additionally, we noted the number of juveniles versus adults and the number of females versus males each night. Each of these 3 plots were surveyed using this method across the season (20 April 2017, 4 May 2017, 11 May 2017, and 24 May 2017). Given that the vegetation became high and dense as the season progressed, our survey method resulted in plots becoming patchy as walking trails formed. To account for this, towards the end of the season, we measured density one more time in newly established plots that had not been previously walked through. Specifically, we moved our plots so that each new plot was directly adjacent to a previous plot and carried out one more round of mark-recapture in the 3 new plots (28 May 2017). This resulted in a total of 5 measures of density and sex ratios across our season. Male and female movement patterns To understand basic male and female movement patterns, we observed the distances that mature males (n =10) and females (n = 10) travelled across a 10-min time period. All observations took place between 2100 and 0000 h. After locating an individual, we carefully marked the location on which the individual was sitting (leaf/plant stem) by placing a small marking of non-toxic paint next to the individual. Each individual was then carefully observed using a red light (poorly perceived by many spiders—Foelix 2011) from over 1 m away for a total of 10 min. Following the completion of the 10-min observation period, the distance from the spider’s starting location to the ending location was measured. Given that the spiders did not necessarily move in a straight line, this start-to-finish measurement represents the minimum amount of distance travelled. Statistical analyses We used the Lincoln-Peterson method to calculate population size, and then used the size of our plots to then calculate density within each plot. To determine whether the likelihood of moving differed between the sexes, we constructed a binomial logistic regression predicting the odds of moving within a 10-min observation period as a function of sex. We assessed the significance of the model using a likelihood ratio test comparing the model to an intercept-only model. Using a Kruskal–Wallis test, we also tested for differences between the sexes in movement distances for only those individuals that moved within the 10-min period. The Kruskal–Wallis test was chosen because visualization of the residuals from a linear model suggested heterogeneous variances. All analyses were conducted in R version 3.3.2. Mathematical model of mating opportunities To predict the number of mating opportunities that a typical male might acquire under natural field conditions, we created a spatially explicit individual-based model of male and female movements and interactions. We parameterized the model using data on movement patterns and demographics in the field (see II. Male and female movement patterns) as well as the frequency of pre and postcopulatory cannibalism observed over the course of the Multiple mating assays (see Appendix A). Because demographic properties change considerably over the course of the season (see Results), we simulated encounters under early and late season conditions to gain insight into how the potential for male multiple mating varies across the growing season. We also briefly consider a version of the model in which movements occur in 3-dimensions (see Appendix A). All modeling was conducted in NetLogo version 6.0.2 (Wilensky 1999). The general model We began by considering a time period T that is sufficiently small so that age-related mortality is negligible and females do not have time to produce offspring and thus become unavailable for mating because they are caring for young (Anderson A, personal observation). We divided the time period into time steps t, such that 0 ≤ t ≤ T. The size of t depends on the temporal resolution of available data on individual movements. Our choice of T implies a static mating pool, aside from losses due to cannibalism and predation (i.e. there are no individuals reaching sexual maturity during the simulation). We initialized each simulation by creating a grid containing N × N patches in which the size of each patch corresponds to the spatial resolution of the demographic data. For each patch, we determined the number of females, males, and juveniles present by drawing random variables i, j, and k from separate Poisson distributions characterized by the mean density (in individuals/patch) of the corresponding demographic group (see Results, Table 1). We then randomized individuals’ positions within the total area of the grid prior to running each simulation. Table 1 The density (individuals/square meter) and proportion of male, female, and juvenile Pisaurina mira within 3 plots measured across the season Time point (date) Plot Density Proportion males Proportion females Proportion juveniles 1) 20 April (early season) 1 0.605 0.069 0.035 0.897 2 0.867 0.050 0.075 0.875 3 0.880 0.049 0.098 0.852 average 0.784 0.056 0.069 0.875 2) 4 May 1 0.560 0.071 0.143 0.786 2 0.707 0.135 0.058 0.808 3 0.523 0.048 0.071 0.881 average 0.597 0.085 0.091 0.825 3) 11 May 1 — — — — 2 0.389 0.192 0.231 0.577 3 0.480 0.143 0.095 0.762 average 0.435 0.168 0.0163 0.670 4) 24 May 1 — — — — 2 0.187 0.250 0.417 0.333 3 0.220 0.471 0.235 0.294 average 0.204 0.361 0.326 0.314 5) 28 May (late season) 1B 0.240 0.600 0.300 0.100 2B 0.338 0.400 0.520 0.080 3B 0.450 0.476 0.476 0.048 average 0.343 0.492 0.432 0.076 Time point (date) Plot Density Proportion males Proportion females Proportion juveniles 1) 20 April (early season) 1 0.605 0.069 0.035 0.897 2 0.867 0.050 0.075 0.875 3 0.880 0.049 0.098 0.852 average 0.784 0.056 0.069 0.875 2) 4 May 1 0.560 0.071 0.143 0.786 2 0.707 0.135 0.058 0.808 3 0.523 0.048 0.071 0.881 average 0.597 0.085 0.091 0.825 3) 11 May 1 — — — — 2 0.389 0.192 0.231 0.577 3 0.480 0.143 0.095 0.762 average 0.435 0.168 0.0163 0.670 4) 24 May 1 — — — — 2 0.187 0.250 0.417 0.333 3 0.220 0.471 0.235 0.294 average 0.204 0.361 0.326 0.314 5) 28 May (late season) 1B 0.240 0.600 0.300 0.100 2B 0.338 0.400 0.520 0.080 3B 0.450 0.476 0.476 0.048 average 0.343 0.492 0.432 0.076 Plot 1B, 2B, and 3B represent the new plots created adjacent to their previous plots. The proportion of females reflects females that are available for copulation (i.e. that do not have egg sacs or spiderlings). View Large Table 1 The density (individuals/square meter) and proportion of male, female, and juvenile Pisaurina mira within 3 plots measured across the season Time point (date) Plot Density Proportion males Proportion females Proportion juveniles 1) 20 April (early season) 1 0.605 0.069 0.035 0.897 2 0.867 0.050 0.075 0.875 3 0.880 0.049 0.098 0.852 average 0.784 0.056 0.069 0.875 2) 4 May 1 0.560 0.071 0.143 0.786 2 0.707 0.135 0.058 0.808 3 0.523 0.048 0.071 0.881 average 0.597 0.085 0.091 0.825 3) 11 May 1 — — — — 2 0.389 0.192 0.231 0.577 3 0.480 0.143 0.095 0.762 average 0.435 0.168 0.0163 0.670 4) 24 May 1 — — — — 2 0.187 0.250 0.417 0.333 3 0.220 0.471 0.235 0.294 average 0.204 0.361 0.326 0.314 5) 28 May (late season) 1B 0.240 0.600 0.300 0.100 2B 0.338 0.400 0.520 0.080 3B 0.450 0.476 0.476 0.048 average 0.343 0.492 0.432 0.076 Time point (date) Plot Density Proportion males Proportion females Proportion juveniles 1) 20 April (early season) 1 0.605 0.069 0.035 0.897 2 0.867 0.050 0.075 0.875 3 0.880 0.049 0.098 0.852 average 0.784 0.056 0.069 0.875 2) 4 May 1 0.560 0.071 0.143 0.786 2 0.707 0.135 0.058 0.808 3 0.523 0.048 0.071 0.881 average 0.597 0.085 0.091 0.825 3) 11 May 1 — — — — 2 0.389 0.192 0.231 0.577 3 0.480 0.143 0.095 0.762 average 0.435 0.168 0.0163 0.670 4) 24 May 1 — — — — 2 0.187 0.250 0.417 0.333 3 0.220 0.471 0.235 0.294 average 0.204 0.361 0.326 0.314 5) 28 May (late season) 1B 0.240 0.600 0.300 0.100 2B 0.338 0.400 0.520 0.080 3B 0.450 0.476 0.476 0.048 average 0.343 0.492 0.432 0.076 Plot 1B, 2B, and 3B represent the new plots created adjacent to their previous plots. The proportion of females reflects females that are available for copulation (i.e. that do not have egg sacs or spiderlings). View Large In each time step, we considered the following sequence of events. First, an individual moves with probability m. Because movement probabilities may differ between sexes, we let mf and mm represent female and male movement probabilities per t, respectively. If an individual moves, the distance moved d is drawn from a normal distribution characterized by mean μ and standard deviation σ. Female and male movement distances are drawn from separate distributions characterized by sex-specific means (μf, μm) and standard deviations (σf, σm) of movement distances. Moving individuals were assumed to move distance d in a straight line but vary their heading randomly within ±45° of their previous heading between movements (i.e. individuals performed correlated random walks). Upon either completing their movement or staying in place, females check for any potential mates within a “mate detection radius” r. The female then cannibalizes a randomly selected male within radius r with probability cpre, which corresponds to the probability of precopulatory cannibalism. If any males remain within radius r after allowing for precopulatory cannibalism, those males then increment an individual-based index of the total number of mating opportunities each male experiences. Thus, encounters resulting in precopulatory cannibalism were not considered mating opportunities. Following a mating opportunity, the female cannibalizes a randomly selected male within radius r with probability cpost, which corresponds to the probability of postcopulatory cannibalism. Consequently, males that achieve a mating opportunity may still die during the encounter and thereby forego all future mating opportunities. Individuals experience mortality due to predation at the end of each t with probability h, where h = 1 – e(−f) and f is the instantaneous rate of predation per t. Parameterizing the model for Pisaurina mira The parameter estimates used for the spatial simulations of P. mira interactions are shown in Supplementary Table A1. We chose T to represent a 10-day period, as this period is short enough that our simplifying assumption of a static mating pool (save for losses due to cannibalism and predation) is reasonable. This 10-day period was subdivided into time steps t of 10-min length, as this is the finest temporal resolution at which we possess data on P. mira movement in the field. Similarly, we set the size of patches in our world equal to the finest spatial resolution of the population density data, which was recorded at a 1-m2 scale. We considered a total spatial area comprised of a 31 × 31 matrix of 1-m2 patches. The model was initialized using the product of the overall population density and the proportions of mature males, mature females, and juveniles of P. mira in the field (Table 1, Supplementary Table A1). To account for seasonal variation in population density and the proportion of mature males and females, we simulated mating interactions under both Early Season and Late Season demographic parameters (Table 1, Supplementary Table A1). A thorough discussion of the source of each parameter estimate is supplied in Appendix A, along with analyses examining the effect of variation in the values of parameters for our qualitative conclusions. Model analysis To determine the number of mating opportunities acquired by the typical male during a 10-day period, we performed 1000 replicate simulations each for Early Season and Late Season demographic conditions and extracted for each simulation the mean number of mating opportunities per male. We then calculated the mean and standard deviation of the mean number of mating opportunities per male across our 1000 replicates for each of the 2 demographic scenarios. To investigate the possibility that the mean number of mating opportunities was inflated by relatively few males with exceptionally high numbers of opportunities, we also calculated the median number of mating opportunities, the proportion of males achieving ≥2 mating opportunities over the 10-day period, and the coefficient of variation of the number of mating opportunities and averaged each over the 1000 simulations for each demographic scenario. RESULTS Male multiple mating and associated fitness Our laboratory assays confirmed that male P. mira will mate with multiple females, although the number of successful matings that males obtained varied (Figure 1). Of the 60 males run in trials, 7 never mated, and a total of 11 males were cannibalized either before or after copulation during one of their mating encounters. Of the males that were not cannibalized and thus had the potential to mate up to 5 times, 8 mated all 5 times. Figure 1 View largeDownload slide (a) The total number of male Pisaurina mira that either mated or did not mate within the first through the fifth mating opportunity with a virgin female. Decreases observed in the total number of males across mating virgin females is due to males being cannibalized either prior or after copulation. (b) The total number of males that achieved 0 through 5 matings with different virgin females. Figure 1 View largeDownload slide (a) The total number of male Pisaurina mira that either mated or did not mate within the first through the fifth mating opportunity with a virgin female. Decreases observed in the total number of males across mating virgin females is due to males being cannibalized either prior or after copulation. (b) The total number of males that achieved 0 through 5 matings with different virgin females. As the number of matings males obtained increased, the total number of offspring fathered by those males also increased ( χ1,512 = 15.342, P < 0.0001; Figure 2). Not all copulations resulted in offspring production (i.e. 48/164 females that mated did not produce any offspring). To determine if males may become sperm limited with increased number of matings we performed a follow-up analysis of the 14 males that successfully mated within their first 3 mating trials (in a row). This analysis did not show significant decreases in offspring numbers across subsequent matings (Mixed effect model: F2,26 = 2.169, P = 0.135). Figure 2 View largeDownload slide The total number of offspring that Pisaurina mira males fathered when achieving 1 through 5 matings with different virgin females. Figure 2 View largeDownload slide The total number of offspring that Pisaurina mira males fathered when achieving 1 through 5 matings with different virgin females. Population demographics Population density The population density varied slightly across our 3 plots, and decreased as the season progressed (Table 1). There were no measurements of population density in Plot 1 for time points 3 and 4 because only one individual was found on the first night of our measurements. Further, this plot was closer to a hiking trail and received some outside human disturbance. The average sex ratio remained close to 50:50 throughout the season, and by our final time point, the majority of individuals located were mature (>90%). Male and female movement patterns Four out of 10 females moved during a 10-min observation period, whereas 9 out of 10 of males moved during a 10-minute period. Females were significantly less likely than males to move ( χ12 = 5.936, P = 0.015), and among those individuals that moved, females moved shorter distances than males (Males: Median = 0.900 m, IQR = 1.15 m; Females: Median = 0.295 m, IQR = 0.16 m; Kruskal–Wallis χ12 = 4.667, P = 0.031). For the purpose of estimating movement distributions for our mathematical model, we also calculated the mean female and male movement distances (μf = 0.31 and μm = 1.10, respectively) and their corresponding standard deviations (σf = 0.20 and σm = 0.83). Mathematical model of mating opportunities Our model predicts that during the Early Season (i.e. when population density is greatest but most individuals are still immature), males should have the opportunity to mate with 7.05 ± 1.10 (mean ± standard deviation) females over a 10-day period (Table 2). In the Late Season (i.e. when population density is lower, but most individuals have matured), males should have the opportunity to mate with 14.21 ± 1.03 (mean ± standard deviation) females during a 10-day period (Table 2). Thus, the effects of decreasing population density as the season progresses appear to be outweighed by the increased proportion of mature individuals. Table 2 Metrics of male multiple mating potential taken from simulations of interactions between male and female Pisaurina mira occurring under early season and late season demographic conditions Phase of growing season Mean females encountered per male Median females encountered per male Proportion encountering ≥2 females CV of females encountered per male Early 7.05 ± 1.10 6.81 ± 1.26 0.90 ± 0.05 0.60 ± 0.07 Late 14.21 ± 1.03 13.42 ± 1.40 0.92 ± 0.02 0.68 ± 0.04 Phase of growing season Mean females encountered per male Median females encountered per male Proportion encountering ≥2 females CV of females encountered per male Early 7.05 ± 1.10 6.81 ± 1.26 0.90 ± 0.05 0.60 ± 0.07 Late 14.21 ± 1.03 13.42 ± 1.40 0.92 ± 0.02 0.68 ± 0.04 Each cell contains the mean ± standard deviation of the metric across 1000 replicate simulations. View Large Table 2 Metrics of male multiple mating potential taken from simulations of interactions between male and female Pisaurina mira occurring under early season and late season demographic conditions Phase of growing season Mean females encountered per male Median females encountered per male Proportion encountering ≥2 females CV of females encountered per male Early 7.05 ± 1.10 6.81 ± 1.26 0.90 ± 0.05 0.60 ± 0.07 Late 14.21 ± 1.03 13.42 ± 1.40 0.92 ± 0.02 0.68 ± 0.04 Phase of growing season Mean females encountered per male Median females encountered per male Proportion encountering ≥2 females CV of females encountered per male Early 7.05 ± 1.10 6.81 ± 1.26 0.90 ± 0.05 0.60 ± 0.07 Late 14.21 ± 1.03 13.42 ± 1.40 0.92 ± 0.02 0.68 ± 0.04 Each cell contains the mean ± standard deviation of the metric across 1000 replicate simulations. View Large Across the 1000 simulations of each demographic scenario, the median number of mating opportunities per male was 6.81 ± 1.26 (mean ± standard deviation) and 13.42 ± 1.40 (mean ± standard deviation) in the early season and late season, respectively (Table 2). The potential for widespread multiple mating is further corroborated by the proportion of males with opportunities to mate with ≥ 2 females, which was 0.90 ± 0.05 (mean ± standard deviation) and 0.92 ± 0.02 (mean ± standard deviation) in the early and late season, respectively (Table 2). Additionally, a low coefficient of variation in the mean number of opportunities among males suggests that the total number of mating opportunities was spread relatively evenly among males within a simulation (Early Season: CV = 0.60 ± 0.07 (mean ± standard deviation), Late Season: CV = 0.68 ± 0.04 (mean ± standard deviation)) (Table 2). These qualitative conclusions hold even if we consider individuals that are randomly distributed in and must move throughout 3-dimensional habitat (see Appendix A). DISCUSSION Unmanipulated male P. mira can, and likely do, mate with multiple females and such multiple mating corresponds to increased offspring numbers for males. Under controlled conditions in the laboratory, we demonstrate that males can and will mate with up to 5 consecutive females. We also show a positive relationship between mating number and offspring number. We used demographic data collected from field surveys to construct a mathematical model that confirms that males in our natural population are likely mating multiply given that male mating opportunities are sufficiently high. Here we will discuss both the findings of our laboratory assays and our model in more detail. Although all males (except males that were cannibalized) were given the opportunity to mate with 5 virgin females over a 10-day period, the number of matings that males achieved ranged from 0 to 5. The majority of males achieved 3 or more matings, and few achieved 0 or 1. Only 8 out of 60 males achieved the full 5 matings. Failed matings were due to a variety of factors including: males not attempting to mate, females resisting male advances and precopulatory cannibalism of the male. Though not all males mated with all 5 females, our data nonetheless demonstrate that unmanipulated male P. mira can and will mate with multiple females when given the opportunity. Such multiple mating can occur within a relatively small time frame as well (relative to the mating season)—i.e. within 72 h. Given that P. mira males do not deplete the sperm store in their pedipalps (i.e. their external sperm storage organ) following a single mating (Anderson and Hebets 2017), they may be capable of multiple mating within an even smaller time window—e.g. a single night. Indeed, even if they were sperm depleted, it has been shown in other spiders (i.e. jumping spiders) that males may be capable of taking up more sperm into their pedipalps fairly quickly after a mating (McGinley et al. 2015). The explicit role of silk wrapping in male multiple mating was beyond the scope of this study as all males wrapped females, but we expect that wrapping females with silk facilitates multiple mating in male P. mira given that silk wrapping was previously shown to allow males to escape postcopulatory cannibalism (Anderson and Hebets 2016). As expected, males that mated with additional females produced more offspring overall. This result is not surprising given that we have no evidence that males become sperm limited within the mating time frame tested within this experiment, other than the fact that more copulations resulted in no offspring than in previous years (Anderson and Hebets 2017; Anderson and Hebets, in review). Given that 7 of the 48 females that did not produce offspring were a male’s first mating partner, and that mating multiple times in a row did not show significant decreases in offspring numbers across subsequent matings, we assume this difference between years is not due to sperm depletion. Within this experiment, males were only paired with virgin females, which likely results in males achieving greater offspring numbers than they would if they mated with previously mated females. We are still in the process of working out sperm priority patterns to determine the paternity share a second or even third male would be able to achieve with a previously mated female. However, given that males will readily attempt to mate and force copulations by tackling and mounting with already mated females in the laboratory (whom are often aggressive and cannibalistic) (Anderson, unpublished data), we propose that males are likely obtaining some benefits from mating with previously mated females. Further, females in the laboratory remate (often by force) ~59% of the time (Anderson, unpublished data), therefore, even if remating rates are lower in the field, it is likely that it is still occurring to some extent. While males of some spider species have mechanisms that allow them to differentiate between mated and non-mated females (e.g. Rypstra et al. 2003; Roberts and Uetz 2005; Stoltz et al. 2007), it is not known whether P. mira males are able to do so. It is possible that the high likelihood of males attempting to mate with non-virgin females in P. mira may simply reflect an inability to differentiate them. Overall, sperm priority patterns and female remating rates within our natural population of P. mira need to be assessed, as the benefits acquired from mating with a non-virgin may be in fact quite minimal. Although our experimental data revealed that males would readily mate multiply in the laboratory, we also wanted to determine that the potential for males to mate multiply exists within our natural population. Utilizing density measures, sex ratios, and sex-specific movement data, we were able to model the average number of mating opportunities males acquire within a 10-day period. We wanted to specifically look at a 10-day period not only to match our laboratory data, but also because we believe this is likely a very conservative estimate of how much time males have for mate searching across their reproductive life span. From this model, we found that within a 10-day period males have the opportunity to mate with approximately 7 females in the early season and approximately 14 females in the late season. Moreover, we found that the opportunity for males to mate with multiple partners likely exists across a wide range of parameter values, suggesting that male P. mira should generally experience positive selection on cannibalism avoidance despite spatial or temporal variation in various aspects of their natural history. Even when considering a highly conservative version of the model in which individuals are dispersed randomly throughout the volume of their habitat (i.e. modeling the habitat in 3 dimensions), the majority of males have the opportunity to mate multiply. Overall, from our model it appears that males have numerous opportunities to seek additional mating partners, making polygyny a likely scenario for our P. mira population. Unfortunately, we are still limited in our knowledge of how frequently females remate in the field, which can strongly influence how many successful matings a male can achieve. From this and previous studies on copulatory silk wrapping in P. mira, it appears that males benefit from copulatory silk wrapping in a number of ways: 1) to increase offspring numbers with individual females by increasing sperm transfer (Anderson and Hebets 2017), and 2) by allowing males to survive matings (Anderson and Hebets, 2016) to obtain increased offspring numbers with additional females (this paper). We suspect that silk wrapping additionally aids males in obtaining matings with previously mated females, but this was not tested in this study. Within the mating system of P. mira, there appears to be a conflict between the reproductive interests of males and females, where males benefit from trying to increase their own offspring production, while females attempt to benefit from cannibalism, which leads to heavier and longer-lived offspring (Anderson and Hebets, in review). Although it appears that the male reproductive strategy often dictates a mating encounter, future studies should consider additional costs of copulatory silk wrapping to females, as well as the potential for cryptic female choice, which may allow females to maintain control over fertilization of her eggs. We thank Lincoln Parks and Recreations for park access and Austin Brooks, Justin Buchanan, and Rowan McGinley for help with spider collections. We also thank members of the Hebets, Shizuka, and Wagner labs for providing feedback on this project. Data accessibility: Analyses reported in this article can be reproduced using the data provided by Anderson et al (2018). REFERENCES Anderson AG , Hebets EA . 2016 . Benefits of size dimorphism and copulatory silk wrapping in the sexually cannibalistic nursery web spider, Pisaurina mira . Biol Lett . 12 : 20150957 . Google Scholar CrossRef Search ADS PubMed Anderson AG , Hebets EA . 2017 . Increased insertion number leads to increased sperm transfer and fertilization success in a nursery web spider . Anim Behav . 132 : 121 – 127 . Google Scholar CrossRef Search ADS Anderson AG , Hebets EA , Bickner BM , Watts JC . 2018 . Data from: males mate with multiple females to increase offspring numbers in a nursery web spider . Dryad Digital Repository . http://dx.doi.org/10.5061/dryad.db0f908. Andersson MB . 1994 . Sexual selection . Princeton (NJ) : Princetion University Press . Barry KL , Holwell GI , Herberstein ME . 2008 . Female praying mantids use sexual cannibalism as a foraging strategy to increase fecundity . Behav Ecol . 19 : 710 – 715 . Google Scholar CrossRef Search ADS Bateman AJ . 1948 . Intra-sexual selection in Drosophila . Heredity (Edinb) . 2 : 349 – 368 . Google Scholar CrossRef Search ADS PubMed Becker E , Riechert S , Singer F . 2005 . Male induction of female quiescence/catalepsis during courtship in the spider, Agelenopsis aperta . Behaviour . 142 : 57 – 70 . Google Scholar CrossRef Search ADS Biaggio MD , Sandomirsky I , Lubin Y , Harari AR , Andrade MCB . 2016 . Copulation with immature females increases male fitness in cannibalistic widow spiders . Biol Lett . 12 : 20160516 . Google Scholar CrossRef Search ADS PubMed Bruce JA , Carico JE . 1988 . Silk use during mating in Pisaurina mira (Walckenaer) (araneae, Pisauridae) . Am Arachnol Soc. 16 : 1 – 4 . Eberhard WG . 1996 . Female control: sexual selection by cryptic female choice . Princeton (NJ) : Princeton University Press . Elgar MA, Crespi BJ. 1992. Ecology and evolution of cannibalism. In: Elgar MA, Crespi BJ, editors. Cannibalism: ecology and evolution among diverse taxa p. 1–12. Oxford University Press. Elgar MA, Schneider JM. 2004. Evolutionary significance of sexual cannibalism. Adv Study Behav. 34:135–163. Elgar MA , Schneider JM , Herberstein ME . 2000 . Female control of paternity in the sexually cannibalistic spider Argiope keyserlingi . Proc R Soc Lond B Biol Sci . 267:2439–2443 . Foelix R . 2011 . Biology of spiders . Cambridge : Harvard University Press . Foellmer MW , Fairbairn DJ . 2003 . Spontaneous male death during copulation in an orb-weaving spider . Proc R Soc Lond B Biol Sci . 270:183–185 . Fromhage L , Schneider JM . 2004 . Safer sex with feeding females: sexual conflict in a cannibalistic spider . Behav Ecol . 16 : 377 – 382 . Google Scholar CrossRef Search ADS McGinley RH , Mendez V , Taylor PW . 2015 . Natural history and display behaviour of Servaea incana, a common and widespread Australian jumping spider (Araneae : Salticidae) . Aust J Zool . 63 : 300 . Google Scholar CrossRef Search ADS Parker GA . 1984 . Sperm competition and the evolution of animal mating strategies . In: Smith RL , editor. Sperm competition and the evolution of animal mating systems . Orlando (FL) : Academic Press . p. 1 – 60 . Google Scholar CrossRef Search ADS Prenter J , MacNeil C , Elwood RW . 2006 . Sexual cannibalism and mate choice . Anim Behav . 71 : 481 – 490 . Google Scholar CrossRef Search ADS Rabaneda-Bueno R , Rodríguez-Gironés MÁ , Aguado-de-la-Paz S , Fernández-Montraveta C , De Mas E , Wise DH , Moya-Laraño J . 2008 . Sexual cannibalism: high incidence in a natural population with benefits to females. Brooks R, editor . PLoS ONE . 3 : e3484 . Google Scholar CrossRef Search ADS PubMed Reed DH , Nicholas AC , Stratton GE . 2007 . Inbreeding levels and prey abundance interact to determine fecundity in natural populations of two species of wolf spider . Community Genet . 8 : 1061 – 1071 . Roberts JA , Uetz GW . 2005 . Information content of female chemical signals in the wolf spider, Schizocosa ocreata: male discrimination of reproductive state and receptivity . Anim Behav . 70 : 217 – 223 . Google Scholar CrossRef Search ADS Rypstra AL , Wieg C , Walker SE , Persons MH . 2003 . Mutual mate assessment in wolf spiders: differences in the cues used by males and females . Ethology . 109 : 315 – 325 . Google Scholar CrossRef Search ADS Schwartz SK , Wagner WE Jr , Hebets EA . 2016 . Males Can Benefit from Sexual Cannibalism Facilitated by Self-Sacrifice . Curr Biol . 26 : 2794 – 2799 . Google Scholar CrossRef Search ADS PubMed Stoltz JA , McNeil JN , Andrade MCB . 2007 . Males assess chemical signals to discriminate just-mated females from virgins in redback spiders . Anim Behav . 74 : 1669 – 1674 . Google Scholar CrossRef Search ADS Toft S , Albo MJ . 2016 . The shield effect: nuptial gifts protect males against pre-copulatory sexual cannibalism . Biol Lett . 12 : 20151082 . Google Scholar CrossRef Search ADS PubMed Trivers R . 1972 . Parental investment and sexual selection . In: Campbell B, editor. Sexual selection and the descent of man . New York : Aldine de Gruyter . p. 136 – 179 . Wilensky U . 1999 . NetLogo. Center for Connected Learning and Computer-Based Modeling . Evanston (IL) : Northwestern University . Available from: http://ccl.northwestern.edu/netlogo/. Young P , Lee KE , Birkhead TR . 1988 . Sexual Cannibalism in the Praying Mantis Hierodula Membranacea . Behaviour . 106 : 112 – 118 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: email@example.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
Behavioral Ecology – Oxford University Press
Published: Apr 10, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera