Behavioral Ecology and Sociobiology

, Volume 62, Issue 1, pp 29–35

Daring females, devoted males, and reversed sexual size dimorphism in the sand-dwelling spider Allocosa brasiliensis (Araneae, Lycosidae)


    • Laboratorio de Etología, Ecología y EvoluciónInstituto de Investigaciones Biológicas Clemente Estable
  • Carmen Viera
    • Sección EntomologíaFacultad de Ciencias
  • Fernando G. Costa
    • Laboratorio de Etología, Ecología y EvoluciónInstituto de Investigaciones Biológicas Clemente Estable
Original Paper

DOI: 10.1007/s00265-007-0435-x

Cite this article as:
Aisenberg, A., Viera, C. & Costa, F.G. Behav Ecol Sociobiol (2007) 62: 29. doi:10.1007/s00265-007-0435-x


Sexual selection theory predicts that a higher investment in offspring will turn females into the selective sex, while males will compete for accessing and courting them. However, there are exceptions to the rule. When males present a high reproductive investment, sex roles can reverse from typical patterns, turning males into the choosy sex, while females locate males and initiate courtship. In many spiders, males are smaller than females, wandering in search of sedentary females and maximizing the number of copulations. In the present study, we present findings on the sand-dwelling wolf spider, Allocosa brasiliensis, evidencing a reversal in typical courtship roles reported for the first time in spiders. Males were bigger than females. Females located males and initiated courtship. Copulation always occurred in male burrows and took place mainly in long burrows. Males donated their burrows to the females after copulation, closing the entrance before leaving with female cooperation from inside. Males would provide females with a secure place for ovipositing, being exposed to predation and diminishing their future mating possibilities until constructing a new burrow. The cost of vacating the burrow and losing the refuge in an unpredictable habitat, such as sand dunes, would explain the courtship roles reversal in this spider species. Results turn A. brasiliensis as a promising model for discussing the determinants of sex roles and the pressures that drive their evolution and maintenance.


Wolf spiderSex rolesSexual size dimorphism reversal


Sex roles reflect female and male contributions to gamete production and maintenance, courtship and mating effort, provision of nuptial gifts and/or other resources associated with reproduction, and parental investment (Trivers 1972; Bonduriansky 2001). Precopulatory asymmetries between sexes in gamete investment will promote asymmetries in mating and post mating investment. Anisogamy will determine that, in general, females are choosy, while males compete for being selected as mating partners. Uncertainty of paternity and higher reproductive rate will induce males to maximize the number of copulations and reduce paternal care (Tallamy 2000; Arnqvist and Rowe 2005). However, in species with high male contribution to reproduction, sex roles and sexual dimorphism can reverse from typical patterns, leading to choosy males and active females that compete and court males (Gwynne 1991; Andersson 1994; Bonduriansky 2001). Sex role reversal can be total, when only males are selective, or partial, when both sexes select their sexual partners (Gwynne 1991; Andersson 1994). Most examples come from species with paternal investment, either by male care of the progeny or delivery of nutritional nuptial gifts to the female (Gwynne 1991; Andersson 1994). Sex role reversal could be more widespread than previously thought; however, few cases have been exhaustively studied. Indeed, various authors have stressed the importance to take on the challenge of studying these atypical cases towards establishing a more robust theory of sexual roles in the animal kingdom (Tallamy 2000; Bonduriansky 2001; Roughgarden et al. 2006). In this study, we report the case of the wolf spider, Allocosa brasiliensis, an inhabitant of coastal sand dunes, showing a reversal in typical sex roles and sexual size dimorphism.

Sexual size dimorphism is considered as a consequence of the differences in the evolutionary history of each sex, driven by pressures imposed by sexual selection, natural selection or both. In most spiders species, females are bigger than males, leading to extreme sexual size dimorphism in certain cases (Vollrath and Parker 1992; Hormiga et al. 2000; Moya-Laraño et al. 2002). There are various interpretations about the origin and maintenance of sexual size dimorphism in spiders: male dwarfism as a consequence of the relaxation of sexual selection for large size caused by high mortality when searching for mates (Vollrath and Parker 1992); female gigantism, possibly in response to fecundity selection (Coddington et al. 1997; Prenter et al. 1999; Hormiga et al. 2000) or smaller size in male web spiders for efficient climbing (Moya-Laraño et al. 2002). Reports on spider species presenting bigger size in males compared to females are very scarce (Prenter et al. 1995; Lang 2001; Schutz and Taborsky 2005).

A. brasiliensis (Araneae, Lycosidae) is a wolf spider adapted to living in sand dunes, distributed along the South American Atlantic Ocean coastline. Individuals are nocturnal and show cryptic whitish color. A. brasiliensis individuals reach adulthood after 1 year of development, live approximately 1 year as adults, and females lay up to 4 egg sacs during the reproductive period (Aisenberg, unpublished data). They construct burrows where they stay during the day and the coldest months, being especially active during summer nights (Costa 1995; Costa et al. 2006). Previous studies using pitfall traps (Costa 1995; Costa et al. 2006) suggested that females were more mobile than males and preliminary data indicated a bigger size in males than females, both facts that can be considered rare in spiders (Vollrath and Parker 1992; Lang 2001; Schutz and Taborsky 2005). Therefore, this sand-dwelling spider was a good candidate for testing the hypotheses about sex role reversal.

The objective of the study was to analyze sexual size dimorphism and describe the courtship and copulation of A. brasiliensis, testing the hypotheses of sex role reversal during courtship for this species. To achieve these objectives, we captured and measured adult individuals at the field and carried out experiments under laboratory conditions, in glass cages with sand as substrate. We analyze data on female and male sizes, evidencing a reversal in typical spider sexual size dimorphism. We describe courtship and copulation for this species, giving details on the interactions between individuals during both events, discussing the courtship sex role reversal hypotheses for this species. Finally, we discuss the behavioral, ecological, and evolutionary pressures that could be driving sex role reversal in A. brasiliensis.

Materials and methods

Fieldwork and breeding of individuals

We collected sub-adult and adult individuals of A. brasiliensis (Petrunkevitch 1910) between October and January 2005 and 2006 in the coastal area of Marindia, Canelones, Uruguay (34° 46′ 49.9″ S, 55° 49′ 34.1″ W). Spiders were captured by hand during the night, using headlamps. We obtained, in total, 90 individuals that were captured either walking or leaning out from their burrows. Individuals were identified, sexed, and measured under a dissecting microscope. We measured the cephalothorax length and width of all adult individuals; measures were considered representative of the total body size in spiders (Eberhard et al. 1998; Moya-Laraño and Cabeza 2003).

Spiders were individually housed in petri dishes of 9.5 cm diameter and 1.5 cm height with sand as substrate and cotton embedded in water. They were fed three times a week with Tenebrio sp. larvae (Coleoptera; Tenebrionidae) and small Blaptica dubia (Blattaria, Blaberidae). The temperature during the experiments was of 22.81 ± 2.88°C (range, 15–31). We monitored individuals and recorded molting occurrence in sub-adults daily, for determining the exact date of reaching adulthood.

Experimental design

For the experiments, we used only virgin females, captured as large sub-adults at the field and raised to adulthood in the laboratory. The males we used in the experiments were either captured as sub-adults and raised to adulthood in the laboratory or captured as adults at the field. Females and males were randomly assigned to the experimental pairs. In each case, we used individuals of at least 10 days of adult age or 10 days after their capture at the field. Males captured as adults at the field did not show significant differences with males bred since sub-adults until adulthood under laboratory conditions, neither in burrow length (t20 = 0.286, P = 0.77) nor in mating success (Fisher’s exact test, P = 0.69). We did not reuse individuals.

We carried a total of 23 experimental trials. For the trials, we used glass cages of 30 cm length, 16 cm width, and 20 cm height with a layer of 15 cm of sand as substrate and cotton embedded in water. The sand was brought from the same capture site. The female and male from each pair were placed in the arena 48 h before the experiment with an opaque barrier dividing the arena in half and separating spiders, allowing burrow construction but avoiding contact between them. Preliminary experiments had shown that individuals constructed their burrows against the glass walls, allowing us to observe and record their behaviors when they were inside their burrows. The day of the experiment, after sunset, we lifted the barrier and recorded the behaviors of both individuals.

Each trial began when we lifted the barrier that divided the glass cage in two. When copulation occurred, the trial ended 30 min after copulation ended (when the male exited the burrow). When copulation did not take place but courtship occurred, the trial ended 1 h after lifting the barrier. If courtship did not occur, the trial concluded after a 30-min period without courtship.

The experiments took place between January 20 to April 29, 2005 and February 6 to March 10, 2006. All the trials began after 6 p.m., coinciding with the period of activity reported for the species (Costa 1995), under red light. The sexual behavior before, during, and after mating was carefully followed by direct observation. In each case, we registered the temperature, occurrence, and dimensions of the burrow (length, width) of each individual. The number of ejaculatory movements was estimated by the hematodochal expansion or by the erection of the spines of male hind legs. For a more exhaustive analysis, we recorded six sexual trials with a video camera (Panasonic S-VHS, NV-M9500EN).

Voucher specimens of both sexes were deposited in the arachnological collection of the Sección Entomología of Facultad de Ciencias, Montevideo, Uruguay.

Statistical analysis

The results were analyzed using Past Palaeontological Statistics version 1.18 (Hammer et al. 2003) and NCSS 2001 (Copyright 2000 Jerry Hintze). Variables with normal distribution (Shapiro–Wilk test) and homogeneity of variances (Levene test) were analyzed with Student t-tests for independent samples. Variables neither following a normal distribution nor having homogeneous variances were analyzed with non-parametric Mann–Whitney U-tests. Frequencies were compared using chi-square tests for independent samples and Fisher’s exact test. We also performed multiple logistic regressions. Whenever necessary, data were log-transformed before the regression analysis.


Sexual size dimorphism reversal evidence: males are bigger than females

Individuals captured in the field showed a reversal in sexual size dimorphism as was expected from previous observations. Male cephalothorax length was 6.90 ± 0.53 mm and female cephalothorax length was 5.58 ± 0.39 mm, presenting a cephalothorax width of 5.66 ± 0.52 mm and 4.65 ± 0.49 mm, respectively. Males were bigger than females (Mann–Whitney U-test: cephalothorax length U = 125, N1 = N2 = 52, P < 0.0001; width U = 142, N1 = N2 = 52, P < 0.0001).

Reversal in typical courtship roles

The occurrence of burrow construction showed significant differences between the sexes; all the males built burrows after 48 h and 15 of 23 females built burrows (chi-square test: \( \chi ^{2}_{1} = 9.68 \), P = 0.002). Male burrows were longer than female burrows, but there were no differences in burrow width between the sexes. These results are shown with the U coefficient and corresponding probability value (Mann–Whitney U-test)(Fig. 1). We investigated a possible correlation between burrow length and body size, but we did not find statistical support either for males (R2 = 0.25, N = 23, F = 0.484, P = 0.494) or for females (R2 = 0.001, N = 15, F = 0.008, P = 0.930).
Fig. 1

Comparisons of female (white bars) and male (black bars) burrows length and width; data are presented as mean ± SD. See text for statistical details

Females were the mobile sex, leaving their own burrows and approaching male burrows in 19 cases (n = 23), whereas males did not approach female burrows in any case (\( X^{2}_{1} = 35.38 \), P < 0.0001). In five cases, females opened male burrows that were closed with silk and sand grains, using their chelicerae, palps, and forelegs. We considered the initial phase of courtship when the female leaned inside the male burrow and performed waving movements alternating legs 1 and 2 and facing the male or when the male performed body shaking inside the burrow. Females initiated courtship in most cases (14 out of 19 experiences with courtship), while males started courtship only in 5 cases, and always after females approached their burrows (\( \chi ^{2}_{1} = 8.53 \), P = 0.003). In all the cases when females initiated courtship, males responded by shaking their bodies and approaching the females. Females alternated leg waving movements with abdominal vibrations, leaning inside the male burrow from the entrance, until finally entering. Males approached females, performing body shaking and touching legs 1 and 2 with those of the females, advancing, and retreating. Then, males drew back to the bottom of the burrow and females followed them. The next behavior turned to be a crucial point for copulation: males exchanged positions with females, sliding under the female chelicerae until they are situated inside the burrow, but on the top (closer to the entrance) with the females located at the bottom of the burrow. After exchanging positions, males faced the females and mounted. Males exchanged places in 11 out of 19 cases with courtship, and copulation took place in 10 of these cases, suggesting the exchange of positions as a good predictor of copulation occurrence. In the other eight cases, the females followed the males into the bottom of the burrow but the exchange of positions did not take place and females turned away from the burrow.

The courtship duration and copulation characteristics are summarized in Table 1. Copulation occurred in 10 cases that coincided with males that had constructed the longest burrows and all the females that had not constructed burrows (Fisher’s exact test, P = 0.04). Results from these comparisons are presented with t (Student t-test) or U (Mann–Whitney U-test) coefficients and the corresponding P values (Fig. 2). We performed a logistic regression with burrow measurements as independent variables and copulation occurrence as dependent variable. The results were not statistically significant either for males (\( \chi ^{2}_{1} = 1.72 \), df = 2, P = 0.422) or for females (\( \chi ^{2}_{1} = 1.83 \), df = 2, P = 0.401). There was no relationship among the open condition of the male burrow entrance and the frequency of copulation (Fisher’s exact test, P = 0.51). We found a direct relationship between male burrow length and courtship duration, but not among other copulation characteristics (Table 1). Copulation always occurred inside the male burrow, in the typical lycosid copulatory position (in this case vertically) with the male mounting on the female’s back, facing opposite to her. The copulatory pattern consisted of multiple ejaculations during a single palpal insertion, alternating the use of each palp, dismounting, performing cleaning behavior on both palps, and mounting again, repeating the sequences several times until final dismount. Females were very active during the whole mating process, approaching males while they cleaned palps and courting them inside the burrow. We found a direct linear relationship between individuals’ size and copulation duration, both for females and males. Larger individuals performed longer copulations (females: R2 = 0.557, N = 10, F = 10.065, P = 0.01; males: R2 = 0.526, N = 10, F = 8.871, P = 0.02).
Table 1

Courtship duration, characteristics of the copulations (N = 10), and results of the linear regression analysis (GLM) with the burrow length as independent variable




Burrow length




Courtship duration (min)






Copulation duration (min)






Number of mounts






Total number of insertions






Total number of ejaculatory movements






Duration of palpal cleaning (min)






Insertion frequency (ins/min)






The asterisk (*) stands for the statistically significant result.
Fig. 2

Burrow length and width when mating took place (white bars) and did not take place (black bars), data are presented as mean ± SD. See text for statistical details

In all the cases, after copulation, males exited their burrows leaving the females inside, and closed the entrances from the outside while females cooperated from the inside (Video S1). Initially, males released silk and transported the silk with sand grains using the chelicerae, and finally dragged sand using the forelegs and palps, until completely sealing the burrow entrance. Males alternated these behaviors with periods of stillness, standing on top of the burrow entrance. These behaviors lasted for 7.1 ± 4.7 min. Females released silk and collaborated in closing the burrow entrance. Finally, males left and females, in all the cases, stayed inside the male burrow, at least for the 1-week period after copulation that was recorded.


The sexual size dimorphism reversal found in A. brasiliensis, jointly considered with the occurrence of active females that look for males and initiate courtship, and males that donate their own burrows to the females after copulation, coincide with the hypothesis of partial courtship role reversal (Gwynne 1991; Andersson 1994). Large male size could be explained in terms of male competition for the best territories for constructing burrows and consequently ensuring access to females, as has been reported for other spider species (Lang 2001), but no evidence is available yet. Large male size could also be explained as an adaptation for constructing longer burrows, as was cited for a Harvestman species with males that provide breeding nests (Mora 1990). Even though our data does not statistically support this last hypothesis, perhaps the period of 48 h given for burrow construction was too brief to detect any tendency. Body size has been cited as an uncheatable signal in sex role reversed species (Kokko 1998). However, in this species, females did not select, at least directly, on this condition. Considering that the females of this species look for and locate males, the smaller size in this sex could also be an adaptation for being the ‘mobile sex’ and avoiding predation. In systems with sex role reversal, the risk of predation could also reverse so that females fall prey more often than males (Gwynne and Bussière 2002). Longer legs in males have been cited as adaptations for the ‘roving phase’ (Foellmer and Fairbairn 2005). Further investigations of leg length in this species and allometric differences between the sexes will help test the hypothesis of small size in females as a search-adapted strategy.

Results showed that females and males have different strategies in relation to burrow construction. Female burrows were constructed less frequently and were shorter than male burrows. Possibly, adult females in search of mates use their burrow just as daytime refuges and specialize in looking for long burrows, which offer a secure place for ovipositing and caring for the egg sac with stable conditions of humidity and temperature. Burrow digging in sand has been considered a costly activity for other spider species (Henschel and Lubin 1992) and burrow depth could be considered an honest signal of male quality, as has been reported for nuptial gifts in other arthropod species (Vahed 1998; Le Bas and Hockham 2005). Burrows could also be functioning as sensory traps for females who could approach and enter burrows just for safety reasons, while males take advantage of this behavior and benefit in increasing attractiveness and mating opportunities by constructing larger and safer refuges. This strategy has already been reported for sand-dwelling arthropods (Christy et al. 2003). However, in A. brasiliensis, the benefits that females would obtain—already discussed before—disagree with male exploitation.

We found a trend indicating that copulation occurs more frequently in longer burrows. However, it was not supported by the logistic regression, indicating that other factors would be determining the probability of copulation. Possibly, and considering that all the males courted, the occurrence of mating with virgin females would be determined mainly by female decisions. The fact that larger females had longer copulations suggests male preference for large-sized females. Nevertheless, male selectivity remains untested.

Females found burrows and initiated courtship before males performed any behavior that could suggest transmission of any visual, acoustic or seismic signal emitted to help females locate them. Preliminary data show that females initiate courtship exclusively when they find burrows with the male inside, but do not display sexual behaviors either when they locate other female burrows or empty burrows (Aisenberg, unpublished data). These findings suggest the existence of male pheromones in the sexual encounter, which have rarely been reported in spiders (Schulz 2004). Further research is needed on this topic.

During courtship, males performed body vibrations implying presumably seismic signals transmission through the sand, as has been reported for other arthropods from sandy habitats (Henschel 2002). The copulatory pattern with one palpal insertion and multiple ejaculations with each palp is considered the primitive pattern for lycosid spiders (Stratton et al. 1996), but the occurrence of several mounts and dismounts during copulation is reported for the first time in the family. This copulatory pattern could be related to the secure copulatory place, which would keep individuals away from predation and/or interruptions by other individuals. Moreover, the pattern of mounts and dismounts could be reflecting more stimulation and copulatory courtship in a cryptic female choice context with the male purpose of inhibiting female remating, as has been reported for other spider species (Eberhard 1996). The fact that larger individuals from both sexes presented longer copulations could imply mutual assessment and selection on large size for males and females, and a consequent major investment in copulation effort in these cases, in agreement with previous reports in spiders (Schutz and Taborsky 2005).

The cooperation among sexes in closing the burrow entrance ensures a better camouflage, complicating the detection by predators as Anoplius pompilid wasps, very frequent in the same areas than A. brasiliensis (Costa 1995). Data from pitfall traps studies (Costa et al. 2006) and present results indicate that A. brasiliensis females are sedentary when carrying egg sacs and until spiderling emergence. This would determine a sequential monogamy system with females remating only after spiderling dispersal. In this way, males would ensure the paternity of at least one clutch, justifying their high investment. Confidence of paternity has been cited as a relevant factor in driving paternal investment (Queller 1997; Hunt and Simmons 2002).

The donation of the male burrow provides the female with a stable and protected place for ovipositing, while males are exposed to predation and limit their reproductive potential until constructing a new burrow. Male discrimination is expected in systems where mating investment compromises future mating opportunities (Bonduriansky 2001; Byrne and Rice 2006). By providing protection to the mother and her egg sac, males may increase progeny survival. Nuptial gifts, in this case the burrow, are considered to reflect courtship and copulation effort and/or paternal investment (Vahed 1998) or a way of reducing costs imposed by the other sex during reproduction (Arnqvist et al. 2003). After leaving their burrows, males would need to feed before constructing the new one. In two cases, at the field, males were seen predating females of their own species (Aisenberg and Costa, personal observations). This astonishing male cannibalistic behavior, which remains to be tested further, could explain the collaboration of both sexes in closing the burrow, ensuring that the mating partner would not be attacked. The burrow can be considered a valuable resource and its delivery implies a considerable cost for males, which could be evaluated by females as reliable signals of courtship and mating effort, and good quality, in agreement with previous reports in arthropods (Tallamy 2000).

Females of A. brasiliensis reach adulthood synchronously (Aisenberg, unpublished data), and according to pitfall trap data, females are more mobile than males (1.5 females per male, whereas other lycosid species show 1.5 males per female) (Costa et al. 2006), suggesting that many females would initially be available for each male. If females remain buried in the male burrow until spiderling emergence, the number of receptive females would reduce, modifying the operational sex ratio along the reproductive period. After the female emergence from the male burrow, they need to feed and find a new place for the next oviposition. These conditions would favor remating for obtaining a long burrow for caring the egg sac, increasing again the number of females ready for copulation. The density of A. brasiliensis in the Uruguayan dunes can be considered high (Costa et al. 2006), allowing high probabilities for locating members from the other sex. However, male remating would be constrained more by female availability and by other activities such as feeding, avoiding predators, and costly burrow construction after each mating.

Unpredictable environments with periods of scarcity of prey, refuges or other resources have been reported as important factors in generating sex role reversal, nuptial gifts, and paternal investment (Karlsson et al. 1997; Lorch 2002). Allocosa species that inhabit Uruguayan coastal dunes are the sole wolf spiders adapted to this habitat where refuges and prey abundance is highly variable and dependent of weather conditions. In most wolf spiders species, females feed intensively before mating and do not feed again until spiderling emergence (Capocasale and Costa 1975; Wager 1995). In this case and additionally, females need to find a secure place for ovipositing. These evolutionary, behavioral, and ecological factors would determine the reversal in courtship sex roles found in the species: females as the mobile sex with the goal of feeding and looking for mates, and males as providers of the mating place and future nest for the progeny. The partial reversal in typical courtship roles found in A. brasiliensis opens multiple avenues to further study the constraints on the roles of each sex and size dimorphism in spiders, challenging typical views on sex-specific determination of roles in the animal kingdom.


Martín Graña, Alfredo Peretti, Fernando Pérez-Miles, and Sergio Martínez for their useful comments on a previous version of the manuscript. We are grateful to Darryl Gwynne, Yael Lubin, Matthias Foellmer, and an anonymous reviewer for their suggestions that substantially improved the manuscript. Rodrigo Postiglioni helped converting the videos to digital format. The study was supported by a grant awarded to AA from PEDECIBA, Facultad de Ciencias, Universidad de la República, Uruguay. Experiments comply with current Uruguayan and institutional laws.

Supplementary material

View video
Video S1

Male closing the burrow entrance and guarding after the copulation, while the female liberates silk and collaborates in closing the entrance from inside the burrow (WMV 1.1 MB)

Copyright information

© Springer-Verlag 2007