Martin Burd, School of Biological Sciences, Monash University, Melbourne, Victoria 3800, Australia. Tel.: +61-3-9905-5667; fax: +61-3-9905-5613; e-mail: email@example.com
Social group size may affect the potential for sperm competition, and this in turn may favour ontogenetic adjustments in testicular mass according to the likely requirements for sperm and spermatophore production. In a number of comparative analyses of testis mass among vertebrate species that differ in mating system or social organization, increasing potential for sperm competition is associated with larger testis size. Intraspecific phenotypic plasticity should be able to produce the same pattern if social group size is heterogenous and reflects differing degrees of average sperm competition, but this intraspecific effect is less well studied. We tested the effect of social groups on both male and female investment in the simultaneously hermaphroditic leech, Helobdella papillornata. Leeches were placed in groups of one, two, four or eight. Sexual investment at the onset of reproductive maturity was quantified as the total testisac volume for male function and total egg volume for female function. We found that testisac volume (statistically adjusted for body size) showed a significant increase with increasing group size. Total egg volume (also adjusted for body size) was unaffected by group size. Our findings indicate adaptive developmental plasticity in male gonad investment in response to the potential for sperm competition.
In many animals, females may have several mates within a reproductive period, creating competition among the sperm of different males for fertilization of eggs. Evolutionary models of sperm competition predict that variation in the average risk or intensity of sperm competition in a population should, under many circumstances, produce variation in the investments that males make in each ejaculate or spermatophore (Parker, 1998). Charnov (1980) considered a similar situation of multiple matings in barnacles as a problem of optimization of sex allocation (see also Charnov, 1996). In general, these models predict that an increase in the average number of competing sperm donors will favour increased resource allocation to male function (Charnov, 1980) or increased investment per ejaculate (Parker, 1998). These models describe evolutionary adjustments of male function in response to the average level of the competitive environment that characterizes a species. Other models of sperm competition envision ‘behavioural’ adjustment of ejaculate size based on information a male has about the likely degree of sperm competition in a particular mating opportunity. For example, if males can perceive that the intensity of competition in a given mating is above the species average, a decreased ejaculate investment may be favoured (i.e. if success is unlikely because competition is intense, a male will do better to make a small immediate expenditure and save resources for obtaining new mates) (Parker et al., 1996).
A third possibility is that investment in male gonads or ejaculate expenditure may be adjusted ontogenetically based on the prospective competitive environment in an individual's population or neighbourhood. Many organisms have a ‘viscous’ population structure, such that social groups in which mating and sperm competition take place vary in size. Larger or more dense groups seem likely to elevate both the average risk of multiple mating and the average number of competing sperm donors per female. If developing males can respond to cues about prospective conditions, then we expect that male investment in these neighbourhoods of different size should mirror the predicted patterns among species with different levels of average risk and intensity of sperm competition. Only the mechanism of adjustment separates these situations: ontogenetic plasticity in the former case and evolutionary change in the latter.
Ejaculate expenditure per se is rarely measured. Instead, the size of gonads relative to body size is often used as a surrogate measure of ejaculate investment. There is some justification for this substitution: a large testis mass may be needed to achieve a high rate of sperm production. Møller (1989) found that mammalian species with relatively high gonado-somatic indices (gonad mass relative to body mass) had relatively large numbers of sperm per ejaculate. Among invertebrates, some hermaphroditic polychaetes (Sella, 1990) and gastropods (Locher & Baur, 1999) require several days to replenish sperm following copulation, and larger testis mass would, presumably, allow more rapid replenishment and shorter intervals between successful copulations.
There is now considerable evidence from many vertebrate and invertebrate groups that testis size varies with the degree of sperm competition. In 31 species of microchiropteran bats (Hosken, 1997) and 17 species of megachiropteran bats (Hosken, 1998), residual testis mass (removing the influence of body size) was positively related to roosting group size. Mating systems or breeding behaviours that would tend to promote sperm competition are associated with larger testes among lemurid and lorisid primates (Kappeler, 1997), among fishes (Stockley et al., 1997), including simultaneously hermaphroditic sea basses (Petersen, 1991), and in two Onthophagus beetle species that differ in the frequency of sneaker males (Simmons et al., 1999). In some cases, there are more direct measures of the frequency of multiple matings, hence sperm competition, among species. Residual testis size was positively related to the frequency of extra-pair paternity in 53 species of birds (Møller & Briskie, 1995), and positively related to the mean female mating frequency (detected from spermatophore counts) in 28 lepidopteran species (Gage, 1994). In a particularly revealing experimental demonstration of the selective effect of sperm competition, Hosken & Ward (2001) bred yellow dung flies, Scathophaga stercoraria, under either a polyandry treatment with three matings per female or a monogamy treatment with a single mating. After 10 generations, males in the polyandrous lineages had significantly greater testis area than males in the monogamous lineages.
There has also been increasing attention to ontogenetic plasticity of investment in spermatogenic tissue. Gage (1995) reared larvae of the Indian meal moth, Plodia interpunctella, at four different densities and found that mating frequency by adult females (hence sperm competition) and residual testis volume in adult males were both greater when they had experienced higher densities in the larval environment. A similar pattern of greater residual testis size in males from larger colonies was found by Brown & Brown (2003) in a comparison among 86 colonies of the cliff swallow, Petrochelidon pyrrhonota.
Simultaneous hermaphrodites present an interesting departure from the standard scenario of sperm competition, because each potential competitor in a neighbourhood group is also a potential mating partner, and because an individual's success as a sperm donor may be linked to its egg investment, which determines its acceptability to another hermaphrodite as a mating partner (Sella, 1985; Michiels, 1998). Furthermore, simultaneous hermaphroditism is more common among ‘lower’ phyla of animals (Michiels, 1998), and perceptual and cognitive abilities needed to assess mating group size and potential sperm competition may be weaker or absent (Charnov, 1987). Indeed, comparison between hermaphrodites and dioecious organisms offers a means of testing the generality of our theoretical understanding of the selective factors acting on reproductive investment. Here, we address the question of how development in varying social environments affects male and female reproductive investment in an annelid, the hermaphroditic leech, Helobdella papillornata.
There is a small but growing body of evidence on the relation of mating group size and reproductive investment in hermaphroditic invertebrates. Raimondi & Martin (1991) studied Catomerus polymerus barnacles in neighbourhoods of either one potential mate or three or more potential mates. Barnacles in the small mating groups had a much greater ratio of egg mass to testis and seminal vesicle mass than did those in the large mating groups, in accord with Charnov's (1980) prediction. Trouvéet al. (1999) infected mice with the trematode Echinostoma caproni, an intestinal parasite, to form groups of one, two, or 20 worms per host. Relative to isolated worms, mean testicular investment was 2.6 times greater for worms in pairs, and three times greater for worms in groups of 20. Female investment by worms in groups of 20 declined to about half that of isolated worms (Trouvéet al., 1999). In contrast, Schärer & Wedekind (2001) found little evidence of plastic adjustment of male and female allocation among singles, pairs, or triplets of the tapeworm Schistocephalus solidus, although a strong body size effect on female investment was observed.
These species are either sessile, with mating neighbourhoods circumscribed by the maximum penis extension, or are internal parasites, which should also experience fairly well defined social group boundaries. Free-living motile organisms may experience more open, but still heterogenous, mating groups. Locher & Baur (2000a) found, contrary to expectations, no effect of sperm competition risk on spermatophore size or sperm delivery in the hermaphroditic terrestrial gastropod Arianta arbustorum. Schärer & Ladurner (2003) created experimental social groups of two, three, four, or eight individuals of the free-living marine platyhelminth, Macrostomum. They further replicated each social group size at two different densities to test whether the number per se or the density of potential competitors affected sexual allocation. Testis area increased significantly with increasing group size, but group density had no significant effect (Schärer & Ladurner, 2003). In this instance, phenotypic plasticity corresponds well to theoretical predictions.
Copulation in both A. arbustorum and Macrostomum sp. involves mutual intromission of copulatory organs and deposition of sperm in a specialized receptacle of the female reproductive tract (Locher & Baur, 2000b; Schärer & Ladurner, 2003). In our study organism, H. papillornata, sperm is donated by dermal implantation of a spermatophore and migration of the sperm through the coelom of the recipient's body. Such differences may affect the frequency of nonreciprocal spermatophore donation, and therefore, the likelihood of competing spermatophores, as well as the size and sperm mixing characteristics of the sperm storage compartment, which should affect optimal sexual allocation (Charnov, 1996; Angeloni et al., 2002). We experimentally imposed different social group sizes during maturation to determine how the potential mating environment affects testis and egg investment at the onset of an individual's first reproductive period.
Helobdella papillornata (Glossiphoniidae: Euhirudinea) is a carnivorous leech that typically occurs under rocks or woody debris in streams and ponds throughout Australia. These leeches are ‘sit-and-wait’ predators that attack invertebrates such as gastropods and insect larvae (Govedich & Davies, 1998). All members of the family Glossiphoniidae exhibit some degree of post-oviposition and post-hatching parental care (Sawyer, 1986; Davies & Govedich, 2001). Helobdella papillornata parents carry eggs and juveniles on their ventral surface and actively feed prey items to their young for approximately 30–50 days following hatching of the eggs.
Although these leeches are motile, their benthic habit and crawling (rather than swimming) locomotion promote aggregations in the field that range from isolated individuals to pairs to groups of several to dozens (F. R. Govedich, personal observation). For comparison, mating aggregations of two to nine individuals have been reported from field observations in Poland of the related glossophoniid leech Theromyzon tessulatum (Wilkialis & Davies, 1980). Mating in leeches occurs by implantation of a spermatophore on the dermal surface of a partner, followed by a complex process of histolysis in the vicinity of the spermatophore and transfer of sperm. Sperm then make their way through the coelomic sinuses to the ova (Myers, 1935). We have observed spermatophore implantations from multiple males in the laboratory, and our evidence from AFLP markers indicates that multiple paternity within broods may occur (F. R. Govedich, unpublished data). Thus, the potential for sperm competition exists in this species. Reciprocal mating allows spermatophores to be placed near the female gonopores of the partner, which may give those sperm better chances of fertilization. However, unreciprocated implantation of spermatophores may occur. Some mating attempts appear to be resisted by the ‘female’. Self-fertilization is possible, because isolated individuals have produced offspring in the laboratory, but our observations suggest that individuals resort to self-fertilization only after a long period in which no partners could be found.
Helobdella papillornata used in this study were obtained from a laboratory population derived from collections made at Aura Vale Lake, 25 km east of Melbourne, Australia (145°23′E, 37°55′S). Experimental groups of leeches were housed in a temperature-controlled room at 22 ± 5 °C in 200 mL of artificial pond water (pH 7.8, 80% O2 saturation, salinity = 36.1 p.p.m., electrical conductivity 95.9 μS) that was changed weekly throughout the duration of the experiment. A diet of two live Potamopyrgus antipodarum snails (2–4 mm in length) per leech per week was provided. Because H. papillornata often feed co-operatively from a single snail, it is unlikely that any individuals monopolized the food supply. However, we checked for treatment effects on body size and adjusted reproductive investment measures to body size to minimize any density effects mediated through body size.
Young, reproductively immature H. papillornata (easily distinguishable by the absence of the white ovisacs that become visible at reproductive maturity) of similar size (6–8 mm length) were segregated in individual containers for 2 weeks at the beginning of the experiment (n = 180). The isolation period ensured that all individuals had experienced nearly identical social environments at the start of the experimental treatments. Following isolation, individuals were randomly assigned to a group of one, two, four, or eight (with 12 replicates of each group size). Six replicates for each group size were killed for destructive determination of testisac volume 9 days after formation of the treatment group, when the first signs of ovisac development were noted. Individual Helobdella undergo a brief protandrous phase before the female organs are mature (Kutschera & Wirtz, 1986), as is common in many simultaneous hermaphrodites (Michiels, 1998). We measured testisac volume just prior to full hermaphroditic competence in order to determine how investment decisions are made in response to the prospective mating environment, before individuals experienced a potentially variable mating history. The remaining six replicates proceeded until all individuals had produced a clutch of eggs. Eggs were produced from 12 to 74 days after the start of treatments (mean ± SD, 33.5 ± 12.9 days).
Testisac volume was measured from serial histological sections of individual leeches. Each individual was fixed in 10% formalin, dehydrated in propanol, cleaned with xylene and embedded in paraffin wax before serial longitudinal sections of 7 μm thickness were cut. Sections were stained with Harris’ haematoxylin, and counter stained with eosin Y. Each section was examined under a dissecting microscope with an ocular micrometer that allowed measurement to <0.01 mm. Body length and width were measured on the largest section, and their product was used as a measure of body size. Total testisac area in each section was determined by treating each testisac as an ellipse (area = πab, with a the semi-major axis and b the semi-minor axis). These areas were multiplied by the thickness of each section (7 μm) to obtain a volume, and the volumes from an individual's serial sections were summed to obtain total testisac volume. Body sizes of histological preparations were commensurate with those of living leeches, and there were no gross deformations of the testisacs, such as separation from surrounding tissue, in the sections. Testisacs could not be discerned in 18 leeches because the body curled during histological preparation and clean longitudinal sections were not obtained. These individuals were removed from the data set.
For those individuals allowed to produce eggs, images of the parental leech and its eggs were taken using an Olympus DP10 digital camera and measurements (to 0.01 mm) were made using HLImage++97 image analysis software (Western Vision Software, Salt Lake City, UT, USA). Eggs were counted and individual egg volumes were calculated by treating eggs as ellipsoids (volume = 4πab2/3, with a the semi-major axis and b the semi-minor axis of the digital image of the egg). Total female investment was quantified as the sum of egg volumes for each individual. Parent size was calculated as the product of the length and width of the individual.
Differences in body size among treatment groups were checked by anova. The effect of treatment group on testisac and egg investment was tested using ancova with body size as a covariate. In this analysis, both testisac and egg volume data were log-transformed to improve the linearity of the relationship with the covariate. We initially tested the homogeneity of slopes by examining the significance of the treatment group × body size interaction. The interaction effect was nonsignificant for both the testisac and egg data, suggesting that slopes were homogeneous across treatments. We then calculated the ancova model without an interaction effect. When the treatment effect was significant, we performed a post-hoc comparison of the adjusted treatment means using the Tukey–Kramer method based on the studentized range (Sokal & Rohlf, 1995).
Individuals used for testisac measurement averaged 5.8 ± 1.39 SD mm in length and 2.11 ± 0.32 SD mm in width. The leeches used to measure egg volume attained larger sizes than those used to measure testisac volume, because they were allowed to grow for a longer time before measurement. These individuals averaged 10.68 ± 2.09 SD mm in length and 3.81 ± 0.73 SD mm in width. Within each sex-role sample, the body size ranges were broadly similar among treatment groups (Fig. 1). However, there were significant differences in mean body size among the leeches used to measure testisac volume (anova, F3,65 = 4.19, P < 0.01), with isolated leeches having greater mean size than those in groups of eight (Tukey post-hoc test, P < 0.01). Mean body size did not differ among treatment groups used for egg volume measurement (anova, F3,85 = 1.79, NS).
The pattern of reproductive investment within each treatment group is given in Table 1. Total testisac volume varied 50-fold across all leeches in the experiment, from 0.0012 to 0.0588 mm3. Clutch sizes ranged from 16 to 60 eggs, and total egg volume varied from 0.47 to 2.00 mm3, an approximately four-fold range.
Table 1. Testisac and egg investment for different treatment group sizes.
Total testisac volume (mm3)
Total egg volume (mm3)
Mean ± SD
Mean ± SD
0.0069 ± 0.0052
1.022 ± 0.299
0.0090 ± 0.0051
1.228 ± 0.347
0.0131 ± 0.0091
1.215 ± 0.287
0.0204 ± 0.0149
1.046 ± 0.338
The effects of social group size on reproductive investment are shown in Fig. 2. The preliminary analysis to examine homogeneity of slopes for the ancova model showed a nonsignificant interaction between the body size covariate and treatment group for both the testisac data (F3,64 = 0.53, P = 0.66) and egg data (F3,82 = 1.19, NS). Therefore, the null hypothesis of equal slopes across treatments was not rejected in either case. The ancova for testisac investment indicated a significant effect of body size (F1,67 = 6.64, P < 0.05) and significant variation among treatment group means (F3,67 = 5.85, P < 0.001). For the model as a whole, R2 = 0.22. Mean testisac volume, adjusted to a common body size, increased monotonically as social group size increased (Fig. 2). The Tukey–Kramer comparisons indicated that mean testisac size was significantly greater in groups of eight than in other groups. In contrast, egg investment did not vary significantly among groups (F3,86 = 2.23, P = 0.09), although egg volume did increase significantly with body size (F1,85 = 6.41, P < 0.05). For the model as a whole, R2 = 0.13. The adjusted means showed no consistent pattern across social group sizes (Fig. 2).
Helobdella papillornata is a likely candidate species for sperm competition. As in other glossiphoniid leeches (Wilkialis & Davies, 1980), the population structure of H. papillornata appears to promote local mating groups of different size. Individuals may receive spermatophores from several mates, and movement of sperm through the coelomic sinuses towards the ovisacs (Myers, 1935) is likely to allow mixing of sperm from multiple donors, so that precedence rules probably approximate ‘fair raffles’ (Parker, 1998). Male reproductive success under these conditions is likely to depend on the amount of sperm transferred in a spermatophore, and thus on the testisac tissue available to produce sperm. Sperm production can be costly, for example, the polychaete annelid Ophryotrocha diadema can take 3 days to replenish sperm supplies following mating (Sella, 1990). If sperm production in H. papillornata is similarly costly, the optimal investment in testisac tissue should vary with the likely degree of sperm competition.
Our results support the idea that hermaphroditic H. papillornata individuals make plastic adjustments to their testisac volume during maturation in response to variation in social group size. As group size varied from one to eight, mean testisac volumes (adjusted for body size) showed a monotonic rise of approximately three-fold (Fig. 2). Such adjustments would be adaptive if the social group size experienced by juveniles predicts the prospective intensity of sperm competition they will face. We do not know the mechanism by which leeches assess their group size, but tactile and chemical cues about the number or density of neighbours are likely sources of this information. For comparison, excretory or secretory chemicals are implicated in attraction and sexual behaviours in hermaphroditic trematodes (Trouvéet al., 1999).
The differing densities in each treatment had some effect on the growth of leeches (Fig. 1). This variation in body size reinforces our interpretation of adaptive adjustment in testisac investment. If a testisac size scaled positively with body size, then larger leeches should also have had larger testisacs. In fact, the large leeches in singletons or pairs had smaller testisacs than the smaller leeches in quartets or octets (Table 1). Thus, when testisac volume is adjusted for body size in ancova, there is a strong pattern of increased male investment with increasing group size (Fig. 2).
There was not, however, any significant change in female allocation with increasing group size (Fig. 2). In the absence of sperm limitation of female fecundity or direct fitness benefits of sperm competition or multiple paternity of broods, we find no obvious adaptive reason for leeches to vary their female investment in relation to mating group size. Although total egg investment did not vary with group size, it did vary with body size of the maternal leech. This positive association of female investment and body size was loose (R2 = 0.13 for the ancova model) but statistically significant. Among a number of hermaphroditic organisms, large individuals tend to a female allocation bias whereas small individuals are more male biased (Petersen & Fischer, 1996; Schärer & Wedekind, 2001; Angeloni & Bradbury, 1999).
Our data do not allow direct comparison of relative male and female investments, because our measurements of investment were made on different individuals and because testisacs and eggs may not represent equivalent investment costs. Nonetheless, egg volume greatly exceeds testisac volume (and including the tiny spermatophores would not greatly alter the comparison), so that female investment seems likely to exceed male investment. The difference may be due to a body-size advantage that provides greater gains for female than for male fitness (Ghiselin, 1969). As H. papillornata body size increases there is likely to be a rise in the resources available for maternal function, including vitellogenesis, in the ventral area available for brooding eggs and young, and possibly in the efficacy of maternal care behaviours, such as ventilation of the clutch. In contrast, paternal success may rise only modestly as body size increases. There are no obvious contests or lekking in which large body size would determine access to mates, although it is possible that size related capacity for egg production affects the acceptability of a potential partner in reciprocal matings (cf. Sella, 1985), thus tying paternal success to an individual's maternal investment (Michiels, 1998). However, body size appears not to be an advantage to male function in simultaneously hermaphroditic snails (DeWitt, 1996) or sea slugs (Angeloni, 2003). We are currently conducting experiments with H. papillornata to determine how body size influences mate choice and realized paternity.
Although the details of mating interactions in leeches that relate group size or density to sperm competition remain to be investigated, it seems likely that local group size at least approximately reflects the probable degree of sperm competition. Our results here indicate that, at the onset of sexual maturity, H. papillornata individuals adjust their testisac investment in relation to group size in a manner that should be adaptive. Interspecific comparative evidence supports the idea that testis mass is related to the degree of polygyny and potential sperm competition in many vertebrates and insects (Birkhead & Parker, 1997). Our results along with those of Trouvéet al. (1999) and Schärer & Ladurner (2003) add evidence that similar intraspecific patterns of male investment occur in hermaphroditic invertebrates.
The authors thank Ian Boundy for assistance in preparing histological sections, and Lisa Angeloni, Bonnie Bain, Eric Charnov and Bill Moser for insightful comments on drafts of the manuscript and for alerting us to relevant literature. This work was supported by an Australian Research Council grant to MB and FRG and by a Monash University Research Fund fellowship to FRG.