Many factors affect an organism’s ability to survive and reproduce. These factors are often called “selection pressures” and include the availability of food and shelter, conditions in the environment such as temperature, and the presence of diseases and predators. Males and females experience different selection pressures so they often evolve to look different – consider, for example, the male deer’s antlers and the peacock’s colourful tail feathers. Such traits arise from a phenomenon called sexual selection, the selection pressures that act on an organism’s ability to obtain a mate.

Measuring sexual selection is not only of interest to scientists looking to understand how evolutionary processes work; it also has wider applications, including in wildlife conservation. For instance, knowing which cues are important for successful reproduction could help efforts to breed endangered animals in captivity and stop them from going extinct.

Scientists study sexual selection in a species by measuring how successful males and females are at mating and reproducing. Past studies have found that a female’s reproductive success mainly depends on there being enough resources available for her to produce eggs, while a male’s success depends on him getting access to these eggs. However, most research into sexual selection has been on species with separate sexes. It is more difficult to measure sexual selection in species – like snails and slugs – where each individual is male and female at the same time. As such, it is not clear if reproductive success in these species, which are known as simultaneous hermaphrodites, depends on the same factors as those species with separate sexes.

To address this, Hoffer et al. measured sexual selection in the great pond snail Lymnaea stagnalis, a simultaneous hermaphrodite. Most studies estimate sexual selection based on measurements taken over several days. Instead, Hoffer et al. observed the great pond snail over a period of eight weeks, which is about a quarter of its reproductive life.

The experiments showed that mating multiple times, especially with multiple partners, overall improves the development of the snail’s offspring. The male part of the great pond snail gains the most reproductive success from repeated mating, whereas the female part may in fact be negatively affected. These negative effects were only seen several weeks into the experiment, and so they show that sexual selection pressures change over time.

Future research is needed to determine what causes the negative effects on the female part of the great pond snail. Overall, these findings stress the need for careful consideration of the time frame over which future measurements of sexual selection take place, not just in hermaphrodites, but in all species.

Darwin defined sexual selection as selection on traits that affect mating success (Darwin, 1871). In doing so, he clearly focused on the obvious secondary sexual characters that often differ between males and females. Classical examples include antlers of deer, long and extravagantly coloured (tail) feathers in birds and traits of that ilk. In recent decades, this definition has been refined, most notable due to the realisation that sexual selection does not only act prior to mating – referred to as pre-copulatory sexual selection (Darwin’s focus) - but also after mating - post-copulatory sexual selection (e.g., Parker, 1970; Eberhard, 1996). Investigating pre-copulatory sexual selection involves measuring mating success, which is generally defined as the number of mates that an individual copulates with (for males MS_m; for females MS_f), and the resulting reproductive success which is measured as the number of offspring that an individual is able to produce (for males RS_m; for females RS_f). The residual variation of the relationship between mating success and reproductive success can then be used as a quantitative proxy for post-copulatory sexual selection. These sexual selection processes are often measured, but with a strong bias towards species with separate sexes, even though simultaneous hermaphrodites are under influence of the same selective pressures (Charnov, 1979; Nakadera and Koene, 2013; Pélissié et al., 2012; Schärer and Pen, 2013).

One, now classical, study on Drosophila melanogaster by Bateman (1948) sparked the pivotal insight that the factors that limit reproductive success for males and females are different. This is referred to as the Darwin-Bateman paradigm (or Bateman principle; reviewed in Dewsbury, 2005). Due to anisogamy, for which sexual selection has been an important driving force (reviewed in Parker and Birkhead, 2013), a clear difference is found in the cost of producing male and female gametes. As a consequence, sperm production is generally not a limiting factor for males, so the number of fathered offspring (male reproductive success, RS_m; Bateman, 1948) depends directly on the number of mates the male can inseminate. In contrast, egg production is highly dependent on available resources and therefore limits the number of offspring the female can produce (i.e., female reproductive success, RS_f; Bateman, 1948).

Even though Bateman’s experiment has been criticized based on experimental design as well as data collection and actual repeatability (e.g., Gowaty et al., 2012; Gowaty et al., 2013), the basics of the Darwin-Bateman paradigm still hold (e.g., Janicke et al., 2016). Therefore, this paradigm has formed the basis for more formalized approaches to measuring and quantifying sexual selection, in which the difference in variance in mating success and reproductive success of males and females can be captured in different measures of sexual selection (e.g., Arnold, 1994; Jones, 2009; see Materials and methods). One insightful measure that has emerged is the sexual selection gradient (also referred to as Bateman gradient), which looks at the linear relationship between mating success and reproductive success (β; e.g., Arnold and Duvall, 1994; Anthes et al., 2010). The steepness of the male and female Bateman gradient can inform about the strength of sexual selection, as has for example been done for polygamous red jungle fowl (Collet et al., 2012), polyandrous rough skinned newts (Jones, 2009), and polygynous, role-reversed pipefish (Jones et al., 2005).

For species with separate sexes, the above approach is relatively straight forward, because one ‘only’ needs to regress mating success against reproductive success for each of the sexes and compare the slopes (β_male and β_female). As pointed out by Gowaty et al. (2012), in order to obtain correct quantifications of sexual selection, independent measures of mating success and reproductive success are needed, which lacked in Bateman’s original study. As furthermore pointed out by Anthes et al. (2010), the quantification of sexual selection is less clear-cut in simultaneous hermaphrodites, because each individual is both male and female at the same time. Hence, besides that mating success in the male role can affect the individual’s male reproductive success, it can also directly affect its female reproductive success and vice versa. These interactions between the sexual functions of the individual are referred to as cross-sex effects (β_mf and β_fm; Anthes et al., 2010) and trigger the legitimate question whether these cross-sex effects cause simultaneous hermaphrodites to deviate from the Darwin-Bateman paradigm.

As illustrated by several recent studies, it is possible to quantify sexual selection in simultaneous hermaphrodites (Ophryotrocha diadema: Lorenzi and Sella, 2008; Biomphalaria glabrata: Anthes et al., 2010; Physa acuta: Pélissié et al., 2012; Macrostomum lignano: Marie-Orleach et al., 2016), with the latter three also applying the sexual selection gradients approach outlined by Anthes et al. (2010). Like in most of the sexual selection gradient studies on separate sexed species, these studies used a very restricted time window (days) within which the strength of sexual selection was estimated. However, many species are long-lived, mate many times and can store and use sperm for extended periods (e.g., Nakadera and Koene, 2013; Nakadera et al., 2014a). Therefore, the relationship between mating success and reproductive success can be expected to change over time, especially when considering that sperm storage becomes important as soon as individuals are no longer virgin, meaning that the degree to which mating success translated into reproductive success might change (Baena and Macías-Ordóñez, 2012; Wacker et al., 2014; Anthes et al., 2016). Here, we present an experiment in which we quantify mating success and reproductive success, using the great pond snail Lymnaea stagnalis, over an eight week period that represents roughly a quarter of its reproductive life in nature.

Our quantification of sexual selection gradients allows us to address several unresolved questions that are of general importance for understanding sexual selection. First, do these simultaneous hermaphrodites conform to the prediction that sexual selection gradients differ for the male and female function? Second, can we detect the predicted cross-sex effects on reproductive success? Third, do sexual selection gradients change depending on the time window of measurement? This latter question, which can be tested given the time frame of our experiment, addresses an issue that has remained experimentally untested in any hermaphroditic species to date (and was only addressed in one separate sexed species: Turnell and Shaw, 2015).

In addition, with the collected data, we can address several remaining questions that are specific to the simultaneous hermaphrodite under investigation, the pond snail L. stagnalis. We can examine whether partner availability is beneficial for offspring quality, something that is predicted based on the finding that multiple mating results in larger investment per egg in this species (Hoffer et al., 2012). Also, by looking at the number of matings in the male and female role, we can determine whether the mating mode of this species is unilateral or relaxed reciprocal (the latter meaning that animals tend to alternate mating roles between successive copulations, possibly with different partners; we already know it is not strict reciprocal: Koene and Ter Maat, 2005). As pointed out by Anthes et al. (2010), this information is important, as the mating mode has implications for how independently sexual selection can act on the two sexual roles of hermaphrodites.

The 200 individuals of the simultaneously hermaphroditic snail L. stagnalis, which were all virgin at the start of the experiment, were divided over three treatment groups: Multiple partners, Single partner and No partner (see Figure 1 and Materials and methods for details). During the whole experiment, a total of 7888 matings were observed in the first two treatments (Figure 1). Of the estimated 343,745 eggs that were produced, 1102 were genotyped (excluding eggs from the selfing No partner treatment). At the end of the experiment, 89 egg masses, containing a total of 8407 eggs, were used to determine hatching and developmental success for each treatment. Focals that died during the experiment were excluded from that point onwards from further analysis. At the end of the experiment, a total of 25 out of 125 (20%) snails died in the Multiple partners treatment. For the Single partner and No partner treatments these numbers were 8 out of 46 (17%) and 3 out of 25 (12%), respectively. Overall survival during the whole experiment was 87.8%, which is within the common range for this species in our experience, and did not differ between the treatment groups (χ²₂ = 1.303, p = 0.521).

Figure 1

Download asset Open asset

Quantification of sexual selection

The measurements of mating activity and reproductive output in the above-mentioned Multiple partners treatment were used to quantify sexual selection by looking at the variance for each sex of one focal within a group of five snails (see Material and methods for details). Due to different limits in terms of their gamete production, variance in the reproductive success of males is expected to be larger than that of females and can be captured in the variance measure I. This measure is defined as the standardized variance in relative reproductive success and its value is indicative of the opportunity for selection (Arnold, 1994; Anthes et al., 2010; Evans and Garcia-Gonzalez, 2016). The opportunity for sexual selection, which is defined as the standardized variance in relative mating success, is captured in the variance measure I_s (Arnold, 1994; Anthes et al., 2010). Comparison of the opportunity for selection values between the male and female role (I_m and I_f) shows that these values are larger for the male role (Figure 2—figure supplement 1 and Figure 2—source data 1). In addition, these values decreased over the course of the experiment. The much lower values for the opportunity for sexual selection (I_sm and I_sf) reveal a similar trend (Figure 2—figure supplement 1 and Figure 2—source data 1).

Subsequently, we compared the sexual selection gradients (β) between and among the sexual roles. Such a gradient is the linear least-squares regression slope of sex-specific relative RS on sex-specific relative MS (Jones, 2009; Klug et al., 2010), thus expressing the expected fitness increase achieved by mating one additional time in a specific sex role (β_m or β_f). Because hermaphrodites express both sex functions within one body, the reproductive efforts in one reproductive function can alter reproductive fitness in the other. To deal with this non-independence of male and female reproduction, we used a multiple regression with MS_m and MS_f as explanatory variables on, respectively, RS_f and RS_m. As pointed out by Anthes et al. (2010), this approach makes possible cross-sex effects explicit. The resulting cross-sex effects (β_mf and β_fm) describe how MS in one sex function changes RS in the other sex function.

In order to make the sexual selection gradient measures comparable across time, we used relative values for both mating success and reproductive success; note that time was divided in weeks because eggs were collected at the end of each week (hence the factor Week below). To analyse the sexual selection gradients from the male perspective, we used a model including the dependent variable relative RS_m and the factors relative MS_m, relative MS_f and Week, plus their interaction, including focal identity as random, repeated factor. This analysis revealed a significant effect for the factor relative MS_m on relative RS_m (F_{1, 145.5} = 7.872; p = 0.0057), while Week and the interaction term Week*relative MS_m showed no significance. This indicates that the male sexual selection gradient, β_mm, is positive and does not change over time, as is also clearly reflected in the regression lines (Figure 2; see β-values in Figure 2—figure supplement 1, the slopes’ confidence intervals in Figure 2—source data 2, and slope comparisons between weeks in Figure 2—source data 3), revealing that continued mating in the male role assures continuous male reproductive success beyond the first week. When looking at the cross-sex effect β_mf in this model, the factor relative MS_f had a significant effect on relative RS_m (F_{1, 145.7} = 29.956; p = 0.0001), which is due to the significant positive relationship between the number of male matings and female reproductive success in the first week, as reflected by the significance of the interaction term Week*relative MS_f (F_{7, 136.1} = 4.159; p = 0.0004); an effect that disappeared afterwards (Figure 2 and Figure 2—figure supplement 1). From the female perspective, running the same model with relative RS_f as dependent variable, for the female role (β_ff), a significant relationship between female mating success and female reproductive success is found (F_{1, 140.6} = 5.101; p = 0.0255), which seems to be caused by the negative trend lines found in the last three weeks of the experiment (Figure 2 and Figure 2—figure supplement 1). For the cross sex effect β_fm the model revealed a significant effect of relative MS_m on relative RS_f (F_{1, 140,5} = 13.523; p = 0.0003), which seem due to the initial positive correlation (in Week 1 and 2; Figure 2 and Figure 2—figure supplement 1), even though the interaction term Week*relative MS_m was not significant.

Figure 2 with 1 supplement see all

Download asset Open asset

Figure 2—source data 1

The calculated values and their confidence interval (CI) for the opportunity for selection (I) and sexual selection (I_s) for both sexual roles, indicated with subscript m or f, over the weeks.

https://doi.org/10.7554/eLife.25139.005

Download elife-25139-fig2-data1-v2.docx

Figure 2—source data 2

Slope comparisons between weeks for all the significant sexual selection gradients shown in Figures 2 and 3.

For the top right of the table this shows the results for the regressions between relative male mating success (MS_m) versus relative male reproductive success (RS_m) and for bottom left (in italics) the results are shown for the regressions between relative principal component 2 (PC2) and relative male reproductive success (RS_m). Each cell of the table comparing two slopes indicated the t-value, degrees of freedom (df) and significance (p). The first value column and row, respectively, indicate the sample size (N), the slope (β) and the standard error (SE) of the compared lines. The cells in the Week 1 row are left empty (-) because this line was not significant for relative MS_m versus relative RS_m and could thus not be compared with the other slopes.

https://doi.org/10.7554/eLife.25139.006

Download elife-25139-fig2-data2-v2.docx

Figure 2—source data 3

Slope and confidence interval of the correlations between the different sexual selection measures.

The values are only given for the significant (p < 0.05, black) and near-significant (p < 0.10, grey) slopes shown in Figures 2 and 3.

https://doi.org/10.7554/eLife.25139.007

Download elife-25139-fig2-data3-v2.docx

Depending on their mating system, male and female mating success may not be fully independent in simultaneous hermaphrodites. Even in unilaterally mating species, playing both roles in a mating encounter in sequence (i.e., reciprocating) could make a multiple regression analysis statistically fragile (Mitchell-Olds and Shaw, 1987). To cope with this potential problem, we follow Anthes et al. (2010) suggestion to replace MS_m and MS_f by their principal components (PC). When comparing the outcomes of the regression analysis with the principle component analysis (PCA) approach, one can see that most findings were corroborated (comparing Figures 2 and 3). For proper interpretation of these graphs, it should be noted that overall mating activity, that is, the correlation component between relative MS_m and relative MS_f is represented by PC2; the sexual bias, that is, the relative difference between MS_m and MS_f, is captured by PC1 (see Figure 3—source data 1 for details; see also Figure 2—source data 2 for slope comparisons between weeks). In other words, the slope β_mPC2 represents β_mm and β_fPC2 represents β_fm, and the cross-sex effects are seen in β_mPC1 (= β_mf) and β_fPC1 (= β_ff; Anthes et al., 2010). One crucial difference to note between the two analytical approaches is that a more female biased mating rate (i.e., higher PC1 values) can negatively affect male reproductive success (negative slope, β_mPC1), which is not seen in the regression analysis between relative MS_f and relative RS_m. Similar negative trend lines are seen for β_fPC1 and β_ff, confirming that a more female biased mating rate may also negatively affect the individual’s female reproductive success (in terms of offspring number, but see below).

Figure 3

Download asset Open asset

Figure 3—source data 1

Results of the principal component analysis (PCA) for each week of the experiment.

PC1 represents the sexual bias (the relative difference between MS_m and MS_f); PC2 represents the overall mating activity (the correlation component between MS_m and MS_f).

https://doi.org/10.7554/eLife.25139.010

Download elife-25139-fig3-data1-v2.docx

Mating mode

During the whole experiment, all the mating interactions were observed. This allowed us to assess whether this species mates fully unilateral or via a form of relaxed reciprocity (Anthes et al., 2010). When looking at the relationship between male and female mating success (respectively, MS_m and MS_f) in the Multiple partner treatment, no significant correlation emerges (Figure 4). We tested this using a GLMM with MS_m and Week as factors, MS_f as dependent variable and focal identity as random factor (MS_m: F_{1, 147,6} = 1.314, p = 0.254; Week: F_{7, 144,6} = 12.421, p < 0.0001; Interaction: n.s.). If animals had alternated roles between matings, either with the same or a different partner, this would have resulted in a positive correlation (note that in the Single partner treatment reciprocity is enforced). The statistical significance of the factor Week simply reflects the cumulative nature of these data and reveals that animals keep mating as male and female throughout the experiment, as also illustrated in Figure 4. For example, mean cumulative mating success at Week 2 averaged around 3 and at Week 8 around 12 matings for each sexual role.

Figure 4 with 1 supplement see all

Download asset Open asset

We also looked at whether there was a clear correlation between male and female reproductive success (respectively, RS_m and RS_f), which might be indicative of overall individual quality. While there was a slight correlation in Week 2 (R = 0.41, N = 24, p = 0.046; Figure 4—figure supplement 1; which might explain the positive relationship between MS_m and RS_f in the first two weeks, see above), this relationship was absent throughout the rest of the experiment (Figure 4—figure supplement 1). This was also tested using a GLMM with RS_m and Week as factors, RS_f as dependent variable and focal identity as random factor (RS_m: F_{1, 151} = 3.115, p = 0.080; Week: F_{7, 146,8} = 28.015, p < 0.0001; Interaction: n.s.). Again, the statistical significance of the factor Week reflects the cumulative nature of these data and reveals that animals keep obtaining male and female reproductive success over the course of the experiment. For example, mean reproductive success at Week 2 was higher for the female than the male role, but these means lie much closer together at Week 8, around 2000 eggs (Figure 1). The reason for male and female reproductive success not being equal is found in the fact that selfing occurs at the start of the experiment, which only contributes to RS_f.

Effect of partner availability

In order to evaluate whether repeated mating is beneficial for offspring in this species, we assessed differences in growth, egg laying and hatching at the end of the experiment between the three treatment groups. There was no difference in body size (shell length: ANOVA: F_{2, 69} = 2.385, p = 0.100; body weight: ANOVA: F_{2, 61} = 1.441, p = 0.245) between the focals of the different treatments. Also, we found no difference in number of eggs between the masses produced by the focals of the different treatments at the end of the experiment (ANOVA: F_{2, 63} = 0.155, p = 0.857). Because we did not follow the development of all the egg masses that were laid on the final day of the experiment, we verified whether there was a difference in the number of eggs that we followed per treatment, but this was not the case (ANOVA: F_{2, 31} = 0.168, p = 0.846).

The overall proportion of hatching success and development scored after 14 days differed between the treatments (χ²₂ = 473.245, p < 0.0001). Using nonparametric multiple comparisons, we found that in the Multiple partners treatment significantly more offspring had reached hatching than in the No partner treatment (Wilcoxon: Z = −2.609, p = 0.009), while the Single partner treatment did not differ significantly from either treatment (Figure 5A). When looking at the proportion of undeveloped offspring, the Single partner and No partner treatments differ significantly from each other (Wilcoxon: Z = 2.060, p = 0.039), while the difference between the Multiple partners and No partner treatment shows a trend (Wilcoxon: Z = 1.714, p = 0.086; Figure 5B).

Figure 5

Download asset Open asset

The study that we present here is the first to look in detail at the effect of repeated mating on reproductive success over time in a simultaneous hermaphrodite. By using a simultaneous hermaphrodite, we could answer several unresolved questions that are relevant for the understanding of sexual selection in general. Firstly, we showed that the potential gain in reproductive success is consistently positive via the male function. Secondly, the existence of cross-sex effects was supported by the sexual selection gradients, emphasizing that their effects are important in the long run. Thirdly, our data clearly showed that sexual selection gradients change over time, which is important for the interpretation of these values. Fourthly, we showed that in this pond snail, repeated mating was beneficial for the development of offspring. Finally, our data confirmed that the investigated hermaphrodite mates unilateral, without any conditional role alternation. In the following, we will briefly discuss each of these conclusions and place these in the larger context of their general implications for sexual selection.

While it becomes clear from the above that it is very informative to run such an experiment for a longer time, this approach also emphasises that it does complicate the interpretation of the sexual selection measures. In short-term experiments, and (imposed) semelparous situations, one can separate pre- from post-copulatory components of sexual selection by looking at the residual variance in reproductive success, i.e. the variance not explained by mating success (this residual can be explained by post-copulatory processes: Rose et al., 2013a, 2013b; Pélissié et al., 2014; Janicke et al., 2015). This is no longer straight-forward in a dataset from a longer running experiment where everyone has mated repeatedly with everyone else and sperm from those previous matings is still in storage. So, even though our data show that at the end of the experiment 19% to 52% of the variance in reproductive success (RS) is explained by mating success (MS; which is essentially pre-copulatory), post-copulatory processes, such as sperm storage and sperm competition, have had time to act on reproductive success (and are thus partially included). Hence, one can still attribute the remaining variance to post-copulatory processes, but this no longer captures all this variance. For example, age now becomes a confounding factor and cannot be fully excluded. These considerations are of biologically relevance because they apply to any species, separate sexed or hermaphroditic, that mates multiple times over an extended reproductive period and stores sperm before producing eggs. The only studies that have been able to deal with this issue partly (though not the factor age), so far, are the ones by Pélissié et al. (2014) and Turnell and Shaw (2015). They achieved this by taking mating order into account, which was facilitated by the short duration of their experiments, and their findings largely corroborate what we observe early on in our experiment. To also be able to take the effect of age into account, a similar experiment to the one presented here would need to be run in which the virgin snails (and their mating partners) have different standardized ages at the start of the eight-week experiment.

Our data reveal that repeated mating, especially with several different partners is beneficial for the hatching success of the offspring. This is especially true when the Multiple partner treatment is compared to the No partner treatment. Hence, this could also suggest that self fertilization, as occurred in the No partner treatment, is detrimental to offspring. Similar effects of selfing have been found for many snail species (e.g., Escobar et al., 2011), and are corroborated by the higher number of undeveloped eggs in that treatment. Note that in this case the non-significant difference between the No partner and Multiple partners treatments, although they seem very different, is due to both sample size and the non-parametric test that needed to be used. Interestingly, results so far had not indicated very clear negative effects of selfing for L. stagnalis (Escobar et al., 2011; Puurtinen et al., 2007), so the fact that we do observe it here may reflect that we looked at the development of offspring in more detail, or that such effects only become apparent in the long run. In addition, while the Single partner treatment differs from neither of the other treatments, it is suggestive of a gradient in which repeated mating may also be beneficial for hatching success when it occurs with the same partner (but to properly answer this question, a Single copulation treatment would need to be included in a follow-up study). Notwithstanding, our data are the first to show that multiple mating does offer a benefit for the development of eggs of L. stagnalis, and are supported by previous work that showed that such eggs are heavier (Hoffer et al., 2012). These findings are also in line with work on other species (e.g., Callosobruchus maculatus: Power and Holman, 2014). It remains to be investigated whether this effect on offspring has any further implications on Lymnaea’s growth and survival.

Multiple mating has already been shown to preferentially occur with different partners (Koene and Ter Maat, 2007) and behavioural studies imply that this species does not alternate mating roles conditionally (i.e., they can swap sexual roles within one mating interaction, but this is not obligatory: Koene and Ter Maat, 2005). This is of importance since it has implications for whether sexual selection can act independently on the two sexual roles of a hermaphrodite (Anthes et al., 2010); see also Arnold, 1994). Given that our current data reveal no relation between MS_m and MS_f, this supports that the mating mode of this species is unilateral (no exchange of sex roles). Hence, male and female strategies can probably be optimized independently in this species. The lack of a clear correlation between RS_m and RS_f also indicates that an individual successful in the male role is not necessarily successful in the female role. The latter is also supported by our finding that reproductive success can be gained constantly via the male function, as shown by the significant male sexual selection gradients (β_mm) throughout the experiment.

Given that this mating system seems largely male-driven, we expected to also see cross-sex effects. These effects should become visible as negative gradients when looking at male or female reproductive success (respectively, β_mf or β_mPC1 and β_fm or β_fPC1). Such negative effects emerged more clearly from the PCA approach, and only near the end of the experimental period (β_fPC1). This may explain why such cross-sex effects were not found in earlier studies that lasted only for several days (Anthes et al., 2010; Pélissié et al., 2012; Pélissié et al., 2014). Moreover, our results highlight that cross-sex effects do have an impact on the reproductive success of this hermaphroditic animal. The cross-sex effects that we observe can potentially be explained by the reported negative effects of seminal fluid proteins on both sexual functions (Koene et al., 2010; Nakadera et al., 2014b), but this remains to be tested directly. Interestingly, we did not find any indication for a trade-off between investing in the two sexes, which would have resulted in a negative β_fm (although a trend is seen in the last weeks of the experiment for β_fPC1). Based on sex allocation theory, one might have expected that increased mating success as a male trades off with female reproductive success (β_fm), or vice versa (β_mf; Anthes et al., 2010; Charnov, 1979; Schärer, 2009).

As mentioned above, other studies on different simultaneous hermaphrodites did not find any cross-sex effects, while they did find that the mating system is mainly driven by the male function (B. glabrata: Anthes et al., 2010 and P. acuta: Pélissié et al., 2012, Pélissié et al., 2014; but see Janicke et al., 2015). An important difference between those studies and our study is that while they followed individuals for 3 to 5 days, we followed them for 56 days (8 weeks). This also allowed us to evaluate the effect of the chosen time frame on the results of such studies. Our data showed that from the start, the main effect, the significant β_mm, is already captured (also with the PCA approach: β_mPC2). This relationship became statistically stronger over time. In contrast, the negative effects only emerged much later into the experiment, which indicates that such effects on reproductive success may only be detectable in the longer run. Hence, short term experiments only take a snap shot of reproductive success (of virgins).

To conclude, we showed that L. stagnalis has a unilateral mating system and that sperm donors gain most reproductive success from repeated mating, even though they seem to lose some reproductive success in the long run (a cross-sex effect from mating frequently as a female). Our data also do reveal that these sperm recipients benefit indirectly from repeated mating, since their egg development and hatching success is higher. Our experiment therefore showed that the experimental time frame is very important for the quantification and interpretation of sexual selection measures, an insight that applies to any mating system with multiple mating.

1. 1. Anthes N
   2. David P
   3. Auld JR
   4. Hoffer JN
   5. Jarne P
   6. Koene JM
   7. Kokko H
   8. Lorenzi MC
   9. Pélissié B
   10. Sprenger D
   11. Staikou A
   12. Schärer L

   (2010) Bateman gradients in hermaphrodites: an extended approach to quantify sexual selection

   The American Naturalist 176:249–263.

   https://doi.org/10.1086/655218

   • PubMed
   • Google Scholar
2. 1. Anthes N
   2. Häderer IK
   3. Michiels NK
   4. Janicke T

   (2016) Measuring and interpreting sexual selection metrics: evaluation and guidelines

   Methods in Ecology and Evolution 11:.

   https://doi.org/10.1111/2041-210X.12707

   • Google Scholar
3. 1. Arnold SJ
   2. Duvall D

   (1994) Animal Mating Systems: a synthesis based on selection Theory

   The American Naturalist 143:317–348.

   https://doi.org/10.1086/285606

   • Google Scholar
4. 1. Arnold SJ

   (1994) Bateman's Principles and the Measurement of Sexual Selection in Plants and Animals

   The American Naturalist 144:S126–S149.

   https://doi.org/10.1086/285656

   • Google Scholar
5. 1. Baena ML
   2. Macías-Ordóñez R

   (2012) Phenology of scramble polygyny in a wild population of chrysomelid beetles: the opportunity for and the strength of sexual selection [corrected]

   PLoS One 7:e38315.

   https://doi.org/10.1371/journal.pone.0038315

   • PubMed
   • Google Scholar
6. 1. Bateman AJ

   (1948) Intra-sexual selection in Drosophila

   Heredity 2:349–368.

   https://doi.org/10.1038/hdy.1948.21

   • PubMed
   • Google Scholar
7. 1. Cain GL

   (1956) Studies on cross-fertilization and self-fertilization in Lymnaea stagnalis appressa say

   The Biological Bulletin 111:45–52.

   https://doi.org/10.2307/1539182

   • Google Scholar
8. 1. Charnov EL

   (1979) Simultaneous hermaphroditism and sexual selection

   PNAS 76:2480–2484.

   https://doi.org/10.1073/pnas.76.5.2480

   • PubMed
   • Google Scholar
9. 1. Collet JM
   2. Dean RF
   3. Worley K
   4. Richardson DS
   5. Pizzari T

   (2014) The measure and significance of Bateman's principles

   Proceedings of the Royal Society B: Biological Sciences 281:20132973.

   https://doi.org/10.1098/rspb.2013.2973

   • PubMed
   • Google Scholar
10. 1. Collet J
    2. Richardson DS
    3. Worley K
    4. Pizzari T

    (2012) Sexual selection and the differential effect of polyandry

    PNAS 109:8641–8645.

    https://doi.org/10.1073/pnas.1200219109

    • PubMed
    • Google Scholar
11. 1. Coutellec MA
    2. Caquet T

    (2011) Heterosis and inbreeding depression in bottlenecked populations: a test in the hermaphroditic freshwater snail Lymnaea stagnalis

    Journal of Evolutionary Biology 24:2248–2257.

    https://doi.org/10.1111/j.1420-9101.2011.02355.x

    • PubMed
    • Google Scholar
12. Book

    1. Darwin C

    (1871)

    The Descent of Man, and Selection in Relation to Sex

    London: Murray.

    • Google Scholar
13. 1. De Boer PACM
    2. Jansen RF
    3. Ter Maat A

    (1996) Copulation in the hermaphroditic snail Lymnaea stagnalis: a review

    Invertebrate Reproduction & Development 30:167–176.

    https://doi.org/10.1080/07924259.1996.9672542

    • Google Scholar
14. 1. Dewsbury DA

    (2005) The Darwin-Bateman paradigm in historical context

    Integrative and Comparative Biology 45:831–837.

    https://doi.org/10.1093/icb/45.5.831

    • PubMed
    • Google Scholar
15. Book

    1. Eberhard WG

    (1996)

    Female Control: Sexual Selection by Cryptic Female Choice

    New Jersey: Princeton University Press.

    • Google Scholar
16. 1. Escobar JS
    2. Auld JR
    3. Correa AC
    4. Alonso JM
    5. Bony YK
    6. Coutellec MA
    7. Koene JM
    8. Pointier JP
    9. Jarne P
    10. David P

    (2011) Patterns of mating-system evolution in hermaphroditic animals: correlations among selfing rate, inbreeding depression, and the timing of reproduction

    Evolution 65:1233–1253.

    https://doi.org/10.1111/j.1558-5646.2011.01218.x

    • PubMed
    • Google Scholar
17. 1. Evans JP
    2. Garcia-Gonzalez F

    (2016) The total opportunity for sexual selection and the integration of pre- and post-mating episodes of sexual selection in a complex world

    Journal of Evolutionary Biology 29:2338–2361.

    https://doi.org/10.1111/jeb.12960

    • PubMed
    • Google Scholar
18. 1. Fritzsche K
    2. Arnqvis G

    (2013) Homage to Bateman: sex roles predict sex differences in sexual selection

    Evolution 67:1926–1936.

    https://doi.org/10.1111/evo.12086

    • PubMed
    • Google Scholar
19. 1. Gowaty PA
    2. Kim YK
    3. Anderson WW

    (2012) No evidence of sexual selection in a repetition of Bateman's classic study of Drosophila melanogaster

    PNAS 109:11740–11745.

    https://doi.org/10.1073/pnas.1207851109

    • PubMed
    • Google Scholar
20. 1. Gowaty PA
    2. Kim YK
    3. Anderson WW

    (2013) Mendel's law reveals fatal flaws in Bateman's 1948 study of mating and fitness

    Fly 7:28–38.

    https://doi.org/10.4161/fly.23505

    • PubMed
    • Google Scholar
21. 1. Gowaty PA
    2. Steinichen R
    3. Anderson WW

    (2003) Indiscriminate females and choosy males: within- and between-species variation in Drosophila

    Evolution 57:2037–2045.

    https://doi.org/10.1111/j.0014-3820.2003.tb00383.x

    • PubMed
    • Google Scholar
22. 1. Hoffer JNA
    2. Schwegler D
    3. Ellers J
    4. Koene JM

    (2012) Mating rate influences female reproductive investment in a simultaneous hermaphrodite, Lymnaea stagnalis

    Animal Behaviour 84:523–529.

    https://doi.org/10.1016/j.anbehav.2012.06.002

    • Google Scholar
23. 1. Janicke T
    2. David P
    3. Chapuis E

    (2015) Environment-dependent sexual selection: Bateman's parameters under varying levels of food availability

    The American Naturalist 185:756–768.

    https://doi.org/10.1086/681128

    • PubMed
    • Google Scholar
24. 1. Janicke T
    2. Häderer IK
    3. Lajeunesse MJ
    4. Anthes N

    (2016) Darwinian sex roles confirmed across the animal kingdom

    Science Advances 2:e1500983.

    https://doi.org/10.1126/sciadv.1500983

    • PubMed
    • Google Scholar
25. 1. Jones AG
    2. Rosenqvist G
    3. Berglund A
    4. Avise JC

    (2005) The measurement of sexual selection using Bateman's principles: an experimental test in the sex-role-reversed pipefish Syngnathus typhle

    Integrative and Comparative Biology 45:874–884.

    https://doi.org/10.1093/icb/45.5.874

    • PubMed
    • Google Scholar
26. 1. Jones AG

    (2009) On the opportunity for sexual selection, the Bateman gradient and the maximum intensity of sexual selection

    Evolution 63:1673–1684.

    https://doi.org/10.1111/j.1558-5646.2009.00664.x

    • PubMed
    • Google Scholar
27. 1. Klug H
    2. Heuschele J
    3. Jennions MD
    4. Kokko H

    (2010) The mismeasurement of sexual selection

    Journal of Evolutionary Biology 23:447–462.

    https://doi.org/10.1111/j.1420-9101.2009.01921.x

    • PubMed
    • Google Scholar
28. 1. Knott KE
    2. Puurtinen M
    3. Kaitala V

    (2003) Primers for nine microsatellite loci in the hermaphroditic snail Lymnaea stagnalis

    Molecular Ecology Notes 3:333–335.

    https://doi.org/10.1046/j.1471-8286.2003.00444.x

    • Google Scholar
29. 1. Koene JM
    2. Montagne-Wajer K
    3. Roelofs D
    4. Ter Maat A

    (2009) The fate of received sperm in the reproductive tract of a hermaphroditic snail and its implications for fertilisation

    Evolutionary Ecology 23:533–543.

    https://doi.org/10.1007/s10682-008-9253-5

    • Google Scholar
30. 1. Koene JM
    2. Sloot W
    3. Montagne-Wajer K
    4. Cummins SF
    5. Degnan BM
    6. Smith JS
    7. Nagle GT
    8. ter Maat A

    (2010) Male accessory gland protein reduces egg laying in a simultaneous hermaphrodite

    PLoS One 5:e10117.

    https://doi.org/10.1371/journal.pone.0010117

    • PubMed
    • Google Scholar
31. 1. Koene JM
    2. Ter Maat A

    (2005) Sex role alternation in the simultaneously hermaphroditic pond snail Lymnaea stagnalis is determined by the availability of seminal fluid

    Animal Behaviour 69:845–850.

    https://doi.org/10.1016/j.anbehav.2004.07.012

    • Google Scholar
32. 1. Koene JM
    2. Ter Maat A

    (2007) Coolidge effect in pond snails: male motivation in a simultaneous hermaphrodite

    BMC Evolutionary Biology 7:212.

    https://doi.org/10.1186/1471-2148-7-212

    • PubMed
    • Google Scholar
33. 1. Lorenzi MC
    2. Sella G

    (2008) A measure of sexual selection in hermaphroditic animals: parentage skew and the opportunity for selection

    Journal of Evolutionary Biology 21:827–833.

    https://doi.org/10.1111/j.1420-9101.2008.01513.x

    • PubMed
    • Google Scholar
34. 1. Marie-Orleach L
    2. Janicke T
    3. Vizoso DB
    4. David P
    5. Schärer L

    (2016) Quantifying episodes of sexual selection: insights from a transparent worm with fluorescent sperm

    Evolution 70:314–328.

    https://doi.org/10.1111/evo.12861

    • PubMed
    • Google Scholar
35. 1. Meeker ND
    2. Hutchinson SA
    3. Ho L
    4. Trede NS

    (2007) Method for isolation of PCR-ready genomic DNA from zebrafish tissues

    BioTechniques 43:610–614.

    https://doi.org/10.2144/000112619

    • PubMed
    • Google Scholar
36. 1. Mitchell-Olds T
    2. Shaw RG

    (1987) Regression analysis of natural selection: statistical inference and biological interpretation

    Evolution 41:1149–1161.

    https://doi.org/10.1111/j.1558-5646.1987.tb02457.x

    • Google Scholar
37. 1. Mobley KB
    2. Jones AG

    (2013) Overcoming statistical bias to estimate genetic mating systems in open populations: a comparison of Bateman's principles between the sexes in a sex-role-reversed pipefish

    Evolution 67:646–660.

    https://doi.org/10.1111/j.1558-5646.2012.01819.x

    • PubMed
    • Google Scholar
38. 1. Nakadera Y
    2. Koene JM

    (2013) Reproductive strategies in hermaphroditic gastropods: conceptual and empirical approaches

    Canadian Journal of Zoology 91:367–381.

    https://doi.org/10.1139/cjz-2012-0272

    • Google Scholar
39. 1. Nakadera Y
    2. Blom C
    3. Koene JM

    (2014a) Duration of sperm storage in the simultaneous hermaphrodite Lymnaea stagnalis

    Journal of Molluscan Studies 80:1–7.

    https://doi.org/10.1093/mollus/eyt049

    • Google Scholar
40. 1. Nakadera Y
    2. Swart EM
    3. Hoffer JN
    4. den Boon O
    5. Ellers J
    6. Koene JM

    (2014b) Receipt of seminal fluid proteins causes reduction of male investment in a simultaneous hermaphrodite

    Current Biology 24:859–862.

    https://doi.org/10.1016/j.cub.2014.02.052

    • PubMed
    • Google Scholar
41. 1. Nakadera Y
    2. Mariën J
    3. Van Straalen NM
    4. Koene JM

    (2017) Multiple mating in natural populations of a simultaneous hermaphrodite, Lymnaea stagnalis

    Journal of Molluscan Studies 83:56–62.

    https://doi.org/10.1093/mollus/eyw043

    • Google Scholar
42. 1. Parker GA
    2. Birkhead TR

    (2013) Polyandry: the history of a revolution

    Philosophical Transactions of the Royal Society B: Biological Sciences 368:20120335.

    https://doi.org/10.1098/rstb.2012.0335

    • PubMed
    • Google Scholar
43. 1. Parker GA

    (1970) Sperm competition and its evolutionary consequences in the insects

    Biological Reviews 45:525–567.

    https://doi.org/10.1111/j.1469-185X.1970.tb01176.x

    • Google Scholar
44. 1. Pischedda A
    2. Rice WR

    (2012) Partitioning sexual selection into its mating success and fertilization success components

    PNAS 109:2049–2053.

    https://doi.org/10.1073/pnas.1110841109

    • Google Scholar
45. 1. Power DJ
    2. Holman L

    (2014) Polyandrous females found fitter populations

    Journal of Evolutionary Biology 27:1948–1955.

    https://doi.org/10.1111/jeb.12448

    • PubMed
    • Google Scholar
46. 1. Puurtinen M
    2. Emily Knott K
    3. Suonpää S
    4. Nissinen K
    5. Kaitala V

    (2007) Predominance of outcrossing in Lymnaea stagnalis despite low apparent fitness costs of self-fertilization

    Journal of Evolutionary Biology 20:901–912.

    https://doi.org/10.1111/j.1420-9101.2007.01312.x

    • PubMed
    • Google Scholar
47. 1. Pélissié B
    2. Jarne P
    3. David P

    (2012) Sexual selection without sexual dimorphism: Bateman gradients in a simultaneous hermaphrodite

    Evolution 66:66–81.

    https://doi.org/10.1111/j.1558-5646.2011.01442.x

    • PubMed
    • Google Scholar
48. 1. Pélissié B
    2. Jarne P
    3. Sarda V
    4. David P

    (2014) Disentangling precopulatory and postcopulatory sexual selection in polyandrous species

    Evolution 68:1320–1331.

    https://doi.org/10.1111/evo.12353

    • PubMed
    • Google Scholar
49. 1. Rose E
    2. Paczolt KA
    3. Jones AG

    (2013a) The contributions of premating and postmating selection episodes to total selection in sex-role-reversed Gulf pipefish

    The American Naturalist 182:410–420.

    https://doi.org/10.1086/671233

    • PubMed
    • Google Scholar
50. 1. Rose E
    2. Paczolt KA
    3. Jones AG

    (2013b) The effects of synthetic estrogen exposure on premating and postmating episodes of selection in sex-role-reversed Gulf pipefish

    Evolutionary Applications 6:1160–1170.

    https://doi.org/10.1111/eva.12093

    • PubMed
    • Google Scholar
51. 1. Schärer L
    2. Pen I

    (2013) Sex allocation and investment into pre- and post-copulatory traits in simultaneous hermaphrodites: the role of polyandry and local sperm competition

    Philosophical Transactions of the Royal Society B: Biological Sciences 368:20120052.

    https://doi.org/10.1098/rstb.2012.0052

    • PubMed
    • Google Scholar
52. 1. Schärer L

    (2009) Tests of sex allocation theory in simultaneously hermaphroditic animals

    Evolution 63:1377–1405.

    https://doi.org/10.1111/j.1558-5646.2009.00669.x

    • PubMed
    • Google Scholar
53. Book

    1. Shuster SM
    2. Wade MJ

    (2003)

    Mating Systems and Strategies: Monographs in Behavior and Ecology

    Princeton: Princeton University Press.

    • Google Scholar
54. 1. Turnell BR
    2. Shaw KL

    (2015) High opportunity for postcopulatory sexual selection under field conditions

    Evolution 69:2094–2104.

    https://doi.org/10.1111/evo.12721

    • PubMed
    • Google Scholar
55. 1. Van Der Steen WJ
    2. Van Den Hoven NP
    3. Jager JC

    (1969) A method for breeding and studying freshwater snails under continuous water change, with some remarks on growth and reproduction in Lymnaea stagnalis (L.)

    Netherlands Journal of Zoology 19:131–139.

    https://doi.org/10.1163/002829669X00170

    • Google Scholar
56. 1. Van Duivenboden YA
    2. Pieneman AW
    3. Ter Maat A

    (1985) Multiple mating suppresses fecundity in the hermaphrodite freshwater snail Lymnaea stagnalis: a laboratory study

    Animal Behaviour 33:1184–1191.

    https://doi.org/10.1016/S0003-3472(85)80179-2

    • Google Scholar
57. 1. Van Duivenboden YA
    2. Ter Maat A

    (1988)

    Mating behaviour of Lymnaea stagnalis

    Malacologia 28:23–64.

    • Google Scholar
58. 1. Wacker S
    2. Amundsen T
    3. Forsgren E
    4. Mobley KB

    (2014) Within-season variation in sexual selection in a fish with dynamic sex roles

    Molecular Ecology 23:3587–3599.

    https://doi.org/10.1111/mec.12826

    • PubMed
    • Google Scholar

Author details

1. Jeroen NA Hoffer

   Section of Animal Ecology, Department of Ecological Science, Faculty of Earth and Life Sciences, Vrije Universiteit, Amsterdam, The Netherlands

   Contribution

   JNAH, Conceptualization, Data curation, Formal analysis, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing

   Competing interests

   The authors declare that no competing interests exist.

2. Janine Mariën

   Section of Animal Ecology, Department of Ecological Science, Faculty of Earth and Life Sciences, Vrije Universiteit, Amsterdam, The Netherlands

   Contribution

   JM, Supervision, Methodology, Writing—original draft

   Competing interests

   The authors declare that no competing interests exist.

3. Jacintha Ellers

   Section of Animal Ecology, Department of Ecological Science, Faculty of Earth and Life Sciences, Vrije Universiteit, Amsterdam, The Netherlands

   Contribution

   JE, Resources, Supervision, Methodology, Writing—original draft, Project administration, Writing—review and editing

   Competing interests

   The authors declare that no competing interests exist.

4. Joris M Koene

   Section of Animal Ecology, Department of Ecological Science, Faculty of Earth and Life Sciences, Vrije Universiteit, Amsterdam, The Netherlands

   Contribution

   JMK, Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   joris.koene@vu.nl

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0001-8188-3439

• 2,565

  views

• 270

  downloads

• 17

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.