SPERM COMPETITIVE ABILITY EVOLVES IN RESPONSE TO EXPERIMENTAL ALTERATION OF OPERATIONAL SEX RATIO

Authors


Abstract

In naturally polygamous organisms such as Drosophila, sperm competitive ability is one of the most important components of male fitness and is expected to evolve in response to varying degrees of male–male competition. Several studies have documented the existence of ample genetic variation in sperm competitive ability of males. However, many experimental evolution studies have found sperm competitive ability to be unresponsive to selection. Even direct selection for increased sperm competitive ability has failed to yield any measurable changes. Here we report the evolution of sperm competitive ability (sperm defense-P1, offense-P2) in a set of replicate populations of Drosophila melanogaster subjected to altered levels of male–male competition (generated by varying the operational sex ratio) for 55–60 generations. Males from populations with female-biased operational sex ratio evolved reduced P1 and P2, without any measurable change in the male reproductive behavior. Males in the male-biased regime evolved increased P1, but there was no significant change in P2. Increase in P1 was associated with an increase in copulation duration, possibly indicating greater ejaculate investment by these males. This study is one of the few to provide empirical evidence for the evolution of sperm competitive ability of males under different levels of male–male competition.

Females of several species of animals mate more than once and store sperm from different males in their genital tract or some specialized storage organ (Lefevre and Jonsson 1962). This leads to “sperm competition,” where sperm from different males compete in the female body for limited opportunity of fertilization (Parker 1970; Wedell et al. 2002). Sperm competitive ability is typically defined and quantified in terms of two components—defense and offense (Parker 1970; Boorman and Parker 1976; Clark et al. 2000; Friberg et al. 2005; Bjork et al. 2007). When a given male is the first mate of a female, the proportion of progeny sired by him gives a measure of sperm defense (P1) and the probability that the female will undergo a second mating is termed fidelity. However, if a male mates with an already mated female, the proportion of progeny fathered by him is termed P2. Together with the probability of mating with an already mated female, P2 is the measure of sperm offense. Males can potentially maximize their sperm competitive ability either by manipulating sperm/ejaculate physiology or by changing female behavior and/or physiology (Snook 2005). Because most of the sexually reproducing species are promiscuous, sperm competition is expected to be widespread (Birkhead and Møller 1998; Birkhead and Pizzari 2002). In addition, sperm competition is predicted to be an important factor in the process of speciation (Parker and Partridge 1998; Simmons 2001) because of its role in driving intersexual conflict (Stockley 1997; Civetta and Clark 2000; Rice 2000; Friberg et al. 2005). Sperm competition and the resulting postcopulatory sexual selection also have the potential to significantly alter male reproductive behavior (Simmons et al. 1993; Cook and Wedell 1996; Gage and Barnard 1996; Wedell and Cook 1999; Bretman et al. 2009, 2010; Nandy and Prasad 2011) and physiology (Wolfner 1997) thereby playing an important role in the evolution of male reproductive and life-history traits.

Empirical studies using model organisms have been a key source of our understanding of the process of evolution of sperm competitive ability. Several studies have documented genetic and phenotypic variation in this trait. Most quantitative genetic studies have shown substantial genetic variation in all components of defense and offense abilities in fruit flies (Clark et al. 1995; Hughes 1997; Civetta and Clark 2000; Friberg et al. 2005; Hughes and Leips 2006). However, while Hughes (1997) found this variance to be mainly nonadditive, Friberg et al. (2005) reported significant amount of additive genetic variance within the population of Drosophila melanogaster. In addition, significant male × male (Clark et al. 2000) and male × female (Clark and Begun 1998; Clark et al. 1999; Miller and Pitnick 2002) interactions were found to influence the outcome of sperm competition.

Despite the existence of strong fitness consequences of sperm competitive ability, attempts to observe evolution of this trait have produced inconsistent results across different studies. When males were allowed to evolve against a fixed target female phenotype (male limited evolution), Rice (1996, 1998) observed increase in offense and defense ability in males. However, using the same ancestral populations, two different studies failed to find any such evolution. Bjork et al. (2007) directly selected for both defense and offense but could not find the predicted response. Another study using the similar “male limited evolution” approach could not reproduce Rice's results (Rice 1996, 1998), despite observing an increase in male fitness (Jiang et al. 2011). Alteration in the levels of sexual selection by changing the operational sex ratio was found to have no significant effect on male reproductive traits (Wigby and Chapman 2004) and sperm competitive ability (Michalczyk et al. 2011). However, removal of sexual selection through experimentally enforced monogamy led to the evolution of male reproductive traits and sperm competitive ability (Hosken et al. 2001; Simmons and García-González 2008). In a relatively recent study, decrease in sperm competitive ability was also observed in mice populations subjected to enforced monogamy (Firman and Simmons 2011).

Here we report the results from an experimental evolution study using a long-term laboratory adapted population of D. melanogaster. We held populations at different operational sex ratios and then assayed the evolutionary response of these populations in terms of different components of sperm competitive ability of the males—defense and offense. Alteration in operational sex ratio is predicted to generate a range of intensities of male–male competition—high (male biased), medium (equal sex ratio), and low (female biased). After 55–60 generations of selection, we quantified sperm defense and offense abilities of males (P1 and P2, respectively) under single-pair and group conditions. We also observed different components of male behavior.

Materials and Methods

MAINTENANCE OF POPULATIONS

The selected populations were derived from a long-term laboratory adapted population of D. melanogaster—LHst (Prasad et al. 2007). LHst is a derivative of LH base population (see Chippindale et al. 2001 for details of the maintenance regime) having a recessive autosomal marker—“scarlet eye” (also see below for a comparison between the two base populations). The LHst population is regularly backcrossed into LH background to ensure that the two populations have identical genetic backgrounds and differ only with respect to the eye color marker. LH is wild type red eyed and LHst is scarlet eyed. Three sex ratio regimes were created—M, male biased (three males : one female); C, equal sex ratio; and F, female biased (one male : three females). Each regime had three replicate populations (M1–3, C1–3, and F1–3). Three replicate populations (C1–3) were derived from LHst and maintained under equal sex ratio for five generations. Each of these populations was then used to derive two other selection regimes—M and F. Thus, populations bearing the same numerical subscript shared a common ancestry and were more closely related to each other compared to populations bearing different numerical subscripts. For example, M1 is more closely related to C1 and F1 than to M2. In addition, during regular maintenance, replicate populations bearing the same numerical subscript were always handled together. Hence replicates bearing the same numerical subscripts were treated as statistical “Blocks” in the analysis. The whole experiment consisted of three statistical blocks (Blocks 1, 2, and 3). As they were derived from LHst population, all of the nine populations (M1–3, C1–3, and F1–3) carried the recessive scarlet-eyed marker. The populations were maintained on 14-day discrete generation cycle, at 25°C and 60% relative humidity (RH) under 12 hours–12 hours light/dark cycle, on standard cornmeal–molasses–yeast food in standard vials (90-mm length × 30-mm diameter). Flies were grown under controlled larval density (140–160 per 8–10 mL food in each vial). Every generation flies were collected as virgins and held in single sex vials (eight individuals per vial) till the 12th day after egg collection. They were then combined following the respective sex ratio regimes in food vials provisioned with 0.47 mg per female of live yeast (adult competition vials). Adult density in adult competition vials was controlled at 32 individuals per vial (M: 24 males and 8 females; C: 16 males and 16 females; F: 8 males and 24 females). Two days later, flies were transferred to oviposition vials and allowed to oviposit for 18 hours. The eggs laid during this window were used to start the next generation. Adult density in the precompetition vials and amount of yeast available in adult-competition vials were kept constant across all the populations. The effective population size (Ne) of all the populations was maintained at 450 following the definition provided by Crow and Kimura (1970), Ne = 4NmNf / (Nm + Nf), where Nm and Nf are number of males and females in the population (see Supporting Information for details).

To equalize any nongenetic parental effect across all the selection regimes, all the populations were passed through one generation of common rearing called “standardization” (see Supporting Information). The flies for the assays were generated from standardized populations.

GENERATION OF EXPERIMENTAL FLIES

All the experimental flies were generated under controlled larval density and standard culture conditions (25°C, 60–80% RH, 12 hours–12 hours light / dark cycle). For each of the nine selected populations, 150 eggs were cultured in 8–10 mL of cornmeal-molasses food per vial. On 10th day after egg collection, adult flies started emerging. Males were collected as very young (≤6 hours after eclosion) virgins during the peak of their eclosion rhythm under light CO2 anesthesia (<3 min exposure). Males were held in single sex vials, at the density of 10 per vial (for group design) and one per vial (for single-pair design) till the mating trials. Two-day-old virgin flies were used to setup the mating trials.

The LH and LHst flies were generated under similar conditions—controlled larval density (140–160 per 10 mL cornmeal molasses food per vial). LHst females were collected as virgins in the same way described earlier and held at a density of eight per vial (for group design) and one per vial (for single-pair design) in single-sex vials for 20 days before the mating trial. LH males, the competitors were collected in the same way and held in single-sex vials at a density of 10 per vial (for group design) and one per vial (for single-sex design). LH males and LHst females were also held for 2 days before the mating trials. The assays were done in the two designs—group and single pair.

GROUP DESIGN

For sperm defense (P1) assay, selection regime males (from one of the nine populations), in groups of 10, were combined with eight virgin LHst females in a mating vial. Nine to 10 such replicate vials were setup for each of the nine populations (M1–3, C1–3, and F1–3). The males were discarded after one round of mating and females were held back in the same vial. Single mating was ensured by direct manual observation and the limited time of exposure (< 60–75 min). This limited exposure is sufficient for only a single round of mating per female and has been successfully adopted previously (Nandy et al. 2012). Females were separated from the males under light CO2-anesthesia and were allowed half an hour to recover from the effect. Following this, the red-eyed (LH) competitor males were introduced into the vials in groups of 10. These vials were left undisturbed for the next 18–20 hours. After this exposure time, males were discarded and females were transferred individually to oviposition test tubes (dimension: 12 mm × 75 mm) provisioned with food under anesthesia. A window of 18 hours was allowed for oviposition, following which the females were discarded and the test tubes were incubated under standard conditions (25°C, 12 hours–12 hours light / dark cycle, 60% RH) for 12 days before freezing them. The progeny was scored for eye color and proportion of scarlet-eyed progeny was taken as a measure of P1 of the selection regime males. Each female in a vial was scored for P1. Females who produced only scarlet-eyed progeny (i.e., P1 = 1) very likely failed to remate and therefore were removed from the analyses. Finally, a vial mean was calculated for each replicate vial using P1 values from all the females (excluding females with P1 = 1) in that vial. Vial means were used as the unit of analysis.

For sperm offense (P2) assay, the design was identical to that of the P1 assay described earlier, except that the first males were taken as LH males and the second males were selection regime males. LHst females were first allowed a single mating (manually observed) with LH males (eight females and 10 males per vial), following which the males were separated and discarded. The second male (i.e., males from one of the nine populations) were introduced after allowing the females to recover from the effect of anesthesia for half an hour to 8–10 vials per population were setup for this part of the experiment. The second males interacted with the females for 18–20 hours, after which they were separated. Females were then transferred to the oviposition test tubes as mentioned in the previous section and allowed to oviposit for 18 hours. Similar to the P1 assay, the proportion of scarlet-eyed progeny was taken as the measure of P2 of the selection regime males. Each female in a vial was scored for P2. Females who produced only red-eyed progeny (i.e., P2 = 0) were removed from the analyses as they very likely failed to remate with the selection line males. Similar to the P1 assay, vial means were calculated for each vial, using P2 values from all the females in that vial (excluding females with P2 = 0). Vial means were used as the unit of analysis. In each vial, proportion of females with P2 > 0 was taken as a measure of remating success (proportion of males that successfully remated with the test-female). Remating success was analyzed using vial means.

SINGLE-PAIR DESIGN: SPERM DEFENSE AND RELATED BEHAVIORS

One virgin LHst female was combined with a single selection regime male in an eight-dram vial provisioned with food. The cotton plug of the vial was push deep into the vial to adjust the space available to the flies to roughly 30 mm × 30 mm. The pair was observed till they finished mating and mating latency (time taken by a pair to start mating after being put together in mating vial) and copulation duration (CD; duration for which the pair remained in copula) were recorded. After the first mating, female was quickly sorted using CO2-anesthesia and the male was discarded. The female was held back into the vial, allowed a recovery time of half an hour before introducing the second male (red eyed LH). Following this the vial is left undisturbed for 44 hours during which competitor LH male mated with the female. This 44-hour exposure closely matched the normal male–female interaction time for the selected populations. After this exposure window, the male was discarded and the female was transferred to test tube provisioned with food and was allowed an oviposition window of 18 hours. The progeny produced during this window was allowed to emerge and then they were scored for their eye color marker. The proportion of scarlet progeny was taken as an estimate of P1 of the selection regime male. A total of 39–44 males from each of the population were assayed for P1. Vials in which females failed to remate (P1 = 1) were excluded for the P1 analysis. Final sample size for P1 analysis was n = 35–43 for each population. Because in this experiment each vial contained a single pair, each vial yielded a single value of mating latency, CD, and P1. This vial value was used as the unit of analysis. Proportion of females that did not remate with the LH males (i.e., produced only scarlet-eyed progeny, P1 = 1) was noted and taken as a measure of remating fidelity of the selection regime males. Remating fidelity was analyzed using population means.

STATISTICAL ANALYSIS

P1 (group and single-pair design), P2, CD, mating latency, and remating success were analyzed using two-factor mixed model analyses of variance (ANOVA) with selection regime as fixed factor crossed with random blocks. We calculated mean remating fidelity for each of the populations and analyzed the data using a two-factor ANOVA with selection regime as fixed factor and block as random factor. All multiple comparisons were done using Tukey's honestly significant difference (HSD). Except mean remating fidelity, all the analyses were done using JMP 10 (SAS Institute, Cary, NC). Mean remating fidelity was analyzed using population mean as unit of analysis in STATISTICA 10 for Windows (StatSoft Inc., Tulsa, OK). Both platforms use Satterthwaite's method of denominator synthesis. Level of significance (α) was taken as 0.05 for all the tests done.

Results

GROUP DESIGN

In the group design, we found a significant effect of selection regime on sperm defense score (P1) as well as sperm offense score (P2) of the selected males (Table 1). Both analyses were consistent across blocks. Multiple comparison using Tukey's HSD showed that M males had the highest P1, which was significantly different from that of F males. C males had intermediate P1, not significantly different from either F or M males (Fig. 1A). F males were found to have significantly lower P2 compared to both C and M males (Fig. 1B). There was no significant effect of selection regime on the remating success (Table 1).

Table 1. Summary of the two-factor mixed model analyses of variance (ANOVA) on P1, P2, and remating success data from the group design assays treating selection regime as fixed factor crossed with random blocks. Vial means were treated as unit of analysis. P-values in bold indicate significant effects
TraitSourceSSMS NumDF NumDF DenF ratioProb > F
P1Selection regime0.0990.0492744.4370.015
 Block0.0170.0092740.7810.462
 Selection regime × block0.0250.0064740.5610.692
P2Selection regime0.0570.0282774.1170.020
 Block0.0160.0082771.1690.316
 Selection regime × block0.0040.0014770.1570.959
Remating successSelection regime0.0120.0062830.3990.672
 Block0.0580.0292831.9380.150
 Selection regime × block0.0140.0034830.2290.921
Figure 1.

Effect of selection regime on (A) P1 and (B) P2 of the selected males competed against ancestral males (LH). Points not sharing common letter are significantly different (determined using Tukey's HSD).

SINGLE-PAIR DESIGN

In the single-pair design assay, selection regime was found to have a significant effect on CD and P1 (Table 2). Tukey's HSD suggested that M males mated for significantly longer duration compared to both F and C males (Fig. 2A). CD for F males was longer than C males but the difference was not significant (Fig. 2A). Multiple comparisons on the P1 data revealed a significant difference between M and F males with M males having the highest P1 and F males having the least P1 (Fig. 2B). C males were again found to have intermediate P1 (Fig. 2B). Neither M nor F males had P1 significantly different from that of C males. P1 results were thus consistent across the group and single-pair design assays (see Figs. 1A and 2B).

Table 2. Summary of two-factor mixed model analysis of variance (ANOVA) on the copulation duration, P1, mating latency, and remating fidelity data from single-pair design assays treating selection regime as fixed factor crossed with random blocks. Except for remating fidelity, the other three traits were analyzed using vial values as the unit of analysis. For remating fidelity, population means were used as the unit of analysis. P-values in bold indicate significant effects
TraitSourceSSMS NumDF NumDF DenF ratioProb > F
Copulation durationSelection regime387.548193.774242532.163<0.001
 Block76.02438.01224256.3090.002
 Selection regime × block48.79812.19944252.0250.090
P1Selection regime0.1370.06823503.2250.041
 Block0.0280.01423500.6680.513
 Selection regime × block0.0220.00543500.2550.907
Mating latencySelection regime28.46114.23024250.9790.377
 Block92.52746.26424253.1810.043
 Selection regime × block33.6668.41744250.5790.678
Remating fidelitySelection regime0.0030.002243.2840.143
 Block0.0010.0003240.5520.614
Figure 2.

Results of single-pair trials: effect of selection regime on (A) copulation duration and (B) P1 of the selected males. Points not sharing common letter are significantly different (determined using Tukey's HSD).

No significant effect (Table 2) of selection regime on mating latency (mean'± SE, F: 0.04'± 0.01; C: 0.09'± 0.01; M: 0.05'± 0.01) and remating fidelity (mean'± SE, F: 0.04'± 0.01; C: 0.09'± 0.01; M: 0.05'± 0.01) was found.

Discussion

Increased male–male competition (male-biased sex ratio in our regime) is likely to select for higher sperm competitive ability. Under relaxation of such competition (female-biased sex ratio in our regime), the traits relevant to sperm competitive ability are expected to degenerate if there is cost of bearing them. Theories of sperm competition predict the evolution of several male traits under the influence of such selection (Parker 1970; Simmons 2001). Empirical studies in a range of taxa also suggest that male behavior and/or physiology can potentially evolve in response to sperm competition experienced by the males (Birkhead and Møller 1998; Simmons 2001). Male reproductive behavior and/or physiology, along with sperm competitive ability have been shown to evolve under various conditions using a number of species (Rice 1996, 1998; Hosken and Ward 2001; Hosken et al. 2001; Pitnick et al. 2001; Simmons and García-González 2008; Firman and Simmons 2011). Here we show the evolution of sperm competitive ability of males under the altered operational sex ratio.

Preadult survivorship differences of the progeny of males across the three selection regimes can potentially lead to differences in the P1 and P2 measures in our assay (Gilchrist and Partridge 1997; García-González 2008). We found no significant difference in the preadult survivorship (egg to adult survival rate) of the M, C, and F flies in a separate experiment (proportion survived, mean'± SE, M: 0.90'± 0.009, C: 0.90'± 0.009, F: 0.87'± 0.009, see Supporting Information). In addition, LHst is an outbred base population with preadult survivorship of about 90%. Hence, the progeny of M, C, and F males with females of the LHst base population are very unlikely to have differential preadult survivorship. Therefore, the observed differences in P1 and P2 of the three selection regimes are representative of their sperm competitive ability. Our results suggest that males from populations with female-biased operational sex ratio (F) evolved reduced sperm competitive ability in terms of both sperm defense and offense. These differences can be caused either by qualitative change in the male ejaculate (Ram and Wolfner 2007; Sirot et al. 2011) or by changes in different components of male-mating behavior, such as quantity of ejaculate invested per copulation (Bretman et al. 2009; Nandy and Prasad 2011). However, we did not observe any significant decline in CD, ability to induce fidelity, and remating success of F males. In a separate assay, we also observed F males to be equally active in courting the females (B. Nandy et al., unpubl. data). The lack of these behavioral changes indicates likely qualitative changes (discussed further later) in the ejaculate of F males. In populations with male-biased sex ratio (M), males evolved increased P1 without inducing higher fidelity. M males did not have increased offense ability (P2).

Our results are largely in line with the predictions of theories of sperm competition (Parker 1970). Under F regime, due to the abundance of mating opportunities, intensity of competition between males, including sperm competition (risk and intensity) can be expected to be low. Such relaxation of selection is likely to cause the loss (or deterioration) of the costly traits that are otherwise advantageous in male–male competition. Male ejaculate contains accessory gland proteins or Acps (Wolfner 2002), which have a substantial manufacturing cost (Chapman and Edward 2011). Most of these proteins have been shown to have post-copulatory effects on females and their evolution is believed to be affected by intersexual conflict and sperm competition (Civetta and Clark 2000; Wolfner 2009). It is possible that, due to the relaxation of the intensity of sperm competition, F males have evolved Acps that are qualitatively and/or quantitatively different. Such changes are likely to make their ejaculate less competitive, causing the decline in P1 and P2 of F males, as observed in our experiment. However, because we did not quantify the ejaculate components in our assay, at this point we cannot confirm this hypothesis. However, it is important to note here that we found no obvious decline in behaviors associated with sperm competitive ability in the F males.

Alternatively, the observed decline in the sperm competitive ability in F males could possibly be due to differential genetic drift in the F populations during the course of selection leading to inbreeding-like effects (Charlesworth and Charlesworth 1987). We attempted to equalize Ne across the three regimes to avoid differential inbreeding. However, the very nature of the mating system and sexual selection in the system (multiple mating by females, last male sperm precedence, etc.) can potentially complicate this equalization method. Alternative methods of calculating Ne (accounting for multiple mating, sperm precedence) has been suggested by Rice and Holland (2005). Following this method, even after considering sperm precedence and female multiple mating, the Ne in all the regimes was greater than 350 (see Supporting Information). Previous studies have shown that laboratory selection experiments, such as ours, where Ne > 100 are very unlikely to suffer effects of inbreeding and drift within the timescales of this study, that is, 50–60 generations (Rice and Holland 2005; Snook et al. 2009; also see Supporting Information). In addition, the regime that experienced smallest Ne (see Supporting Information) in our experiment was M and not F. However, M populations showed increase in components of fitness, thereby indicating the absence of any confounding effect of inbreeding in the experiment.

Males of M regime were found to have evolved increased P1 (defense) even though they did not induce higher fidelity to their mates. However, we found M males to copulate with virgin females for significantly longer duration. Previous studies using the same (Nandy and Prasad 2011) as well as different (Bretman et al. 2009, 2010) populations of D. melanogaster have shown a positive correlation between CD and P1. In absence of further quantitative data on sperm and/or Acps, it is, however, not clear whether this is caused by increased transfer of sperm or other components of ejaculate or both (Gilchrist and Partridge 2000; Sirot et al. 2011).

We did not find any change in P2 of the M males. Although it is difficult to predict the reason for this, we cite a few possibilities. First, even in our baseline population, according to previous studies (Boorman and Parker 1976; Friberg et al. 2005; Bjork et al. 2007), P2 is usually very high—close to 0.8 (i.e., 80% of the progeny is sired by the second male). Further increase in P2 might be too costly for males to evolve even under M regime. Second, under severely high risk and intensity of sperm competition (which can be expected to represent our M regime) it might not be worth investing more in P2 (Enqvist and Reinhold 2006).

A number of studies have reported the evolution of decreased sperm competitive ability and testes size in response to complete removal of sexual selection by experimentally enforced monogamy (relative to polygamy) in a variety of model organisms (Hosken et al. 2001; Pitnick et al. 2001; Simmons and García-González 2008; Firman and Simmons 2011). However, Wigby and Chapman (2004) altered levels of sexual selection by altering operational sex ratio in D. melanogaster but did not find the evolution of testis size. Michalczyk et al. (2011) maintained populations of Tribolium castaneum under male-biased and female-biased sex ratio and found that while the males from the male-biased regime had higher mating success compared to males from the female-biased regime, their sperm competitive ability did not differ significantly. However, in our study, sperm competitive ability of the males evolved in response to manipulation of operational sex ratio. Although Rice (1996, 1998) observed evolution of sperm competitive ability in response to “male limited selection,” Jiang et al. (2011) did not find any such evidence following the same approach of experimental evolution. Direct selection of sperm competitive ability by Bjork et al. (2007) did not cause any change in the offense and defense ability. However, Friberg et al. (2005) detected small but measurable additive variation with respect to all components of offense and defense. Bjork et al. (2007) suggested that complex interactions with rival males and females could possibly explain their failure to find response to the selection. Jiang et al. (2011) suggested yet another novel explanation that included a possible loophole in the experimental evolution studies using long-term laboratory adapted populations. Long-term laboratory domestication under a controlled condition (specifically for LH population), leading to strong directional selection on the relevant traits, can potentially erode additive genetic variation (Jiang et al. 2011). Our results at least to some extent contradict these two studies (Bjork et al. 2007; Jiang et al. 2011). We did find significant effect of selection regime on sperm competitive abilities of males after 60 generations of selection, indicating the presence of substantial additive genetic variation in the ancestral population. There are several differences between our study and the two (Bjork et al 2007; Jiang et al. 2011) mentioned earlier, potentially explaining such different results. (a) Given that the outcome of sperm competition depends on the complex male–male and male–female interactions (Clark 2000; Miller and Pitnick 2002), direct selection on only males either for P1/P2 or fitness might not lead to the expected selection response (Bjork et al. 2007). Our selection, however, was of a more multifarious type, potentially selecting for a suite of traits in both the sexes. (b) Earlier studies (Bjork et al. 2007; Jiang et al. 2011) looked only at increasing defense and offense abilities. Our F regime (female-biased sex ratio) relaxed the selection of male–male competition and thus the population could evolve in the direction (i.e., decrease of defense and offense) not studied in these previous experiments. As discussed earlier, it is possible that evolution in one direction (i.e., increased P2) is too costly to evolve, whereas evolution in the other direction is possible.

In summary, we have shown in this study that sperm competitive ability (both defense and offense), at least in fruit flies, can potentially undergo adaptive evolution in response to the changes in the operational sex ratio. Due to the cost of maintenance, traits related to sperm competitive ability undergo degeneration upon relaxation of competition among males. We have also shown increase in sperm defense ability under male-biased selection condition that represents increased intensity of competition among males.

ACKNOWLEDGMENTS

We thank M. Mallet and S. Bedhomme for helpful comments on the manuscript. We thank A. Joshi, A. K. Chippindale, and S. Bedhomme for helpful discussions about statistical analyses. We also thank the handling editor and the two anonymous reviewers for their helpful remarks that greatly improved the clarity of the manuscript. We thank Indian Institute of Science Education and Research Mohali and Department of Science and Technology, Govt. of India for continued funding for the project. We are also thankful to the A. Gupta, M. Samant, S. Sen, and V. Shenoy for their help in the lab. BN and VG thank the Council for Scientific and Industrial Research, Govt. of India, for financial assistance in the form of a Senior Research Fellowship and Junior Research Fellowship respectively. ASZ thanks the Department of Science and Technology, Govt. of India, for financial assistance in the form of Kishore Vaigyanik Protsahan Yojana fellowship. PC thanks Department of Science and Technology, Govt. of India for INSPIRE scholarship.

Ancillary