SEARCH

SEARCH BY CITATION

Keywords:

  • female remating;
  • good genes;
  • indirect benefits;
  • sexual conflict;
  • sexual selection

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Sexual conflict theory is based on the observation that females of many species are harmed through their interactions with males. Direct harm to females, however, can potentially be counterbalanced by indirect genetic benefits, where females make up for a reduction in offspring quantity by an increase in offspring quality through a generic increase in offspring fitness (good genes) and/or one restricted to the context of sexual selection (sexy sons). Here, we quantify the magnitude of the good genes mechanism of indirect benefits in a laboratory-adapted population of Drosophila melanogaster. We find that despite high-standing genetic variance for fitness, females gain at most only a modest benefit through the good genes form of indirect benefits – far too little to counterbalance the direct cost of male-induced harm.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Females are harmed by their interactions with their mates in a wide diversity of taxa (McKinney et al., 1983; Arnqvist, 1992; Crudgington & Siva-Jothy, 2000; Hosken et al., 2001; Stutt & Siva-Jothy, 2001; Arnqvist & Rowe, 2002). This phenomenon has been studied most extensively in the Drosophila melanogaster laboratory model system (Cohet & David, 1976; Partridge et al., 1987; Chapman et al., 1995; Rice, 1996, 2000; Partridge & Hurst, 1998; Holland & Rice, 1999; Chapman, 2001). Observations such as these form the foundation for the field of sexual conflict. However, indirect genetic benefits (through increased quality of offspring) can potentially offset the observed direct harm that males cause to their mates (Cordero & Eberhard, 2003). Theoretical work has indicated that such indirect benefits are likely to be small (Kirkpatrick & Barton, 1997; Cameron et al., 2003; Chapman et al., 2003). In agreement with this theoretical work, a comprehensive review of data available on good genes effects revealed that it has, on average, a small effect on offspring fitness (∼1.5%; Moller & Alatalo, 1999), but some studies suggest that indirect benefits can at least sometimes equal or outweigh direct costs (e.g. Head et al., 2005; Rundle et al., 2007; Priest et al., 2008).

A recent experiment in the D. melanogaster laboratory model system (Stewart et al., 2005; using the outbred LHM base population) indicates that indirect benefits do not counterbalance direct, male-induced harm to females in this environment, despite substantial genetic variation for fitness in both sexes (Rice et al., 2006). Orteiza et al. (2005) specifically tested for the presence of an ‘attractive sons’ benefit, in the context of remating (which is harmful to females in the LHM population, Kuijper et al., 2006) by comparing the fitness of sons derived from primary (first mates) and secondary sires (subsequent mates). They found no significant difference between the mating and fertilization success of sons resulting from the two classes of sires, indicating that any indirect benefit females may obtain by ‘trading up’ (i.e. mating again with more attractive males in order to obtain more attractive sons) is more than offset by the direct costs associated with interacting and mating with males (although see Head et al., 2005 for a counterexample in a cricket model system). Here we quantify the other major form of indirect benefits, ‘good genes’, in this model system.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Fly stocks

The base population of D. melanogaster used for all experiments described here was LHM. It is a large outbred population that has adapted to laboratory conditions for over 400 generations. Here, we provide only a brief description of LHM’s culturing protocol because it has been described in detail elsewhere (Rice et al., 2005, 2006). During each discrete generation (2 weeks), flies are reared in three consecutive sets of 56 vials. In the first set, called ‘juvenile competition’ vials, eggs that were laid the previous day are manually culled to 150–200 per vial for a total of about 10 000 eggs used to begin each generation. Progeny from these eggs remain in the juvenile competition vials for the larval, pupal and early adult stages of life. On day 12, when virtually all flies are sexually mature, flies are mixed among juvenile competition vials and 1792 individuals are randomly selected (under CO2 anaesthesia). These flies are transferred to a second set of ‘adult competition’ vials (at 16 pairs per vial), where they remain for 2 days. During this time, females compete for a limited supply of live yeast (10 mg), which strongly influences their lifetime fecundity (Stewart et al., 2005), and males compete to fertilize the females’ eggs. Most females have mated before they are transferred to the adult competition vials, and most mate again during this stage of their life history (Rice et al., 2005, 2006). After the 2-day adult competition stage, the flies are transferred to the third set of ‘oviposition’ vials, without live yeast, where they remain for 18 h before being discarded. Only eggs laid in the oviposition vials are used to found the next generation. Mortality during the juvenile competition phase of the life cycle averages 10–15%, whereas mortality during the adult competition and oviposition stages is negligible. The experiments described below are designed to match, as closely as possible, the standard culturing process of the LHM population. In addition to the LHM population, we also used a replica of this population, into which the recessive brown-eye colour marker (bw) has been recurrently backcrossed (Rice et al., 2006).

Experimental design

The logic of our experimental design was to first mate a female to a randomly selected sire (no mate choice) and then allow her to mate again (by selecting a mate from a pool of 16 potential sires), thereby providing an opportunity to ‘trade up’ and secure ‘good genes’ for her offspring – compared with what she received from the randomly assigned sire. To exclude any possibility of an ‘attractive sons’ indirect benefit, we compared only the fitness of daughters sired by the randomly assigned primary sire with the fitness of daughters sired by the secondary sire that was chosen in the context of sexual selection. Because previous work indicated that nearly all standing fitness variation in the LHM population is due to variation in adult fitness, rather than variation in egg-to-adult viability (Chippindale et al., 2001), we focused here exclusively on indirect benefits due to variation in the adult fitness of progeny.

Primary sires

In the first treatment, called primary sires, each of 16 brown-eyed virgin females (bwbw, 2–3 days old) was individually combined with a single red-eyed male (bw+bw+, 2–3 days old) in a 4.2-dram ‘first mating’ test tube containing 3 mL of standard culturing medium for 2 h (Fig. 1, top, first mating). Prior work had shown that this 2-h mating period is sufficient for nearly every female to mate (Stewart et al., 2005) and only mate once (Rice, 1996; Holland & Rice, 1999; Kuijper et al., 2006). After the first mating, the red-eyed male was discarded and the 16 brown-eyed females were pooled and transferred to a ‘second mating’ vial containing 16 brown-eyed (bwbw) males (all 2–3 days of age; Fig. 1, top, second mating; corresponds to the 2-day adult competition phase of the LHM life cycle). During the 2 days in the second mating vial, the brown-eyed females had an opportunity to ‘trade up’ by mating again with any one of the 16 males. Prior work showed that nearly all females remate once during this period (Stewart et al., 2005), but rarely remate more than once (Morrow et al., 2005). After 2 days in the second mating vial, the brown-eyed females were transferred to individual brood development vials (a 4.2-dram test-tube containing 3 mL of standard culture medium; corresponding to the 18-h oviposition phase of the LHM life cycle) where they laid eggs for 18 h, after which the females were discarded. Twelve days later, when the LHM population would normally be transferred to the adult competition phase of their lifecycle, the progeny were scored for eye colour (Fig. 1, brood development). Broods with both red-eyed (bw+bw) and brown-eyed (bwbw) progeny indicated that a female mated to both the primary (red-eyed) male as well as a secondary (brown-eyed) male. To equalize any maternal effects that might be produced in response to the number of times that a female mated, we assayed only red-eyed daughters from broods that contained both eye colour phenotypes. If there was a maternal effect on progeny due to the female receiving a double dose of seminal fluid proteins, using offspring from a single mating might affect the fitness of daughters from secondary sires and, hence, either mask or exaggerate potential indirect benefits of good genes. Red-eyed (bw+bw) daughters sired by the first, red-eyed male (bw+bw+) were collected and assayed for fitness as described below.

image

Figure 1.  Experimental protocol to obtain, and assay for fitness, daughters from (i) randomly assigned primary sires; and (ii) secondary sires that successfully competed in the context of sexual selection. The shaded boxes and bold text highlight the differences between the primary and secondary sire treatments. Details of the protocol are described in the text.

Download figure to PowerPoint

From the group of all red-eyed daughters produced by all doubly mated females from each second mating vial, two adult competition vials were constructed, each containing eight randomly selected red-eyed daughters, eight brown-eyed competitor females and 16 red-eyed males for 2 days (Fig. 1, adult competition). In those cases where there were fewer than 16 daughters, additional brown-eyed females were added so that the total number of females in each adult competition vial was maintained at 16. After the 2-day adult competition stage, the red-eyed daughters were placed separately into individual oviposition vials (as described above for their mothers) for 18 h (Fig. 1, oviposition), and total lifetime fecundity was measured as the number of eggs each female produced.

Secondary sires

The second treatment, secondary sires, was carried out identically to the primary sires treatment, except that the eye colour of the primary and secondary sires was reversed (Fig. 1, bottom, first and second matings), so that the assayed red-eyed daughters were sired by males that were successful in the context of sexual selection, rather than being randomly assigned to females. By comparing the lifetime reproduction of red-eyed females from the two treatments, we were able to test for any change in offspring quality associated with females choosing to mate a second time and possibly trading up by mating with a male of higher genetic quality than the first mate. Both treatments were simultaneously carried out in a single block of 20 replicates each.

Statistics

Because the distributions of female fecundity (fitness) deviated from a normal distribution, nonparametric bootstrapping (10 000 iterations) was used to carry out all statistical analyses. Bootstrapped 95% confidence intervals (95% CI) are reported to indicate the uncertainty in our estimates of effect sizes (Colegrave & Ruxton, 2003). Reported P-values represent the maximal value of α, such that a 100 × (1 − α)% CI supported the null hypothesis. Bootstrapping was carried out using PopTools (http://www.cse.csiro.au/poptools).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Daughters resulting from matings with primary sires had an average lifetime fecundity of 32.47 eggs per female (95% CI 30.01–34.99). Daughters resulting from matings with secondary sires had an average lifetime fecundity of 31.74 eggs per female (95% CI 29.62–33.88). The average difference (primary sire–secondary sire) between the two treatments was 0.73 eggs per female (95% CI −2.59 to 3.92). Even though the point estimate indicates a small (−2.25%) decline in fitness for offspring from secondary sires, the two treatments did not significantly differ from each other (= 0.66), indicating no substantial difference in the fitness of daughters from either class of sires. The upper bound of our bootstrap 95% confidence interval of the effect size of the good genes benefit obtained through remating indicates that this potential benefit does not exceed 6.1%.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Our experimental design compared the fitness of the daughters derived from randomly assigned primary sires with those produced from secondary sires that competed in the context of sexual selection. If sexual selection caused females to gain indirect benefits via the good genes mechanism, then daughters from secondary sires should have had higher lifetime fecundity than those from primary sires. This pattern was not manifest in our data, and the narrow 95% confidence interval for the effect size indicates that any potential good genes benefit that was undetected due to insufficient statistical power was no larger than about +6%. Because in this population the cost to females of interacting with males has been shown to be at least −19% of their lifetime fecundity (Rice et al., 2006), indirect benefits through good genes is insufficient to counterbalance such harm. When we combine the results presented here based on the ‘good genes’ mechanism of indirect benefits, with a previous study of the ‘attractive sons’ mechanism (Orteiza et al., 2005), the total benefits from both types of indirect benefits are far too small to compensate for male-induced harm to their mates (Fig. 2), which is in accordance with a previous study of the LHM population using experimental evolution (Stewart et al., 2005).

image

Figure 2.  Summary of minimal direct costs and maximal indirect benefits to females from interactions with males (expanded from Rice et al., 2006).

Download figure to PowerPoint

There are several factors that could influence the interpretation of our results. First, the trade-off in females between direct costs and indirect benefits is likely to depend on the environmental context in which male–female interactions occur. Our study followed the ‘laboratory island analysis’ approach (Rice et al., 2005, 2006) in which we studied a large outbred population that was adapted to competitive laboratory conditions for a protracted period of time (over 400 generations). Such a locally adapted population is considered to be of interest in-and-of-itself (like a natural island population), and its attributes are not presumed to be identical to those of the natural population from which it was founded. Substantial adaptation to the laboratory environment is expected for most quantitative traits (including net fitness; Frankham & Loebel, 1992), but oligogenic adaptations that depend on rare mutations may be absent. Although some naturally occurring male sexual signals and female preferences may not function in the laboratory environment, past studies have shown that traits strongly influencing sexual selection can rapidly evolve in an adaptive manner in both the laboratory (e.g. Endler, 1980) and the field (e.g. Reznick et al., 1990). Because of the potential for such rapid evolution, we consider our assay to be a meaningful – but not definitive – assay of the potential for indirect benefits to offset direct harm from males, albeit in a population that has had hundreds, rather than many thousands, of generations to become adapted to its present environment.

A second consideration when interpreting our results is that alternative experimental designs might have been more powerful in detecting indirect benefits from female choice. For example, a potentially more powerful assay would have provided females with more choices of males and/or less female–female competition for the limited supply of high-fitness males. However, our assay was designed to closely match the environmental conditions to which the females had adapted. So, although our experiments may have missed some potential indirect benefits that females could achieve in nature, they should nonetheless reflect the level of indirect benefits that females would actually achieve during their coevolution with males in the LHM environment.

A third consideration is that we did not measure the hatch rate of eggs laid by daughters of primary and secondary sires. If daughters from secondary sires had substantially higher fertility, this could provide an advantage to remating that we did not measure. However, the average egg hatch rate in the LHM population is extremely high (97%; Long & Pischedda, 2005); so, there is very little scope for an increase in this parameter to counterbalance the large cost of multiple mating (Linder & Rice, 2005; Kuijper et al., 2006) in this population.

A final consideration is that our study of indirect benefits was restricted to the measures of the fitness of daughters and did not include that of sons. We focused on the fitness of daughters to ensure that any indirect benefits that we observed were not due to sexual selection. If good genes benefits were larger in sons compared with that in daughters, then our measure of indirect benefits would be biased downward. A previous experiment utilizing a similar experimental design and the same LHM population (Orteiza et al., 2005), however, indicated that such a bias, if present, was small. This study found no significant indirect benefits via trading up (with a 95% upper bound effect size of ∼3%) when the fitness of sons was measured in the context of sexual selection. This 3% maximum effect size would include nearly all indirect benefits through both good genes and sexy sons size because: (i) in the LHM population, juvenile survival is high (85–90% egg-to-adult viability), with low heritable variation (Chippindale et al., 2001); and (ii) adult mortality is nearly zero (Rice et al., 2005). Nonetheless, our finding of no measurable indirect benefits for the LHM population differs from the positive findings of other studies (e.g. Head et al., 2005; Rundle et al., 2007). There is also recent evidence that a ‘cross-generational effect’ could compensate for the male-induced harm that females receive when they mate with more than one male (Priest et al., 2008). However, a multigeneration study on our base population has demonstrated that this process cannot fully compensate for the cost of mating in our population (Stewart et al., 2005).

Our study, like that of Holland (2002) and Brown et al. (2004) on the same LHM base population, provides little support for the idea that females benefit substantially from indirect benefits in the laboratory environment (although see Rundle et al., 2007; where a larger effect was found, but the direct costs of male–female interactions were unmeasured). But do these results tell us anything about the capacity of indirect benefits to counterbalance male harm in nature? Certainly nature is far more complex, owing to the presence of spatial and temporal heterogeneities, competitors, predators and so on. However, we think that the critical factors in testing for substantial levels of indirect benefits are: (i) standing genetic variation for fitness; and (ii) the ability of males to signal their genetic quality to females. Because of its continuous large size and outbreeding, the LHM population has retained substantial standing genetic variation for fitness, despite having adapted to the laboratory environment over the last 400+ generations (reviewed in Rice et al., 2005, 2006). Also the LHM population, during the adult competition phase of its life cycle, is reared at low densities (16 pairs per vial, with vials placed on their sides to provide more horizontal space for courtship), which permits the males’ elaborate courtship repertoire to be expressed. Nonetheless, we cannot rule out the possibility that indirect benefits may be as large as 6%; so, in situations where male-induced harm to females is much smaller than found in our laboratory model system, indirect benefits may be capable of counterbalancing male-induced harm to females. We also cannot rule out the possibility that the good genes benefit is more substantial in populations where mortality is higher than in our laboratory population. We can conclude, however, that indirect benefits do not come even close to compensating females for the direct harm from their mates in at least one model system: a population with high-standing genetic variation for fitness that has adapted to a relatively simple laboratory environment for hundreds of generations.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This work was supported by two grants from the National Sciences Foundation (DEB-0128780 and DEB-0111613). We thank Ted Morrow, Jodie Linder and Urban Friberg for their helpful comments on earlier versions of this manuscript and Giles Wemmbly-Hogg for support during data collection.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References