* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Condition-dependent mutation rates and sexual selection
Group selection wikipedia , lookup
Human genetic variation wikipedia , lookup
Oncogenomics wikipedia , lookup
Genome (book) wikipedia , lookup
Genetic drift wikipedia , lookup
Biology and sexual orientation wikipedia , lookup
Sexual dimorphism wikipedia , lookup
Koinophilia wikipedia , lookup
Frameshift mutation wikipedia , lookup
Sexual selection wikipedia , lookup
Population genetics wikipedia , lookup
doi:10.1111/j.1420-9101.2008.01683.x Condition-dependent mutation rates and sexual selection S. COTTON Research Department of Genetics, Evolution & Environment, University College London, London, UK Keywords: Abstract condition; good genes; mutation rate; ornament; sexual selection. ‘Good genes’ models of sexual selection show that females can gain indirect benefits for their offspring if male ornaments are condition-dependent signals of genetic quality. Recurrent deleterious mutation is viewed as a major contributor to variance in genetic quality, and previous theoretical treatments of ‘good genes’ processes have assumed that the influx of new mutations is constant. I propose that this assumption is too simplistic, and that mutation rates vary in ways that are important for sexual selection. Recent data have shown that individuals in poor condition can have higher mutation rates, and I argue that if both male sexual ornaments and mutation rates are conditiondependent, then females can use male ornamentation to evaluate their mate’s mutation rate. As most mutations are deleterious, females benefit from choosing well-ornamented mates, as they are less likely to contribute germline-derived mutations to offspring. I discuss some of the evolutionary ramifications of condition-dependent mutation rates and sexual selection. Introduction Genetic models of sexual selection show that females can gain indirect benefits in terms of ‘good genes’ for their offspring if they show mating preference for ornaments that signal male genetic quality (Iwasa et al., 1991; Iwasa & Pomiankowski, 1994, 1999; Houle & Kondrashov, 2002). The more genetic variance that exists in male quality, the more females gain from being choosy, as mating with genetically high-quality males endows their offspring with better than average genotypes. A variety of mechanisms for generating genetic variation in male quality have been suggested, including the continual loss of genetic adaptation as a result of co-evolution with parasites (Hamilton & Zuk, 1982) and recurrent deleterious mutation (Houle, 1991; Burt, 1995). The latter has featured heavily in previous models. However, these theoretical treatments have assumed that the influx of new mutations is constant both across generations and between individuals within the same generation (Iwasa et al., 1991; Iwasa & Pomiankowski, 1994; Houle & Kondrashov, 2002). Here, I suggest that this assumption Correspondence: Samuel Cotton, Research Department of Genetics, Evolution & Environment, University College London, Wolfson House, 4 Stephenson Way, London NW1 2HE, UK. Tel.: +44 20 7679 5116; fax: +44 20 7679 5052; e-mail: [email protected] may be too simplistic and propose that mutation rates may vary in ways that are important for sexual selection. Far from being static and inflexible, mutation rates are highly variable and responsive to selection (Baer et al., 2007). Although mutations arise in all cell types, the most important mutational events for evolution are those that occur in the germline, as these are inherited by offspring. The majority of mutations are deleterious to fitness; so, some opposing force must maintain the mutation rate above zero. Selection can favour the spread of mutator alleles that elevate mutation rates, as a result of rare beneficial mutations that they create and hitch-hike with (e.g. Sniegowski et al., 1997; Taddei et al., 1997). However, this is likely to occur only in asexual populations where there is no recombination to unlink mutator alleles and beneficial mutations (Johnson, 1999). In sexual populations, recombination drastically reduces hitch-hiking of mutator alleles (Johnson, 1999) and the alternative, more pervasive, explanation for a nonzero mutation rate is that of a physiological constraint, as maintaining a low mutation rate is likely to be costly (Sniegowski et al., 2000; Agrawal & Wang, 2008). Evolved mutation rates are therefore an optimal balance between the costs of DNA fidelity and the predominantly deleterious consequences of mutation (Sniegowski et al., 2000). ª 2009 THE AUTHOR. J. EVOL. BIOL. 22 (2008) 899–906 JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 899 900 S. COTTON If the costs of mutational repair and DNA fidelity are evolutionarily important, then mutation rates should be sensitive to environmental and genetic stress, as the degree of stress will influence the relative costs of DNA replication fidelity and repair. In this respect, mutation rates are similar to other genomic traits that are condition-dependent, such as recombination (Parsons, 1988; Lucht et al., 2002). Transposition rates of selfish genetic elements, and hence mutational events that result from their incision, increase significantly in the presence of external stressors (McClintock, 1984; Grandbastien, 1998) and populations maintained under environmental stress evolve higher mutation rates in both prokaryotes (Bjedov et al., 2003) and eukaryotes (Goho & Bell, 2000). Goho & Bell (2000) estimated that mutation rates were 10- to 40-fold greater in Chlamydomonas populations maintained in mildly stressful environments compared with that in controls reared under benign conditions. Recently, Agrawal & Wang (2008) demonstrated that such patterns also exist at the level of the individual by showing for the first time that DNA repair is condition-dependent in a multicellular eukaryote. In Drosophila melanogaster, maternal DNA repair mechanisms can repair damaged DNA in sperm after fertilization. Females were mated to mutagenized males and the daughters were screened for recessive lethals on the paternally inherited X chromosome. Females maintained on a low-quality diet transmitted significantly more (!30%) sex-linked recessive lethals to their offspring than did females reared on a high-quality diet (Agrawal & Wang, 2008). Together, these results support the hypothesis that individuals in poor condition (as a result of detrimental internal or external factors) have higher mutation rates, through either increased incidences of DNA damage and ⁄ or reduced ability to repair such damage. If mutation rates scale negatively with individual quality, then they are expected to covary with other condition-dependent traits. There is a large body of evidence demonstrating that male sexual ornaments are expressed in a condition-dependent fashion, with the highest quality individuals in the best condition displaying the largest and most extravagant ornaments (Andersson, 1994; Johnstone, 1995; Cotton et al., 2004a). For example, in stalk-eyed flies the male ornament (elongated eye-stalks) becomes proportionately smaller as phenotypic condition is reduced by nutritional stress (David et al., 1998; Cotton et al., 2004b), dessication and heat shock (Bjorksten et al., 2001). Similarly, plumage ornaments are condition dependent in many avian species, becoming less gaudy or elaborate as the bearer’s phenotypic quality declines (e.g. Hill, 2000), and the frequency and complexity of many bird and orthopteran songs are reduced when individuals are stressed (e.g. Scheuber et al., 2003; Spencer et al., 2003). Sexual ornaments also reflect variation in genetic quality. For instance, male carotenoid colouration and sexual displays exhibit marked inbreeding depression in guppies (Sheridan & Pomiankowski, 1997; Van Oosterhout et al., 2003), consistent with being sensitive to deleterious mutation loads. Here, I propose, with the aid of a simple simulation, that, if both male sexual ornaments and mutation rates are condition-dependent, then the degree of ornament exaggeration will be revealing of the mutation rate of the bearer. Females may therefore be able to exploit male ornamentation to evaluate their mate’s germline mutation rate. As most mutations are deleterious, females will benefit indirectly from choosing wellornamented males as mates because they are less likely to contribute germline-derived mutations to offspring. So, in conjunction with additional genetic benefits of mate choice that result from pre-existing standing genetic variance in male quality (Iwasa et al., 1991; Iwasa & Pomiankowski, 1994, 1999; Houle & Kondrashov, 2002), condition-dependent mutation rates may elevate the overall level of heritable quality variation in males and lead to stronger selection on both male ornaments and female mate preferences. The model To demonstrate the potential for variation in phenotypic condition to create variation in the germline mutation rate, I use a simple simulation in which I follow Blumenstiel’s (2007) proposal that germline mutation rates are sensitive to investment into, and efficiency of, DNA replication and repair, such that ! "1=k E li ¼ ; ð1Þ Ii where li is the mean genome-wide deleterious mutation rate per haploid gamete in individual i. Under this model, the cost of maintaining l close to zero becomes prohibitively high and approaches infinity (Sniegowski et al., 2000). Ii reflects the investment by individual i into maintaining DNA fidelity. E and k are linear and exponential scaling parameters, respectively, that determine the cost of maintaining a particular mutation rate for individual i. As E (and ⁄ or k) increase(s) in magnitude, the greater is the cost of maintaining a given mutation rate. E (and ⁄ or k) can also be viewed as parameter(s) describing the efficiency of DNA replication fidelity and repair. The consequences of variation in Ii and E on li are shown in Fig. 1. If mutation rates are condition-dependent, then I is expected to increase in proportion with individual condition. Let us assume that Ii increases in an exponential fashion with condition, Ci, Ii ¼ ðImax % Imin ÞCia ; ð2Þ where Imax and Imin are the maximum and minimum levels of investment into mutation repair, respectively, and a is the scaling exponent. When a = 1, investment ª 2009 THE AUTHOR. J. EVOL. BIOL. 22 (2008) 899–906 JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY Mutation rates and sexual selection 0.5 0.1 0.05 sire’s genetic quality through its ability to provide a female with mutation-free gametes. If mutations are rare, distributed randomly across gametes and not subject to haploid selection, then the probability of syngamy and fertilization involving a mutation-free gamete from individual i (P[0]i) can be given by the first term in the Poisson series, 0.01 0.005 0.001 P½0(i ¼ Fig. 1 The consequences of variation in investment in DNA replication fidelity (Ii) across different values of E on the mean genomewide deleterious mutation rate per haploid gamete in individual i (li), as expected if li ¼ ðE=Ii Þ1=k ; k = 1.0. into DNA repair and replication fidelity increases linearly with condition, whereas when a > 1or a < 1, the rate of investment increases or decreases, respectively, with condition. To investigate the effect of condition-dependent mutation rates on any genetic benefits of mate choice, I use a series of simple simulations. Males were assigned a condition value (Ci) drawn at random from a normal distribution and standardized within the range 0.01 £ Ci £ 1. A male i also possesses a sexual ornament, whose size (Si) is determined by a condition-dependent exaggeration away from a natural selection optimum trait value (t), Si ¼ ðt þ t 0 Ci Þ þ e; ð3Þ where t ¢ is the degree of condition dependence and e is a normally distributed error with a mean of zero and a standard deviation of 1. Biologically, ! reflects sexual signalling inefficiency either through imperfect signalling of condition by ornaments or through perceptual errors made by females in their assessment of male ornaments. Females in the population exhibit an open-ended psychophysical mate preference function (Lande, 1981); the probability of a male i being chosen as a mate (Pmate,i) is contingent on its ornamental phenotype and the strength of female preference (y), eySi Pmate;i ¼ Pn yS i i¼1 e 901 ð4Þ where n is the number of males (here arbitrarily n = 100). Females mate at random when y = 0 and show stronger mating preferences for males with large ornaments when y increases above zero. I consider offspring fitness to be a function of the number of deleterious mutations, with mutation-free zygotes having the highest potential fitness. In order to keep the model as general as possible, I ignore the fitness effects of mutated gametes and instead define a potential 1 : eli ð5Þ The greatest genetic benefits will be obtained by mating with males bearing the highest values of P[0]i. This definition of P[0]i only represents a male’s true genetic quality if all new mutations are dominant lethals. Nonetheless, given that P[0]i scales negatively with overall gametic mutation load, then it is likely to be a useful proxy of a male’s gametic quality under more relaxed assumptions about mutation selection coefficients and dominance. Note that P[0]i is also a conservative index of male genetic quality, precisely because it does not account for the (magnitude of the) deleterious fitness consequences of gametes containing one or multiple mutations, which are increasingly likely in individuals with higher l. The probability (P[0]mating) of a female receiving a mutation-free sperm from a mating, given its preference function is, P½0(mating ¼ n X i¼1 ðPmate;i P½0(i Þ: ð6Þ Females with the highest P[0]mating values produce offspring with the fewest deleterious mutations. To explore whether sexual selection can favour males that produce the fewest mutations, I compared the populations with no sexual selection [random mating (y = 0) and no condition-dependent ornaments (t¢ = 0)], with those in which females preferred well-ornamented males (e.g. y = 0.5) and ornaments were revealing of male condition (e.g. t¢ = 5). To explore variation in the form of condition dependence of mutation rates I ran simulations under different scaling exponents that reflected an increasing, unchanging and decreasing rate of investment into mutational repair processes with increasing condition (a = 0.5, 1 and 2 respectively). Similarly, simulations were repeated over two values of E that reflected weakly and strongly declining mutation rates with increasing investment (E = 0.05 and 0.005 respectively). Each set of simulations was repeated 10 times and P[0]mating values from the different mating regimes were compared using Wilcoxon tests. Results If both male sexual ornaments and mutation rates are condition-dependent, then we observe that high-quality males with large ornaments have the lowest mutation rates and the greatest probability of delivering mutationfree gametes to females (Fig. 2a). Poor quality males have ª 2009 THE AUTHOR. J. EVOL. BIOL. 22 (2008) 899–906 JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 902 S. COTTON high mutation rates, small ornaments and produce fewer mutation-free sperm. By contrast, if male sexual traits are not condition-dependent (t¢ = 0), then ornament size does not reveal the underlying mutation rate and thus cannot be used to predict the probability of receiving mutation-free gametes (Fig. 2b). Relative to random-mating females, females that preferred well-ornamented males gained significant genetic benefits in terms of their offspring acquiring fewer mutations (higher P[0]mating values; Fig. 3). Although the benefits of sexual selection were small – but nonetheless significant – when investment into mutational repair increased with condition at a diminishing rate (e.g. when a = 0.5) and ⁄ or when mutation rates declined strongly with increasing investment (e.g. E = 0.005; Fig. 3b), they became quite large when investment into mutational repair increased with condition at an unchanging or increasing rate (e.g. when a ‡ 1; Fig. 3a). These results arose primarily because females showing preference avoid mating with the few very low quality males with very high mutation rates. Fig. 2 Sample simulations showing (a) condition-dependent ornaments (t¢ = 5, upper panel) can reveal condition-dependent mutation rates (li, central panel) and can be used to predict the probability of receiving a mutation-free gamete during a mating (P[0]i, lower panel). Removing the condition dependence of ornaments (b; t¢ = 0, upper panel) eliminates their utility as signals of mutation rate (central panel) and hence P[0]i (lower panel). Additional parameter values: E = 0.05, k = 1.0, Imax = 1.0, Imin = 0.001, a = 1.0, t = 4, n = 100. ª 2009 THE AUTHOR. J. EVOL. BIOL. 22 (2008) 899–906 JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY Mutation rates and sexual selection (a) (b) Fig. 3 The probability of a female receiving male gametes with no mutations (P[0]mating) under random mating (open bars; y = 0) and sexual selection via female mating preference for well-ornamented males (shaded bars; y = 0.5), when the decline in mutation rate with increasing investment was weak (a; E = 0.05) or strong (b; E = 0.005). Results are given for three values of a, the exponent relating investment to mutational repair with condition. Bars represent median values (±interquartile ranges) from 10 simulations. Asterisks denote the significance of Wilcoxon tests; ***P < 0.001, **P < 0.01, *P < 0.05. Additional parameter values: k = 1.0, Imax = 1.0, Imin = 0.001, t = 4, n = 100. Discussion Previous genetic models demonstrating that the handicap principle can work have assumed that the rate of influx of new mutations is constant (e.g. Iwasa et al., 1991; Iwasa & Pomiankowski, 1994; Houle & Kondrashov, 2002). Here I suggest, with reference to recent experimental evidence, that this assumption is too simplistic and propose an additional factor favouring sexual selection. Mutation rates are highly variable (Baer et al., 2007) and variation in individual condition has been shown to explain significant levels of intraspecific variation in mutation rates, with poor quality individuals having higher mutation rates than individuals in good phenotypic condition (Agrawal & Wang, 2008). Male sexual ornaments also covary with phenotypic condition in most systems studied (Andersson, 1994; Johnstone, 1995; Cotton et al., 2004a). The common reliance of 903 these two traits (mutation rates and ornamentation) on condition means that a male’s sexual trait size may be revealing of its propensity for generating new mutations. If mutations are predominantly deleterious and occur with increasing frequency in low condition individuals, then in addition to signalling any pre-existing genetic variance in quality, ornaments will also reflect the risk of endowing the offspring with recently acquired mutations. Given the relative infancy of investigations into condition-dependent mutation rates, it is currently unknown how widespread the phenomenon is, but the available evidence is from diverse taxa suggesting that it could be near ubiquitous (Goho & Bell, 2000; Agrawal & Wang, 2008). Likewise, it is not clear what proportion of intraspecific variance in mutation rates is explained by condition, although Agrawal & Wang (2008) report that a 30% increase in mutation in low condition individuals is accompanied by a 30% reduction in female fecundity, suggesting that the proportion may be quite high. It is also difficult to make specific predictions or conclusions, as most parameters in my simulations are unknown for real biological systems. However, a few general remarks can be made. The form of the relationships between: (i) investment into DNA repair and condition and (ii) investment and realized mutation rate, have large effects on the degree of potential benefits to be gained by females. The benefits are the greatest when the rate of investment into DNA repair increases with condition and when mutation rates decline more slowly with increasing investment. As with previous ‘good genes’ models (Iwasa et al., 1991; Iwasa & Pomiankowski, 1994), the reliability of signalling of such genetic benefits is the greatest when the degree of ornament condition dependence is high. I used a simplified measure of gamete quality, P[0]i, the probability of syngamy and fertilization involving a mutation-free gamete. However, if mutations are rare and occur at random in the genome, then the probability of receiving a heavily mutated gamete will be proportionately greater for poorly ornamented, low-quality mates with a higher mean genome-wide deleterious mutation rates. Even if loci act independently, offspring fitness (xm) is expected to decline rapidly with increasing numbers of mutations (m) because xm ¼ ð1 % sÞm (Agrawal, 2002). So P[0]i is also likely to be a highly conservative index of gamete quality. Interaction among alleles and loci will also have large influences on offspring fitness. For example, synergistic epistasis will lead to greater than additive reductions in offspring fitness if multiple mutations occur in the same gamete, as is more likely in individuals with higher average mutation rates. So, condition dependence of mutation rates, in conjunction with additive and nonadditive allelic and locus effects, may generate more genetic variance in quality than previously thought. Condition-dependent mutation may also contribute to solving the paradox of the lek (Borgia, 1979; Taylor & ª 2009 THE AUTHOR. J. EVOL. BIOL. 22 (2008) 899–906 JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 904 S. COTTON Williams, 1982). Theory predicts that directional selection will deplete genetic variation in fitness, as favoured alleles quickly spread to fixation (Maynard Smith, 1978; Charlesworth, 1987), and indirect selection on female preference as a result of ‘good genes’ is predicted to be rather low (!3.5%; Kirkpatrick & Barton, 1997). So why do females, which receive only genes during mating, continue to discriminate between males if there are so few genetic benefits of choice (Borgia, 1979; Taylor & Williams, 1982; Kirkpatrick & Ryan, 1991)? The dominant resolution to this paradox has been the hypothesis that, although the loss of adaptation per locus (for example, through recurrent deleterious mutation) is low, many genes underlie the condition so that the total amount of genetic variance introduced into the population across all loci is sufficient to drive ‘good genes’ sexual selection (Pomiankowski & Møller, 1995; Rowe & Houle, 1996). Under the assumptions that surround eqn (1), condition-dependent mutation results in poor quality individuals harbouring proportionately more mutations in their germline (Fig. 1). This (nonlinear) covariation in the influx of new mutations with condition elevates the overall level of genetic variance in offspring quality, and may therefore provide a stronger basis for ‘good genes’ sexual selection. Handicap theory requires that ornaments are costly, as this maintains signal honesty (Grafen, 1990; Iwasa et al., 1991; Iwasa & Pomiankowski, 1994, 1999). However, the cost of having a large ornament may impinge on viability so that well-ornamented individuals have lower phenotypic condition than those with small ornaments, despite having overall higher fitness (sensu Kokko, 2001). Under such circumstances well-ornamented males might be expected to have higher germline mutation rates than those with smaller ornaments, and thus females may weaken their preference for wellornamented males if they provide a higher complement of mutated gametes. Although some studies suggest that well-ornamented males may have lower phenotypic condition, the majority have shown that individuals in good phenotypic condition have the largest ornaments (reviewed in Andersson, 1994; Johnstone, 1995; Jennions et al., 2001; Cotton et al., 2004a) suggesting that the positive relationship assumed in this paper is common in nature. The signalling of condition-dependent mutation rates by male ornaments is a unique hypothesis because, whereas previous theory has suggested that ornaments reflect genotype quality of the bearer, the one presented in this paper suggests that ornaments may signal the genotypic quality of offspring. ‘Good genes’ models of sexual selection traditionally require that there is a genetic basis to condition-dependent ornament expression (Pomiankowski & Møller, 1995; Rowe & Houle, 1996; Cotton et al., 2004a; Hunt et al., 2004; Tomkins et al., 2004); the absence of genetic variance in the trait is usually taken to mean that ‘good genes’ effects are not important. However, if mutation rates are phenotypically condition-dependent (Agrawal & Wang, 2008), then females will still obtain genetic benefits for their offspring even if ornament expression does not have a genetic basis (i.e. when their expression is contingent only on phenotypic rather than genetic quality variation). So, ‘good genes’ sexual selection may be more cryptic than previously thought. Moreover, although conditiondependent mutations may influence ‘good genes’ to the extent that they are inherited by progeny, they may also lead to reduced sperm viability and ⁄ or fertilization ability leading to direct fitness consequences. So, a low sperm mutation load can also be classified as a direct benefit to the female as it improves the likelihood of successful fertilization, an important consideration if females are sperm limited (Sheldon, 1994; Arnqvist & Nilsson, 2000; Wedell et al., 2002). The traditional distinction between direct (material) and indirect (genetic) benefits of mate choice may therefore be rather more blurred than previously thought. Throughout this paper, I have concentrated on germline mutation rates, as these are the most important for evolution and ‘good genes’ sexual selection. However, condition-dependent mutations will also occur at a higher frequency in the soma of poor quality individuals. Although themselves being evolutionary ‘dead-ends’, mutations in the soma may still play an important role in sexual selection if they affect the fitness of their bearer, an extreme example being the heightened progression of cancers associated with elevated somatic mutation rates (Frank & Nowak, 2004). If somatic mutations reduce male health and vigour then such individuals will have lower condition and hence lower attractiveness, as a result of condition-dependent ornamentation. Moreover, an increased somatic mutation load may inhibit the performance of poor quality males still further, rendering them less able to provide females with resources, such as parental care, providing an additional advantage to females preferring wellornamented males. Condition- or fitness-dependent mutation rates have been suggested to increase the twofold cost of sex, with sexual populations having a greater number of mutations at equilibrium (Agrawal, 2002). In asexual populations, equilibrium mean fitness is contingent only on the mutation rate of the least mutationally loaded class, and as these individuals are of high fitness they also have low mutation rates. Hence, asexuals are expected to have higher equilibrium fitness than sexual populations (Agrawal, 2002). However, if conditiondependent male ornamentation, or male mating success, reflects male mutation rates, then sexual selection will result in preferred males transmitting the fewest mutations to their offspring, thereby lowering the mutational load. So, it seems plausible that any disadvantages to sex that arise from condition-dependent mutation rates may be overcome by sexual ª 2009 THE AUTHOR. J. EVOL. BIOL. 22 (2008) 899–906 JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY Mutation rates and sexual selection selection, in much the same way that Agrawal (2001) and Siller (2001) argued that the twofold cost of sex can be overcome by sexual selection purging existing mutations in the (male) population. Some similarities may be drawn between the current hypothesis and that of female preference based on agedependent male germline mutation (e.g. Radwan, 2003). Germline mutation rates are expected to increase with male age as a result of the higher number of sperm cell divisions (Ellegren, 2007); so, females should decrease the mutation load of their progeny by avoiding matings with older males. Current tests of this hypothesis are equivocal, suggesting that female choice for low (agedependent) germline mutation loads is weak (Radwan, 2003 and references therein). However, considerable variation in condition is likely to exist within and between male age classes and thus condition-dependent mutation variation may obscure any age-dependent mutation effects in natural populations. For example, older males are often in a better condition as a result of viability selection (Kokko, 1998) and negative correlations between age and condition will have contrasting consequences on mutation rates, rendering age or condition, in isolation, as poor indicators of germline mutation load. Studies of condition- and age-dependent mutation rates are in their infancy; so, future work should attempt to determine the relative importance of these two forces. Finally, condition-dependent DNA repair processes may also contribute towards the elevated male-biased mutation rates observed in sexually selected species (Bartosch-Härlid et al., 2003; Ellegren, 2007). Such biases are thought to arise because the greater number of cell divisions in the male germline increases the likelihood of DNA replication errors occurring during the production of sperm, relative to ova (Ellegren, 2007; Hedrick, 2007). Male-biased mutation rates tend to be higher in sexually selected species, as sperm competition increases the production of sperm, and therefore the propensity for mutation (Bartosch-Härlid et al., 2003; Blumenstiel, 2007; Ellegren, 2007). However, similar patterns may arise in species with sexual ornaments. Although an evolutionary relationship between ornaments and overall mutation rates has been proposed before (Petrie & Roberts, 2006; but see Cotton & Pomiankowski, 2007), I suggest that condition dependence of DNA repair will lead to an association between ornaments and the degree of male bias in mutation rate. Ornaments are costly (Kotiaho, 2001) and hence are likely to reduce condition in males relative to females, which lack such costly structures ⁄ displays. If this were so, then we would expect males, on average, to suffer higher mutation rates than females, leading to maledriven evolution. Relationships between ornaments and male-biased mutation rates have yet to be studied empirically; so, future work would profit from addressing this expectation. 905 Acknowledgments This work was supported by a NERC (UK) Fellowship. The author thanks K. Fowler and A. Pomiankowski for useful discussion and two anonymous reviewers for constructive comments on a previous version of the manuscript. References Agrawal, A.F. 2001. Sexual selection and the maintenance of sexual reproduction. Nature 411: 692–695. Agrawal, A.F. 2002. Genetic loads under fitness-dependent mutation rates. J. Evol. Biol. 15: 1004–1010. Agrawal, A.F. & Wang, A.D. 2008. Increased transmission of mutations by low-condition females: evidence for conditiondependent DNA repair. PLoS Biol. 6(2): e30. Andersson, M. 1994. Sexual Selection. Princeton University Press, Princeton, NJ. Arnqvist, G. & Nilsson, T. 2000. The evolution of polyandry: multiple mating and female fitness in insects. Anim. Behav. 60: 145–164. Baer, C.F., Miyamoto, M.M. & Denver, D.R. 2007. Mutation rate variation in multicellular eukaryotes: causes and consequences. Nat. Rev. Gen. 8: 619–631. Bartosch-Härlid, A., Berlin, S., Smith, N.G.C., Møller, A.P. & Ellegren, H. 2003. Life history and the male mutation bias. Evolution 57: 2398–2406. Bjedov, I., Tenaillon, O., Gérard, B., Souza, V., Denamur, E., Radman, M., Taddei, F. & Matic, I. 2003. Stress-induced mutagenesis in bacteria. Science 300: 1404–1409. Bjorksten, T.A., Pomiankowski, A. & Fowler, K. 2001. Temperature shock during development fails to increase the fluctuating asymmetry of a sexual trait in stalk-eyed flies. Proc. R. Soc. Lond. B 268: 1503–1510. Blumenstiel, J.P. 2007. Sperm competition can drive a malebiased mutation rate. J. Theor. Biol. 249: 624–632. Borgia, G. 1979. Sexual selection and the evolution of mating systems. In: Sexual Selection and Reproductive Competition in Insects (M. Blum & A. Blum, eds), pp. 19–80. Academic Press, New York. Burt, A. 1995. The evolution of fitness. Evolution 49: 1–8. Charlesworth, B. 1987. The heritability of fitness. In: Sexual Selection: Testing the Alternatives (J.W. Bradbury & M. Andersson, eds), pp. 21–40. John Wiley, Chichester. Cotton, S. & Pomiankowski, A. 2007. Sexually selected mutation rates. Heredity 98: 185–186. Cotton, S., Fowler, K. & Pomiankowski, A. 2004a. Do sexual ornaments demonstrate heightened condition-dependent expression as predicted by the handicap hypothesis? Proc. R. Soc. Lond. B 271: 771–783. Cotton, S., Fowler, K. & Pomiankowski, A. 2004b. Condition dependence of sexual ornament size and variation in the stalkeyed fly Cyrtodiopsis dalmanni (Diptera: Diopsidae). Evolution 58: 1038–1046. David, P., Hingle, A., Greig, D., Rutherford, A., Pomiankowski, A. & Fowler, K. 1998. Male sexual ornament size but not asymmetry reflects condition in stalk-eyed flies. Proc. R. Soc. Lond. B 265: 2211–2216. Ellegren, H. 2007. Characteristics, causes and evolutionary consequences of male-biased mutation. Proc. R. Soc. Lond. B 274: 1–10. ª 2009 THE AUTHOR. J. EVOL. BIOL. 22 (2008) 899–906 JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 906 S. COTTON Frank, S.A. & Nowak, M.A. 2004. Problems of somatic mutation and cancer. BioEssays 26: 291–299. Goho, S. & Bell, G. 2000. Mild environmental stress elicits mutations affecting fitness in Chlamydomonas. Proc. R. Soc. Lond. B 267: 123–129. Grafen, A. 1990. Sexual selection unhandicapped by the Fisher process. J. Theor. Biol. 144: 473–516. Grandbastien, M.A. 1998. Activation of plant retrotransposons under stress conditions. Trends Plant Sci. 3: 181–187. Hamilton, W.D. & Zuk, M. 1982. Heritable true fitness and bright birds: a role for parasites? Science 218: 384–387. Hedrick, P.W. 2007. Sex: differences in mutation, recombination, selection, gene flow, and tic drift. Evolution 61: 2750– 2771. Hill, G.E. 2000. Energetic constraints on expression of carotenoid-based plumage coloration. J. Avian Biol. 31: 559–566. Houle, D. 1991. Genetic covariance of fitness correlates: what genetic correlations are made of and why it matters. Evolution 45: 630–648. Houle, D. & Kondrashov, A.S. 2002. Coevolution of costly mate choice and condition-dependent display of good genes. Proc. R. Soc. Lond. B 269: 97–104. Hunt, J., Bussière, L.F., Jennions, M.D. & Brooks, R. 2004. What is genetic quality? Trends Ecol. Evol. 19: 329–333. Iwasa, Y. & Pomiankowski, A. 1994. The evolution of mate preferences for multiple handicaps. Evolution 48: 853–867. Iwasa, Y. & Pomiankowski, A. 1999. Good parent and good genes models of handicap evolution. J. Theor. Biol. 200: 97– 109. Iwasa, Y., Pomiankowski, A. & Nee, S. 1991. The evolution of costly mate preferences. II. The ‘handicap’ principle. Evolution 45: 1431–1442. Jennions, M.D., Møller, A.P. & Petrie, M. 2001. Sexually selected traits and adult survival: a meta-analysis. Q. Rev. Biol. 76: 3–36. Johnson, T. 1999. Beneficial mutations, hitchhiking and the evolution of mutation rates in sexual populations. Genetics 151: 1621–1631. Johnstone, R.A. 1995. Sexual selection, honest advertisement and the handicap principle: reviewing the evidence. Biol. Rev. 70: 1–65. Kirkpatrick, M. & Barton, N.H. 1997. The strength of indirect selection on female mating preferences. Proc. Natl Acad. Sci. USA 94: 1282–1286. Kirkpatrick, M. & Ryan, M.J. 1991. The evolution of mating preferences and the paradox of the lek. Nature 350: 33–38. Kokko, H. 1998. Good genes, old age and life-history trade-offs. Evol. Ecol. 12: 739–750. Kokko, H. 2001. Fisherian and ‘good genes’ benefits of mate choice: how (not) to distinguish between them. Ecol. Lett. 4: 322–326. Kotiaho, J.S. 2001. Costs of sexual traits: a mismatch between theoretical considerations and empirical evidence. Biol. Rev. 76: 365–376. Lande, R. 1981. Models of speciation by sexual selection on polygenic characters. Proc. Natl Acad. Sci. USA 78: 3721–3725. Lucht, J.M., Mauch-Mani, B., Steiner, H.Y., Metraux, J.P., Ryals, J. & Hohn, B. 2002. Pathogen stress increases somatic recombination frequency in Arabidopsis. Nat. Genet. 30: 311– 314. Maynard Smith, J. 1978. The Evolution of Sex. Cambridge University Press, Cambridge, UK. McClintock, B. 1984. The significance of responses of the genome to challenge. Science 226: 792–801. Parsons, P.A. 1988. Evolutionary rates: effects of stress upon recombination. Biol. J. Linn. Soc. 35: 49–68. Petrie, M. & Roberts, G. 2006. Sexual selection and the evolution of evolvability. Heredity 98: 198–205. Pomiankowski, A. & Møller, A.P. 1995. A resolution of the Lek paradox. Proc. R. Soc. Lond. B 260: 21–29. Radwan, J. 2003. Male age, germline mutations and the benefits of polyandry. Ecol. Lett. 6: 581–586. Rowe, L. & Houle, D. 1996. The Lek paradox and the capture of genetic variance by condition dependent traits. Proc. R. Soc. Lond. B 263: 1415–1421. Scheuber, H., Jacot, A. & Brinkhof, M.W.G. 2003. Conditiondependence of a multicomponent sexual signal in the field cricket Gryllus campestris. Anim. Behav. 65: 721–727. Sheldon, B.C. 1994. Male phenotype, fertility, and the pursuit of extra-pair copulations by female birds. Proc. R. Soc. Lond. B 257: 25–30. Sheridan, L. & Pomiankowski, A. 1997. Fluctuating asymmetry, spot asymmetry and inbreeding depression in the sexual coloration of male guppy fish. Heredity 79: 515–523. Siller, S. 2001. Sexual selection and the maintenance of sex. Nature 411: 689–692. Sniegowski, P.D., Gerrish, P.J. & Lenski, R.E. 1997. Evolution of high mutation rates in experimental populations of E. coli. Nature 387: 703–705. Sniegowski, P.D., Gerrish, P.J., Johnson, T. & Shaver, A. 2000. The evolution of mutation rates: separating causes from consequences. BioEssays 22: 1057–1066. Spencer, K.A., Buchanan, K.L., Goldsmith, A.R. & Catchpole, C.K. 2003. Song as an honest signal of developmental stress in the zebra finch (Taeniopygia guttata). Horm. Behav. 44: 132– 139. Taddei, F., Radman, M., Maynard-Smith, J., Toupance, B., Gouyon, P.H. & Godelle, B. 1997. Role of mutator alleles in adaptive evolution. Nature 387: 700–702. Taylor, P.D. & Williams, G.C. 1982. The lek paradox is not resolved. Theor. Popul. Biol. 22: 392–409. Tomkins, J.L., Radwan, J., Kotiaho, J.S. & Tregenza, T. 2004. Genic capture and resolving the lek paradox. Trends Ecol. Evol. 19: 323–328. Van Oosterhout, C., Trigg, R.E., Carvalho, G.R., Magurran, A.E. & Hauser, L. 2003. Inbreeding depression and genetic load of sexually selected traits: how the guppy lost its spots. J. Evol. Biol. 16: 273–281. Wedell, N., Gage, M.J.G. & Parker, G.A. 2002. Sperm competition, male prudence and sperm-limited females. Trends Ecol. Evol. 7: 313–320. Received 24 September 2008; revised 18 November 2008; accepted 25 November 2008 ª 2009 THE AUTHOR. J. EVOL. BIOL. 22 (2008) 899–906 JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY