In many species, females mate with multiple partners, meaning that sexual selection on male traits operates across a spectrum that encompasses the competition for mates (that is, before mating) and fertilizations (after mating). Despite being inextricably linked, pre- and postcopulatory sexual selection are typically studied independently, and we know almost nothing about how sexual selection operates across this divide. Here we bridge this knowledge gap using the livebearing fish Poecilia reticulata. We show that both selective episodes, as well as their covariance, explain a significant component of variance in male reproductive fitness. Moreover, linear and nonlinear selection simultaneously act on pre- and postcopulatory traits, and interact to generate multiple phenotypes with similar fitness.

No abstract

No abstract

In polyandrous species, a male's reproductive success depends on his fertilization capability and traits enhancing competitive fertilization success will be under strong, directional selection. This leads to the prediction that these traits should show stronger condition dependence and larger genetic variance than other traits subject to weaker or stabilizing selection. While empirical evidence of condition dependence in postcopulatory traits is increasing, the comparison between sexually selected and ‘control’ traits is often based on untested assumption concerning the different strength of selection acting on these traits. Furthermore, information on selection in the past is essential, as both condition dependence and genetic variance of a trait are likely to be influenced by the pattern of selection acting historically on it. Using the guppy (Poecilia reticulata), a livebearing fish with high levels of multiple paternity, we performed three independent experiments on three ejaculate quality traits, sperm number, velocity, and size, which have been previously shown to be subject to strong, intermediate, and weak directional postcopulatory selection, respectively. First, we conducted an inbreeding experiment to determine the pattern of selection in the past. Second, we used a diet restriction experiment to estimate their level of condition dependence. Third, we used a half-sib/full-sib mating design to estimate the coefficients of additive genetic variance (CVA) underlying these traits. Additionally, using a simulated predator evasion test, we showed that both inbreeding and diet restriction significantly reduced condition. According to predictions, sperm number showed higher inbreeding depression, stronger condition dependence, and larger CVA than sperm velocity and sperm size. The lack of significant genetic correlation between sperm number and velocity suggests that the former may respond to selection independently one from other ejaculate quality traits. Finally, the association between sperm number and condition suggests that this trait may mediate the genetic benefits of polyandry which have been shown in this species.