The interaction between sex-specific selection and local adaptation in species without separate sexes

Abstract

Local adaptation in hermaphrodite species can be based on a variety of fitness components, including survival, as well as both female and male sex-functions within individuals. When selection via female and male fitness components varies spatially (e.g. due to environmental heterogeneity), local adaptation will depend, in part, on variation in selection through each fitness component, and the extent to which genetic trade-offs between sex-functions maintain genetic variation necessary for adaptation. Local adaptation will also depend on the hermaphrodite mating system because self-fertilization alters several key factors influencing selection and the maintenance of genetic variance underlying trade-offs between the sex-functions (sexually antagonistic polymorphism). As a first step to guide intuition regarding sex-specific adaptation in hermaphrodites, we develop a simple theoretical model incorporating the essential features of hermaphrodite mating and adaptation in a spatially heterogeneous environment, and explore the interaction between sex-specific selection, self-fertilization and local adaptation. Our results suggest that opportunities for sex-specific local adaptation in hermaphrodites depend strongly on the extent of self-fertilization and inbreeding depression. Using our model as a conceptual framework, we provide a broad overview of the literature on sex-specific selection and local adaptation in hermaphroditic plants and animals, emphasizing promising future directions in light of our theoretical predictions.This article is part of the theme issue 'Linking local adaptation with the evolution of sex differences'.

Keywords: hermaphrodites; intra-locus sexual conflict; local adaptation; mixed mating systems; sexually antagonistic selection; variable selection.

Figure 1.

Spatially heterogeneous selection expands the total parameter space where sexually antagonistic polymorphism is maintained for predominantly outcrossing populations (C < ½) under both additive (a,b) and partially recessive (h_f = h_m = ¼; e,f) sexually antagonistic fitness effects, but has the reverse effect on predominantly selfing populations unless selection is very strong (C ≥ ½; c,d for additive effects; g,h, for dominance reversal). Analytical results (solid lines) are based on numerical evaluation of at the boundary equilibria (, ; see electronic supplementary material, equations A4.1a,b) for 10⁵ simulated pairs of selection coefficients (one value for s_f, one for s_m, for each patch) drawn independently from uniform distributions with minimum = 0, and maximum defined by max(s_f, s_m). Hence, max(s_f, s_m) defines the size of the square portion of s_f × s_m parameter space being sampled uniformly and ranges from weak selection only (max(s_f, s_m) = 0.025) to all of plausible selection parameter space (max(s_f, s_m) = 1). Plots show the proportion of randomly drawn pairs of selection coefficients for which both alleles can invade when rare as a function of max(s) for a single population with constant selection (black line), and populations with 2–5 patches in which selection may differ (greyscale lines) with no inbreeding depression (δ = 0). Results for deterministic simulations of the exact genotypic recursions based on 10⁴ simulated pairs of selection coefficients for each patch are plotted over the lines (greyscale points) and show the proportion of parameter sets for which simulations converged on a polymorphic equilibrium. Tight correspondence between the simulation and analytic results indicates that the model predictions are robust to the assumption of weak selection. Effects of inbreeding depression (δ > 0) are presented in electronic supplementary material, appendix A, figure A2.