No abstract

Inbreeding avoidance and asymmetric competition over resources have both been identified as factors favouring the evolution of sex-biased dispersal. It has also been recognized that sex-specific costs of dispersal would select for sex-biased dispersal, but there is little quantitative information on this aspect. In this paper we explore (i) the quantitative relationship between cost-asymmetry and a bias in dispersal, (ii) the influence of demographic stochasticity on this effect, and (iii) how inbreeding and cost-asymmetry interact in their effect on sex-specific dispersal. We adjust an existing analytical model to account for sex-specific costs of dispersal. Based on numerical calculations we predict a severe bias in dispersal already for small differences in dispersal costs. We corroborate these predictions in individual-based simulations, but show that demographic stochasticity generally leads to more balanced dispersal. In combination with inbreeding, cost asymmetries will usually determine which of the two sexes becomes the more dispersive.

The optimal probability and distance of dispersal largely depend on the risk to end up in unsuitable habitat. This risk is highest close to the habitat's edge and consequently, optimal dispersal probability and distance should decline towards the habitat's border. This selection should lead to the emergence of spatial gradients in dispersal strategies. However, gene flow caused by dispersal itself is counteracting local adaptation. Using an individual based model we investigate the evolution of local adaptations of dispersal probability and distance within a single, circular, habitat patch. We compare evolved dispersal probabilities and distances for six different dispersal kernels (two negative exponential kernels, two skewed kernels, nearest neighbour dispersal and global dispersal) in patches of different size. For all kernels a positive correlation between patch size and dispersal probability emerges. However, a minimum patch size is necessary to allow for local adaptation of dispersal strategies within patches. Beyond this minimum patch area the difference in mean dispersal distance between center and edge increases linearly with patch radius, but the intensity of local adaptation depends on the dispersal kernel. Except for global and nearest neighbour dispersal, the evolved spatial pattern are qualitatively similar for both, mean dispersal probability and distance. We conclude, that inspite of the gene‐flow originating from dispersal local adaptation of dispersal strategies is possible if a habitat is of sufficient size. This presumably holds for any realistic type of dispersal kernel.

Inbreeding depression, asymmetries in costs or benefits of dispersal, and the mating system have been identified as potential factors underlying the evolution of sexbiased dispersal. We use individual-based simulations to explore how the mating system and demographic stochasticity influence the evolution of sex-specific dispersal in a metapopulation with females competing over breeding sites, and males over mating opportunities. Comparison of simulation results for random mating with those for a harem system (locally, a single male sires all offspring) reveal that even extreme variance in local male reproductive success (extreme male competition) does not induce male-biased dispersal. The latter evolves if between-patch variance in reproductive success is larger for males than females. This can emerge due to demographic stochasticity if habitat patches are small. More generally, members of a group of individuals experiencing higher spatio-temporal variance in fitness expectations may evolve to disperse with greater probability than others.

We analyze the simultaneous evolution of emigration and settlement decisions for actively dispersing species differing in their ability to assess population density. Using an individual-based model we simulate dispersal as a multi-step (patch to patch) movement in a world consisting of habitat patches surrounded by a hostile matrix. Each such step is associated with the same mortality risk. Our simulations show that individuals following an informed strategy, where emigration (and settlement) probability depends on local population density, evolve a lower (natal) emigration propensity but disperse over significantly larger distances - i.e. postpone settlement longer - than individuals performing density-independent emigration. This holds especially when variation in environmental conditions is spatially correlated. Both effects can be traced to the informed individuals' ability to better exploit existing heterogeneity in reproductive chances. Yet, already moderate distance-dependent dispersal costs prevent the evolution of multi-step (long-distance) dispersal, irrespective of the dispersal strategy.