Theoretical work predicts that dispersal and self-fertilisation (selfing) should always be negatively correlated and the Good Coloniser Syndrome (GCS) of high dispersal and selfing should not occur when both traits are free to evolve. This contradicts positive relationships between selfing and dispersal in empirical data. Critically, previous work assumes density of adults is spatially and temporally homogeneous, so selfing results in homogeneity in propagule production and competition, which

more »

... iminates the benefit of dispersal for escaping from local resource competition. We investigate the joint evolution of dispersal and selfing in a demographically structured metapopulation model where local density varies due to stochastic extinction-recolonisation dynamics. Increasing local extinction rate reduces local density across the metapopulation, which favours high selfing to mitigate mate limitation, but increases heterogeneity in density, which favours high dispersal for escape from competition. Together, these effects produce a positive relationship between selfing and dispersal, and evolution of the GCS. Nevertheless, the relationship between selfing and dispersal is context-dependent, as varying dispersal cost yields a negative relationship. Our results imply that if spatiotemporal heterogeneity in environmental suitability increases towards the range edge, the GCS may evolve there, favouring further range expansion (Cf. Bakers Law).