The need for reproductive assurance during dispersal, along with the pressure of local mate competition, means that the importance of frequent or repeated colonization is implicit in the sexual-system evolution literature. However, to date there have been few empirical tests of the association between colonization and the sexual system in plants. Here we provide such a test by comparing occupancy and abundance of populations of the European plant Mercurialis annua across regions characterized by different sexual systems. Specifically, we predicted that monomorphic, hermaphroditic populations, which are thought to have evolved under selection for reproductive assurance during repeated bouts of colonization, would be smaller and their suitable habitat less frequently occupied than dimorphic populations, where males co-occur with either females or hermaphrodites. We show that both of these predictions are upheld. We evaluate our results against competing hypotheses for the occupancy-abundance relationship and conclude that they are most consistent with the metapopulation model for sexual-system variation in M. annua.