Population genetic studies of the evolution of breeding systems in flowering plants are reviewed. The selective advantage of a gene's increasing the selfing rate is stressed. In the evolution of outbreeding mechanisms, some strong disadvantage to selfing must therefore be acting; it is suggested that this disadvantage is inbreeding depression. Populations with no absolute barrier to selfing, and with intermediate levels of self-fertilization, appear to be the most likely starting state for the evolution of outbreeding mechanisms. There is some evidence for inbreeding depression in such populations. The evolution of distyly and dioecy are considered in some detail. An explanation for the existence of supergenes controlling these systems is proposed. The breakdown of distyly and tristyly are also considered. The evolution of recombination rates in selfing and outcrossing species is examined briefly.