The evolutionary dynamics of plant mating systems: how bias for studying 'interesting' plant reproductive systems could backfire.

Background and aims: An "abominable mystery": angiosperm sexual systems have been a source of both interest and frustration for the botanical community since Darwin. The evolutionary stability, overall frequency, and distribution of self-fertilization and mixed-mating systems have been explored in a variety of studies. However, there has been no recent study which directly addresses our knowledge of mating systems across families, the adequacy of existing data, or the potential for biases.

Scope: Here we present an updated dataset of mating systems across flowering plants covering 6,781 species and 212 families based on a synthesis of existing reviews and an original literature review using Web of Science. We assess the adequacy of this data by evaluating for bias indicating enrichment of certain families or sexual systems.

Key results: We find that the vast majority of our data on mating systems comes from a small number of disproportionally sampled families, and that families with significant proportions of dioecious or monoecious species are much more likely to be undersampled.

Conclusions: Our results show that the frequency of selfing in angiosperms is overestimated, possibly due to increased research interest in selfing and mixed-mating systems. This suggests that systematic study bias may mean we know less about this vital facet of plant life than we think.