Sexual specialization and inbreeding avoidance in the evolution of dioecy

Dioecy has evolved independently, many times, among unrelated taxa. It also appears to have evolved along two contrasting pathways: (1) from hermaphroditism via monoecy to dioecy and (2) from hermaphroditism via gynodioecy to dioecy. Most dioecious plants have close cosexual relatives with some means of promoting outcrossing (e.g., herkogamy, dichogamy, self-incompatibility, or monoecy). To the extent that these devices prevent inbreeding, the evolution of dioecy in these species cannot logically be attributed to selection for outcrossing. In these cases, the evolution of dioecy is, we believe, due to selection for sexual specialization. However, in other species, that lack outbreeding close relatives, dioecy may have evolved from gynodioecy (males and hermaphrodites) as an outbreeding device. Subsequent disruptive selection and selection for sexual specialization may have also shaped the evolution of dioecy from gynodioecy in these species, resulting in two genetically determined, constant sex morphs. Both pathways for the evolution of dioecy require the operation of disruptive selection, though the gynodioecy route involves more restrictive disruptive selection and a genetic designation of gender. In contrast, the monoecy route is not dependent on the genetic designation of two sex morphs, but, rather, allows the possibility of sexual intermediates and sexual lability. Both pathways produce one morph in which maleness is suppressed and another in which the female function is negligible or nonexistent-the reproductive mode recognized as dioecy. Evidence is presented here to support the thesis that instances of sexual lability, the presence of an array of sexual intermediates, sex-switching, and sexual niche segregation can be explained in terms of the pathway that was taken in the evolution of a particular dioecious species. In addition, the degree of sexual dimorphism seen in dioecious species is correlated with mode of pollination (insect- or wind-pollinated) and other ecologicaldecades of debate concerning causes for the evolution of dioecy may simply reflect the fact that different research groups have studied different types of plants. We believe that our thesis is in accord with the finding that most dioecious species arose via monoecy and that sex-changing and sexual lability occur only among the species that arose by that pathway. Whereas in the dioecious species that arose via gynodioecy, we have heard of no reports of sexual lability or sex-changing, or of sexual niche partitioning, whether that is based on environmental sex determination or patchy selection that favors males and females (or differential flowering of males and females) in contrasting microsites. However, we would caution at this point that we have emphasized the role of pollination biology in sexual selection and may not have paid sufficient attention to other ecological factors such as differential predation and competitive abilities, which may also be agents of sexual selection that favor sexual specialization (Cox, 1982). Thus both IA and sexual selection have been important in the evolution of plant reproductive systems, but the degree to which each has participated depends critically on functional aspects of the species' reproductive biology and their interplay with the environment.