Five Evolutionary Pathways to Subdioecy

It appears that subdioecy in the seed plants has evolved repeatedly, by at least five evolutionary pathways. These are: (1) from hermaphroditism via gynodioecy; (2) from monoecy; (3) from distyly; (4) from hermaphroditism via andromonoecy and androdioecy; and (5) directly from hermaphroditism. Pathway 5 has been little studied. This paper summarizes the evidence for the five pathways, and gives a genetic model which represents both hermaphrodite or monoecious populations, and their direct differentiation toward dioecy. It is emphasized that there is a morphological, physiological, and evolutionary continuum between hermaphroditism and dioecy, so that some populations cannot easily be placed into discrete classes. For example, populations of Vaccinium angustifoblium may have small proportions of males or females, together with many hermaphrodites, which may vary in fertility. The fertility genes are apparently not organized into an XX/XY chromosome system. In the model recessive gene a reduces pollen but increases ovule fertility, whereas gene B acts in the reverse fashion, and is dominant or partially dominant for ovule and dominant for pollen fertility. Dominant B (with respect to ovule production) does not allow overdominance, but partially dominant B may result in overdominance in polymorphic equilibrium populations. Numerical results sometimes gave equilibria with all genotypes present for free recombination, but with only genotypes present for free recombination, but with only genotypes AB/AB or YY, AB/ab or XY, and ab/ab or XX present for complete linkage. The YY and XY types differentiate toward the male, and XX toward the female condition. Such results occur both in the presence and absence of overdominance, so that selection for heterozygotes, although it apparently aids differentiation towards subdioecy, is not essential. Such differentiation may be aided also when the above model is modified by incorporating outbreeding advantage, such that all offspring of crosses have greater viability than offspring of selfings. The above model assumes equal outcrossing rates for the ovules of all genotypes, but pollen outcrossing rates differ among genotypes and are frequency dependent. A new outcrossing rate is defined as the number of offspring derived from crossing through both ovules and pollen, as a proportion of all offspring. Types with more pollen but fewer ovules may have higher outcrossing rates than the reverse type for this definition of outcrossing. The selective forces apparently responsible for the evolution of dioecy are reviewed, and several selection models, involving fertility variation, resource allocation, sexual selection, and overdominance, are shown to be equivalent. It is held that outbreeding advantage is sometimes but not always an important selective force.