A reexamination of the pollen-donation hypothesis in an experimental population of Asclepias exaltata

The evolution of large floral displays in hermaphroditic flowering plants has been attributed to natural selection acting to enhance male, rather than female, reproductive success. Proponents of the "pollen-donation hypothesis" have assumed that maternal resources, rather than levels of effective pollination, limit fruit set. We investigated the pollen-donation hypothesis in an experimental population of poke milkweed, Asclepias exaltata, where effective pollination did not limit fruit set. Specifically, we examined the effects of flower number per plant, and flower number per umbel on male reproductive success (number of fruits sired) and female reproductive success (number of fruits matured). In 1990, a paternity analysis was performed on fruits collected from 53 plants whose inflorescences were not manipulated. Flower number per plant was significantly correlated with male success, but not with plant gender. Flower number per plant was also significantly correlated with female success, but umbel number and stem number per plant together explained more than half (58%) the variation in female success. The percentage of fruit set was not significantly correlated with flower number per plant. Plants with large floral displays did not disproportionately increase in male reproductive success, relative to female success, as predicted by the pollen-donation hypothesis. In 1991, the effect of flower number per umbel on male and female reproductive success was investigated. Flower number per umbel was manipulated on four umbels per plant by removing flowers to leave 6, 12, or 18 flowers in each umbel. Plants with the largest umbels effectively pollinated twice as many flowers on other plants, but produced only 1.35 times as many fruits as plants with 6 and 12 flowers per umbel. Relative maleness of plants with large umbels was nearly twice that of small and medium umbels. Although these observations are consistent with the pollen-donation hypothesis at the level of umbels, they are problematic, because much of the variation in flower number perumbel exists within, rather than among, plants in natural populations. Thus, plants consist of both reproductively male (large) and female (small) inflorescences, which act to increase total reproductive success. It is therefore inappropriate to explain the evolution of large floral displays in milkweeds solely in terms of potential male reproductive success.