Gender specialization of flowers within inflorescences of hermaphroditic plants

Although hermaphroditic flowers have both male and female parts, each flower may not allocate the same amount of reproductive resources to male and female functions. I show that, in Aquilegia caerulea, early-opening flowers allocate more of their reproductive resources to female function relative to later-opening flowers. This pattern persists at both the pre- and post-pollination stages. I illustrate how the pattern of variation in sex allocation among flowers can be explained mostly by a decrease in the absolute amount of reproductive resources allocated to female function in later-opening flowers. For the male function, I present patterns of variation among flowers in the number of anthers, total pollen production, anther size and pollen grain size. For the female function, I examine patterns of variation among flowers in the number of carpels, the number of ovules per carpel, and in fruit and seed set. I experimentally demonstrate that neither variation in pollen abundance nor interfloral competition for fruit-developing resources can satisfactorily explain the pattern of variation in seed set among flowers in Aquilegia caerulea.; I develop a mathematical model to illustrate the factors that can influence patterns of variation in sex allocation among flowers within inflorescences of hermaphroditic plants. I show that, through their influence on the mating environment of pollen, dichogamy and pollinator directionality will select for variation in sex allocation among flowers. Dichogamy is the temporal separation of male and female functions within a flower. Pollinator directionality refers to the movement pattern of pollinators along inflorescences. The model predicts that, in sequential bloomers, protandry will select for greater female bias in early--relative to late--opening flowers. The opposite pattern is expected with protogyny. In simultaneous bloomers, dichogamy will select for variation in sex allocation among individuals in the population but not among flowers within inflorescences. If pollinators tend to move up inflorescences, top-position flowers will be more male-biased relative to bottom-position flowers. The reverse sex allocation pattern is expected when pollinators move down inflorescences.