FLORAL DISPLAY, POLLINATOR DISCRIMINATION, AND FEMALE REPRODUCTIVE SUCCESS IN 2 MONOECIOUS BEGONIA SPECIES

In many plant species with unisexual flowers, the female flowers lack pollinator rewards and are pollinated by visitors who mistake the fema le flowers for rewarding male flowers. Here we present the results of observations and experiments designed to investigate pollinator discri mination and its consequences for female reproductive success in two m onoecious Begonia species pollinated by deceit. Male and female flower s of B. urophylla are similar in shape, but female flowers are smaller than male flowers, while in B. tonduzii, male and female flowers have strikingly different shapes, but their petaloid tepals are similar in total area. Flower visitors displayed marked discrimination against t he rewardless female flowers of both species, Seven bee species were o bserved on B. urophylla, and for all species taken together the relati ve visitation rate to male flowers was 8.6 times that to female flower s. The bee Trigona fulviventris accounted for 95% of all visits to B. tonduzii and visited male flowers 15.4 times as often as female flower s. The flower visitors of B, urophylla varied significantly in their d egree of discrimination against female flowers, with visitation rates to male flowers 3.7-21.8 times those to female flowers. The two most a bundant flower visitors displayed the greatest discrimination against female flowers and, although they comprised nearly 54% of all visits, were responsible for only 15% of all visits to female flowers. Thus, t he relative abundance of different flower visitors on B. urophylla is a poor predictor of pollinator importance. In both species the inflore scence sex ratio was male-biased early in the season and female-biased later. Hand-pollination experiments revealed that fruit set in B. ton duzii was pollinator limited both during peak and late flowering, with supplemental pollination having its greatest effect late in the seaso n when male flowers were rare. In contrast, fruit set was only weakly affected by supplemental hand-pollination in B. urophylla. The signifi cant pollinator limitation observed in B. tonduzii and the lack of it in B, urophylla are consistent with our observation that the visitatio n rate to female flowers was much higher in B. urophylla than in B. to nduzii, which was due in part to the lower level of pollinator discrim ination against female flowers of B. urophylla. Our finding that femal e reproductive success is limited by pollinator visitation in B. tondu zii indicates an opportunity for the evolution of increased resemblanc e between male and female flowers in this species.