Seed Set and Wasp Predation in Dioecious Ficus variegata from an Australian Wet Tropical Forest

A RECENT REVIEW (Bronstein 1992) of the interaction between figs (Ficus species, Moraceae) and their pollinators (Agaonidae: Hymenoptera) evaluated some potential conflicts of interest hat might constrain the evolution of the mutualism. One apparent conflict ensues when female wasps pollinate flowers while depositing their eggs in a fraction of fig ovaries. Some developing seeds are totally consumed by wasp larvae, and consequently, wasps are both pollinators and seed predators. Variable style lengths of fig flowers were thought to regulate the donation of ovaries to wasps in exchange for seeds in monoecious figs, since the ovaries of short-styled flowers are considered accessible to wasp ovipositors while the ovaries of long-styled flowers are generally not (Janzen 1979a). In monoecious pecies, however, most styles are of intermediate l ngth, and there is insufficient evidence to conclude that the fate of fig ovaries is determined solely by style and ovipositor lengths (Compton & Nefdt 1990, Bronstein 1992). Heterostyly in gynodioecious figs generates a different conflict of interest han for the monoecious figs and their pollinators. In gynodioecious fig populations, female trees produce figs containing only longstyled, seed-bearing flowers and hermaphroditic trees produce figs containing short-styled, wasp-bearing female flowers and pollen-producing male flowers (Verkerke 1989). Hermaphroditic figs are functionally male, producing few or no seeds due to intense predation by wasp larvae that become the pollen vectors as adults. Hundreds of Ficus species with gynodioecious morphology are thus functionally dioecious (Berg 1989). The evolutionary conflict between gynodioecious figs and agaonid wasps arises from the separation of seed and wasp production between two sexes of trees. Wasps leave no offspring in female figs, but have access to every ovary in male figs. Thus, female trees are lethal to their pollinators, whereas the pollinators of male trees are reproductively successful (Valdeyron & Lloyd 1979). Kjellberg et al. (1987) supposed that intense selection against the pollinators of female figs could favor wasps capable of distinguishing the male figs from the females, but asynchronous flowering phenology between trees of different sexes could eliminate the opportunity for wasps to discriminate against females. Grafen & Godfray (1991) argued alternatively that male and female figs would be under strong selection to mimic each other to an extent that prevents wasps from discriminating between sexes. Hypotheses that attempt to explain the evolutionary stability of the gynodioecious fig/pollinator interaction (Kjellberg et al. 1987, Grafen & Godfray 1991) cannot be fully evaluated until more empirical data are available, especially for gynodioecious pecies in tropical habitats. An initial step is to compare the reproductive output of a fig species and its pollinator, as did Janzen (1979b) when he asked what proportion of monoecious fig ovaries are forfeited to wasps in exchange for seeds. Corlett et al. (1990) compared seed set and wasp production between monoecious and dioecious figs in Singapore. Our study compares the reproductive output of female and male figs in Ficus variegata Blume in Queensland, Australia. We measured levels of seed set and wasp predation in female and male trees in a population of F. variegata at Cape Tribulation, Australia (16?05'S, 145?27'E). F. variegata is gynodioecious and is pollinated by Ceratosolen appendiculatus Mayr (Condit 1969). Our study area is situated in a coastal lowland basin which receives over 4000 mm of rainfall annually. A humid and seasonally wet climate supports luxuriant vegetation dassified as complex mesophyll vine forest (Tracey 1982). F. variegata is very common along watercourses in the basin, especially where natural regeneration has occurred following dearing of the vegetation for cattle pasture. We examined the fruit of 11 male and 9 female trees from a population of more than 200 marked individuals along the banks of Mason's Creek. Figs were sampled during January and February of 1993, coincident with the annual peak period of