Print PageLoss and fragmentation in native vegetation cover are thought to be one of the major drivers of widespread declines in pollination success. However, the specific mechanisms through which these landscape changes are driving declines remain unknown; little is known about the relative contribution of landscape composition versus configuration on disruption of pollination services. We tested the relative importance of landscape composition versus configuration on the reproductive success of Heliconia tortuosa, a hummingbird pollinated forest herb, near Las Cruces, Costa Rica. We used a stratified random sampling design to select sites across orthogonal gradients in patch size, amount of forest, and elevation (N=34 patches [166 plants]). We examined proportion of successful fruits out of the total flowers attempted and number of seeds produced out of the total possible (given the number of flowers). Using linear mixed models, with 'patch' as a random effect, we modeled the effects of patch size, proportion of forest within 1 km and the interaction between these variables. We controlled statistically for the influence of elevation.
Results/Conclusions
We found support for a positive influence of patch size on proportion of successful fruits. The top model for proportion of successful fruits was the additive relationship of patch size and distance to edge. Patch size had a cumulative importance weight of 0.6. After controlling for elevation, the top model for proportion of seeds also included patch size. Patch size had a cumulative importance weight of 0.61. Patch size was clearly more important than proportion of forest (2.03 times more plausible given the data and candidate models). It appears that as patch size increases plants are setting more seeds. We hypothesize that these differences in reproductive success are a result of differential rates of pollinator movements across a gradient in landscape configuration (i.e., patch size).
phone 202-833-8773