Many plant invaders rely on seeds for dispersal.  Therefore, factors controlling seed production, such as pollinator services, could influence the invasion process.  Pollinator interactions are sensitive to extrinsic environmental factors such as light, which can vary over small spatial scales.  We explored the variability of pollinator services to the partially self-compatible invasive shrub Lonicera maackii among two very different, but adjacent, light environments: forest edge and forest interior.  Pollinators that rely on visual cues or favorable temperatures may avoid low light and temperature environments in the forest interior.  Furthermore, the composition of visitors in the two habitats may differ in pollination effectiveness.  We tested the hypothesis that edge plants receive more visits per flower and higher stigmatic pollen loads than forest interior plants.  We also tested whether pollen limitation of fruit and seed production was greater in forest interior habitats than to forest edges.  In two urban forests in Columbus, OH USA, fruit and seed production were compared between flowers pollinated naturally by ambient pollinators and flowers hand pollinated with outcross pollen.  We recorded pollinator composition, visitation rates to flowers, and the number of pollen grains deposited on stigmas to elucidate the mechanism of reproductive differences.

Results/Conclusions

Pollinators of Lonicera maackii represented bees of eight genera.  Honey bees were most frequent, followed by small generalist native bees.   Visitation rates per flower at one site were higher on the edge as predicted but this pattern was not consistent at the other site.  High densities of honey bees on the edge explained most of this variation.  Edge flowers consistently had 2.36-fold higher stigmatic pollen loads than interior flowers.  Lonicera maackii exhibited significant pollen limitation of fruit production across habitats, sites and years. Flowers supplemented with outcross pollen were 73% more likely to produce fruits and contained 15% more seeds per fruit than open pollinated flowers. We found no evidence that forest interior plants were more pollen limited in their fruit or seed production than forest edge plants, however.  Fruit and seed production for hand-pollinated forest interior flowers was significantly lower than on the edge, indicating underlying resource limitation of reproduction of forest interior plants.  Incongruence between the indicators of pollinator service and pollen limitation data sets likely reflects pollen quality differences across the two habitats.  Along the forest edge honeybees visited more flowers per plant and likely transferred more geitonogamous pollen than visitors frequenting the forest interior.