Because flowers are the major unit of angiosperm reproduction, positive and negative floral interactions can have significant impacts on plant fitness. Pollinators are well-studied mutualists that interact with flowers, but there are also many kinds of antagonists that use flowers, including nectar robbers (insects that take nectar without pollinating) and florivores (insects that eat flowers). However, the evolution of floral traits is frequently interpreted solely in terms of interactions with pollinators, even though floral antagonists can also exert selection on floral traits. Floral antagonists may compete among themselves or with mutualists for flowers as a resource. Although interactions between multiple leaf antagonists are commonly examined, very few studies have manipulated multiple floral interactions, despite their potential to interact with one another and have non-additive effects on plant fitness. We manipulated pollination, nectar robbing, and florivory in a factorial design in the annual Impatiens capensis to examine combined effects on plant growth, floral traits, subsequent floral mutualist and antagonist interactions, and plant reproduction.
Results/Conclusions
Artificial florivory was the only treatment that affected plant reproduction, significantly reducing the number of outcrossing flowers, total fruits, and maturing fruits but having no effect on fruit or seed quality or on selfing reproduction. This effect of florivory on outcrossed reproduction was not due to reduced pollinator visitation, suggesting plants may change allocation strategies due to florivory. Hand pollination had little to no effect on any measure of reproduction, indicating I. capensis plants were not pollen limited. None of the treatments affected any measure of plant growth. However, floral treatments did affect leaf herbivory; artificial nectar robbing marginally decreased percent leaf damage, and florivory decreased leaf damage in the presence of hand pollination but increased leaf damage without hand pollination. Together, the effects on leaf herbivory indicate there may be systemic responses to floral interactions, either through induction of defenses or tradeoffs in allocation to leaves and flowers. However, floral treatments had no effect on subsequent natural floral interactions; none of the treatments affected subsequent florivory, pollinator visitation, nectar robbing, or nectar thieves. Overall, this study shows that interactions between flowers and insects can have strong effects both on plant reproduction and on systemic plant resistance, and that antagonistic floral interactions may exert stronger selection on floral traits than interactions with pollinators.