Many plants use defense chemicals to reduce the harmful effect of insect antagonists. While the defense responses of plants to leaf herbivory have been well-studied, much less is known about defense responses in floral tissues. This is surprising since floral tissues are generally more tightly linked with plant reproduction than leaves. Anthocyanins are secondary chemicals expressed in flowers that can both provide pigmentation to attract pollinators and potentially serve as defenses against florivores. I conducted a manipulative experiment to determine whether floral damage induces anthocyanins in Impatiens capensis flowers, whether induction is systemic or local, the time frame over which induction occurs, and whether induction reduces subsequent florivory. To do this, I damaged one flower per plant at three randomly assigned damage treatments (control, 30%, and 60% tissue removal) and then collected subsequent flowers at set time intervals and specific locations on the plant for analysis of red color and anthocyanin content.
Results/Conclusions
Impatiens capensis increased anthocyanins in subsequent flowers, but only for plants in the 30% damage treatment and only on the same branch as the damaged flower. This suggests that anthocyanins can be locally induced, but only at moderate levels of floral damage. Because 30% damage induced anthocyanins but 60% damage did not, there may be a threshold response to floral damage, above which plants no longer allocate to floral defenses. Our results also show that post-treatment florivory on damaged plants was one third that of control plants, suggesting that damage to a single flower can induce changes that deter subsequent whole-plant florivory. Since the anthocyanin response was localized and the subsequent florivore response was systemic, there may be other floral defense chemicals that have been induced systemically after floral damage. Analysis of red flower color revealed that plants either produced predominantly red or yellow flowers, and that the redness of either flower type was unaffected by the damage treatment. Our results suggest that more study is needed to understand potential induction of defenses in flowers and whether our current understanding of induced defenses in leaves is applicable to flowers.