Floral traits, pollinator behavior, and plant reproduction: tests of natural and sexual selection in the hummingbird-pollinated herb Chrysothemis friedrichsthaliana

Abstract

Pollinators may promote floral evolution by selectively visiting preferred floral phenotypes. Herbivores may also influence the evolution of floral traits, by reducing the fitness of poorly-defended plants, such as those without protective sepals. In this dissertation, I investigate the reproductive and evolutionary consequences of floral visitors for two floral traits of the Neotropical herb Chrysothemis friedrichsthaliana (Gesneriaceae). I focus on the trait male-biased nectar production, which results from coincident temporal shifts in nectar production rates and sexual function within flowers. First, I review theory relevant to male- and female-biased nectar production, and I generate hypotheses to explain its evolution in 41+ representative species. Next, I test natural and sexual selection hypotheses using C. friedrichsthaliana and its hummingbird pollinator Phaethornis striigularis (Trochilidae). Sexual selection theory predicts that secondary sexual traits (like nectar) serve to increase male more than female pollination success, and in accordance, (1) nectar production was consistently greater during the male phase; (2) visits by P. striigularis were sometimes limiting, indicating that plants had to compete for visits; (3) hummingbirds made more and longer visits to unmanipulated male- versus female-phase flowers; (4) female fecundity was maximized by one visit, whereas male fecundity continued to increase with additional visits. Using nectar manipulations, I also showed that P. striigularis discriminated against unrewarding real and artificial flowers, suggesting that preferences are based partially on nectar. Natural selection for inbreeding avoidance predicts high costs of inbreeding and specific pollinator responses to reward distributions, yet neither prediction was met in C. friedrichsthaliana. Thus, sexual selection, but not inbreeding avoidance, likely maintains male-biased nectar production in this species. I used a second floral trait, the cup-like, liquid-filled calyx of C. friedrichsthaliana, to test the hypothesis that sepal morphology functions predominately for herbivore defense. Experimental calyx draining showed that a liquid barrier over buds halved flower loss to a highly detrimental microlepidopteran herbivore (Alucitidae). Thus, C. friedrichthaliana's water calyx promotes reproduction by protecting developing buds. This dissertation is the first to comprehensively describe the mechanistic basis for both traits' functions. It also lays the groundwork for future, novel explorations into natural and sexual selection on flowers.