Reproductive biology of flowering plants on Pohnpei

Oceanic islands have long been considered natural laboratories of evolution. Present-day biotas of oceanic islands can provide better understanding of the effects of island age, size, and distance from continents on trait evolution. For plants, small, isolated oceanic islands should have the strongest establishment filter, relative to larger, less remote islands. A key step in understanding the establishment filter on such islands is to characterize the reproductive biology of current inhabitants relative to mainland source areas. In this dissertation, I investigate the reproductive biology of 84 angiosperm species that are considered indigenous (resident) or naturalized (recent colonist) on Pohnpei Island, Micronesia. I then ask how they differ from their continental mainland relatives. In the first chapter, I asked if there was a difference between breeding systems of native versus naturalized species on Pohnpei, and I found most taxa had some degree of autogamy and there were no significant differences between several determinants of island breeding systems. In the second chapter, I compared Pohnpei island breeding systems with those of their closest continental mainland relatives. I asked if there were differences in: (1) self-compatibility status, and (2) sexual systems. I found that Pohnpeian species had higher rates of autonomous self-pollination and self-fertilization, whereas their mainland relatives had more outcrossed breeding systems even though all were self-compatible. I also found that all Pohnpeian species and all but three continental mainland relatives had bisexual flowers and, when documented, hermaphroditic sexual systems. The high potential for self-pollination could be due to depauperate pollinators on Pohnpei island, as also indicated by my finding of high levels of pollen limitation. Finally, in the third chapter, I tested two hypotheses that arise from my finding of higher levels of autogamy on islands. First, floral sizes on islands should be smaller than on mainlands, due to less reliance on pollinator attraction. Second, floral allometry should shift such that island styles are proportionally shorter than styles of mainland flowers, due to relaxed sexual selection on style length on islands. Contrary to most studies, island flower size (petal length) was not smaller than that of continental mainland relatives. On the other hand, style length was significantly smaller in island species relative to their mainland sister taxa. These mainland sister taxa did not have unusual petal or style lengths compared to the species means of each of their mainland genera. The similar patterns across many unrelated plant colonizations of an island, strongly support the conclusion that smaller style lengths evolved independently of petal size post establishment