Cryptic dioecy and evolution of pollen morphology in Thalictrum

The majority of flowering plants are hermaphroditic, but repeated evolution of separate sexes (dioecy) in diverse lineages indicates that this mating system is favored in a variety of ecological and selective environments. For my dissertation research, I studied species within the flowering-plant genus Thalictrum (Ranunculaceae), which includes an uncommon diversity of reproductive systems, including independent and repeated evolutionary events that yielded wind-pollinated and dioecious lineages from insect-pollinated and hermaphroditic ancestors. Thalictrum macrostylum is a morphologically androdioecious species; populations are consistently dimorphic with male plants that bear staminate flowers (producing pollen) and morphologically hermaphroditic plants whose flowers contain both stamens and pistils (male and female reproductive structures, respectively). I demonstrate that T. macrostylum is cryptically dioecious: pollen from morphological hermaphrodites (cryptic females) is sterile. I also establish that T. macrostylum is sexually dimorphic for some traits in directions consistent with other dioecious angiosperms; other traits do not differ significantly between the sexes, which may indicate that the species recently evolved separate sexes. Further, I refute the commonly cited hypothesis---pollinator attraction---for the maintenance of these sterile structures in females. The species is not significantly insect-pollinated, and the presence of stamens in flowers of cryptic females does not enhance female reproductive success. Data also refute a proposed beneficial effect of sterile stamens on the aerodynamics of pollen deposition. Additionally, I survey the genus Thalictrum to test whether evolution in pollen morphology (specifically grain size and aperture number) is correlated with shifts in pollination mode (wind or insect), sexual system, geographic distribution, or floral morphology. I find evidence of selection for higher aperture numbers in insect-pollinated species and for smaller size in wind-pollinated species. I also uncover instances of correlated evolution between pollination mode and floral morphology. In the final chapter, I discuss the implications of my work on cryptic dioecy as a whole and look toward future avenues of study related to the evolution and maintenance of cryptic dioecy.