Abstract
A combination of field experiments and allozyme studies was used to test whether two spring-ephemeral plants growing in eastern North America, Trillium erectum and T. grandiflorum (Liliaceae), exhibited an optimal outcrossing distance for fruit and seed production. Furthermore, the spatial genetic structure of the Trillium populations was examined in light of the outcrossing results. In field experiments, recipient plants were pollinated with either self pollen or with pollen from donors growing 1 m, 10 m, 100 m, and 1500 m away. These distances represented crosses between individuals growing within the same population (1 m, 10 m, and 100 m crosses) and between individuals growing in different populations (1500 m crosses). Self-pollinated T. erectum and T. grandiflorum produced 71% and 89% fewer seeds, respectively, than all other outcrossing treatments. However, there were no significant differences among outcrossing treatments for fruit or seed production. As neither T. erectum nor T. grandiflorum exhibited an optimal outcrossing distance for fruit or seed production, it was predicted that populations of the two would not demonstrate strong spatial genetic structure or isolation by distance. The allozyme results only partially supported the patterns revealed in the outcrossing treatments. Populations of T. erectum and T. grandiflorum showed moderate spatial genetic differentiation based on F-statistics, and only T. grandiflorum exhibited significant isolation by distance based on spatial autocorrelation analyses. The lack of optimal outcrossing distances and the patterns of allozyme variation in T. erectum and T. grandiflorum populations may be attributed to a number of factors, including active seed dispersal by ants, rare long-distance gene-flow events, post-pollination and post-fertilization selection, and/or the severity of inbreeding depression.