Doctor of Philosophy (PhD)


Biological Sciences

Document Type



Seed dispersal and seedling establishment – the two stages in seedling recruitment – set the spatiotemporal distribution of new individuals in plant communities. Diversity often increases at the seed to seedling transition, making it critical for species coexistence. Debate continues regarding the effects of each stage on the community structure of diverse forests. Neutral theories postulate a strong role of dispersal, whereas niche-differentiation theories suggest that environmental conditions may be more important. This dissertation tested the effects of dispersal, competition and predation on the structure of the seedling layer in a pristine Amazonian rainforest. Seed-addition experiments broadly tested the relative importance of dispersal and environmental conditions on seedling community structure. Dispersal treatments explained more variance in community structure than did environmental conditions. This was the first variance- partitioning study to show that dispersal affects not only seedling density, but also diversity and species composition. Two more narrowly focused studies tested the intensity of competition among seedlings, and examined the effects of various mammalian predators on seedling recruitment. Evidence for inter-seedling competition was weak: individual growth and survival rates were generally unrelated to stem density, and seedlings’ zones of influence rarely overlapped substantially. As predators, small and medium-sized mammals reduced seedling density, whereas large mammals had no detectable effects. Furthermore, small mammals generated a rare-species advantage, the fundamental element of frequency dependence. Integrating the three studies, we suggest that dispersal is more important for seedling community structure than are environmental conditions. Given the low density of seedlings in xii Neotropical forests, we infer that competition among tree seedlings is largely irrelevant to their recruitment. Seed predators operated in a distinctly non-neutral manner, preferentially removing seeds of common and large-seeded species. Despite the powerful effects of predation, dispersal explained more variance in seedling recruitment than did all aspects of environmental variation (including predation). Taken together, the results of these three experiments support a view that, at least for young plants, and at small scales, dispersal may more strongly influence the species composition of tropical trees than environmental conditions, consistent with predictions from neutral models.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Kyle E. Harms