The forest archipelago of the East Usambara Moutains, Tanzania, holds the richest remnants of 30 million year old rainforests that once extended across Africa. Endemic tree species, comprising 24% of the tree flora, include overstory and understory species of all successional stages and a variety of pollination and dispersal systems. This investigation tests potential reasons why Leptonychia usambarensis (Sterculiaceae), an endemic bird-dispersed successional tree, reproduces less well in forest fragments < 30 ha than in nearby continuous forest of 3500 ha. Forest fragmentation in tropical systems may have biotic or abiotic consequences for trees. Biotic consequences include pollinator or disperser limitation from loss or reduction of insects, birds or mammals that pollinate flowers or disperse seeds, reduction of seed or seedling survival from pathogens or animals that are more prevalent or more devastating in disturbed habitats, and inbreeding depression among small isolated populations of only a few related individuals. Abiotic effects can be numerous, but usually reflect responses to altered light regimens with the increase in edge to area ratio that occurs with reduction of forest patch size. The proposed study would extend for one year an ongoing investigation of the consequences of forest fragmentation. Our results to date show that birds that disperse seeds of this species in continuous forest are reduced in numbers or absent in fragments, that fewer seeds are dispersed away from parent trees, that more seedlings occur near parents than further away in fragments than in continuous forest, and that juvenile plants more than one year old recruit less well in fragments than in extensive forest. Remaining issues in need of evaluation require a reciprocal seedling transplant experiment that will test potential inbreeding effects and habitat and edge effects on regenerating seedlings. Two thousand establishing seedlings from 20 continuous forest and 20 fragment trees will be transplanted to edge and interior locations to test: (1) presence or absence of potential inbreeding effects under competitive field conditions, and (2) edge and interior habitat suitability of seedlings from continuous forest trees in both extensive and fragmented habitats. If either or both effects are shown, further support will be requested to dissect their causes and consequences for this and other endemic and pan-African trees species. The broader impacts of this study relate to a clearer understanding of the causes of diversity and rarity in tropical forests. These forests typically contain a few common species together with many species that are present in low abundance or are rare in any given unit area. This fact makes tropical trees especially vulnerable to forest fragmentation, leaving the majority in low densities in small relict patches. This study has been the first to clearly demonstrate that tree species in tropical forest fragments are dispersal limited, and it may be the first to clearly document an environment by seed source interaction that would alert ecologists to the ubiquitous threat of inbreeding depression in small relict stands. The study supports graduate training of an African student.
