Print PageForest fragmentation is pervasive in tropical landscapes, and animal-dispersed, non-pioneer species are one of the groups most at risk from fragmentation. Early regeneration stages are often the most vulnerable, but fragmentation may differentially affect various components of regeneration. The objective of this study was to evaluate forest fragmentation impacts on multiple early regeneration components of an animal-dispersed, non-pioneer canopy tree, Tapirira mexicana, in a highly fragmented region of southern Costa Rica. The following hypotheses were tested in different-sized forest fragments: (1) pre-dispersal predation and fungal infection will be lower in smaller fragments; (2) inherent seed germination will be poorer in smaller fragments; (3) secondary dispersal and post-dispersal predation will be lower in smaller fragments; (4) in situ seed longevity will be shorter in smaller fragments; and (5) 1-yr seedling establishment will be reduced with decreasing distance from forest edges. Seeds were cut open and inspected for pre-dispersal predation and fungal infection. Inherent germination was examined by germination tests in a screen house. Post-dispersal fate of seeds was tracked using threaded fruits. In situ seed longevity was recorded by coloration of fruits. Finally, number of 1-yr seedlings in transects was plotted against distance from edges for any evidence of edge effects.
Results/Conclusions
Pre-dispersal predation was a major filter for regeneration of T. mexicana. Over 60% of the seeds were predated by fly (Anastrepha; Tephritidae) larvae before dispersal, which was significantly higher in smaller fragments than the largest forest fragment in the immediate area. However fungal infection was generally low with no difference among different-sized forest fragments. Another major filter was germination. Even seeds free of pre-dispersal predation or infection failed to germinate under full radiation in the screen house. Longevity of seeds was generally short and even in the forest floor seed lost its viability within 10 days on average, which did not differ among forest fragments. These facts suggest that seeds of T. mexicana are recalcitrant without dormancy and when heated or desiccated, lose their viability very quickly. Once fruits were on the ground, both secondary dispersal and post-dispersal predation were low. Finally, seedling establishment increased linearly with distance from forest edge, which suggests an evidence for edge effects. Although seedlings of T. mexicana are currently fairly abundant in forest interior due to its high fecundity, this study shows the vulnerable components of regeneration that may drive populations down in small fragments.
