Recruitment limitation after mast-seeding in two African rain forest trees.

Seed and establishment limitation can have a major role in determining plant species' abundances and distributions in communities. Their relative importance, however, remains uncertain and controversial, especially for trees in forests where density-dependent mortality of seeds and seedlings may be common. In a primary African rain forest, we directly tested the strength of each limitation by using seeds of the tree species Microberlinia bisulcata and Tetraberlinia bifoliolata at eight addition levels and by following establishment over six weeks. Local background seed rain was also measured. Seedling recruitment was higher in seed-addition quadrats than in control quadrats, indicating seed limitation in both species. However, fitting the Beverton-Holt model indicated that establishment limitation was consistently 2-4 times stronger than seed limitation for Microberlinia, whereas seed limitation greatly exceeded establishment limitation for Tetraberlinia. Strong density dependence was operating in the short seed-to-seedling transition for Microberlinia, whereas it was almost negligible for Tetraberlinia. Although early postdispersal mortality was very high for both species (> 80%), they may achieve local codominance as a result of differing strengths of seed limitation vs. establishment limitation. Assessing the importance of seed limitation for tree populations requires a knowledge of species-specific seed rain as well as a reliable recruitment function. The outcome of early establishment processes also needs to be seen in the context of later stages of tree dynamics.