Two previous hypotheses to account for the coexistence of many tree species (rather than a single-species dominance) and the low density and uniform dispersion of mature trees in tropical forests were those propounded by Janzen (1970) and Connell (1971).
Both invoked seed/seedling predation and Connell emphasized that predation was more important than interspecific competition in preventing single-species dominance. Doubtless, seed predation is widespread and intense in many tropical trees. This is borne out by the oft repeated failure in many cases to collect an intact seed beeneath the parent tree (see Hubbell, 1980).
This is surprising in view of the equally common observation that adult trees frequently grow in clumps of conspecific individuals of all sizes. According to Hubbell, spacing of conspecific adult trees does not, by itself, explain the large number of coexisting tree species in tropical forests unless the trees occur quite far away from each other. In actual observation, the inter-tree distances are quite small.
Thus it seems that factors other than tree spacing limit tree growth. Probably, all microhabitats are not equally suitable for growth and survival of the trees. A due consideration of the interacting processes between a tree species and its seed predators and dispersal agents prompted Hubbell to deduce high variance in the intensity of seed predation and the effectiveness of seed dispersal from tree to tree and from year to year.
He is of the opinion that if even a small fraction of seeds near the parent escape seed predators, then the maximum of the population recruitment curve will lie near the parent and not at some intermediate distance between two adult trees.
This is because many more seeds happen to fall near the parent than are dispersed to greater distances, and this is how one can explain why adult trees are often clumped despite the fact that most of the seeds falling near the parents are eaten by predators (Hubbell, 1980).