Experimental manipulation of seed shadows of an Afrotropical tree determines drivers of recruitment

The loss of animals in tropical forests may alter seed dispersal patterns and reduce seedling recruitment of tree species, but direct experimental evidence is scarce. We manipulated dispersal patterns of Manilkara mabokeensis, a monkey-dispersed tree, to assess the extent to which spatial distributions of seeds drive seedling recruitment. Based on the natural seed shadow, we created seed distributions with seeds deposited under the canopy ("no dispersal"), with declining density from the tree ("natural dispersal"), and at uniform densities ("good dispersal"). These distributions mimicked dispersal patterns that could occur with the extirpation of monkeys, low levels of hunting, and high rates of seed dispersal. We monitored seedling emergence and survival for 18 months and recorded the number of leaves and damage to leaves. "Good dispersal" increased seedling survival by 26%, and "no dispersal" decreased survival by 78%, relative to "natural dispersal." Using a mixed-effects survival model, we decoupled the distance and density components of the seed shadow: seedling survival depended on the seed density, but not on the distance from the tree. Although community seedling diversity tended to decrease with longer dispersal distances, we found no conclusive evidence that patterns of seed dispersal influence the diversity of the seedling community. Local seed dispersal does affect seedling recruitment and survival, with better dispersal resulting in higher seedling recruitment; hence the loss of dispersal services that comes with the reduction or extirpation of seed dispersers will decrease regeneration of some tree species.