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.

Fruit availability, frugivore satiation and seed removal in 2 primate-dispersed tree species.

During a mast-fruiting event we investigated spatial variability in fruit availability, consumption, and seed removal at two sympatric tree species, Manilkara bidentata and M. huberi (Sapotaceae) at Nouragues Natural Reserve, French Guiana. We addressed the question of how Manilkara density and fruits at the community level might be major causes of variability in feeding assemblages between tree species. We thus explored how the frugivore assemblages differed between forest patches with contrasting relative Manilkara density and fruiting context. During the daytime, Alouatta seniculus was more often observed in M. huberi crowns at Petit Plateau (PP) with the greatest density of Manilkara spp. and the lowest fruit diversity and availability, whereas Cebus apella and Saguinus midas were more often observed in M. bidentata crowns at both Grand Plateau (GP), with a lowest density of M. bidentata and overall greater fruit supply, and PP. Overall, nearly 53% and 15% of the M. bidentata seed crop at GP and PP, respectively, and about 47% of the M. huberi seed crop were removed, otherwise either spit out or defecated beneath trees, or dropped in fruits. Small-bodied primates concentrated fallen seeds beneath parent trees while large-bodied primate species removed and dispersed more seeds away from parents. However, among the latter, satiated A. seniculus wasted seeds under conspecific trees at PP. Variations in feeding assemblages, seed removal rates and fates possibly reflected interactions with extra-generic fruit species at the community level, according to feeding choice, habitat preferences and ranging patterns of primate species.

Genetic structure and mating system of Manilkara huberi (Ducke) A. Chev., a heavily logged Amazonian timber species.

In this work, we report on the population genetic structure of the endangered tree species Manilkara huberi, an Amazonian tree species intensely exploited due to the high density and resistance of its wood. We investigated the patterns of spatial distribution, genetic structure, and mating system using 7 microsatellite loci and here discuss the consequences for conservation and management of the species. To examine the population genetic structure, 481 adult trees and 810 seedlings were sampled from an area of 200 ha from a natural population in FLONA Tapajós, PA, Brazil. We found relatively high and consistent inbreeding levels (intrapopulation fixation index [f] 0.175 and 0.240) and a significant spatial genetic structure up to a radius of approximately 300 m, most likely due to a limited seed and pollen flow. The multilocus (tm) population outcrossing rate was high (0.995), suggesting that the species is predominantly allogamous with a pollen flow restricted to 47 m. These results suggest that M. huberi is spatially structured, consistent with a model of isolation by distance. Fragmentation may therefore cause the loss of subpopulations, suggesting that management programs for production and conservation should include large areas. The genetic data also revealed that for ex situ conservation, seeds should be collected from more than 175 maternal trees, in order to keep an effective population size of 500. Furthermore, as the species is widely distributed across the Amazon Forest, samples should include several populations in order to represent the highest genetic diversity possible. These results provide a blueprint to guide the production and conservation management policies of this valuable timber species.