Individual and neigborhood effects on tropical tree pollination

Background/Question/Methods

Despite conservation efforts, tropical trees face constant threats, with land conversion to agriculture and pasture continuing to be the primary drivers of global deforestation. Therefore, it is essential to develop an understanding of the processes influencing the reproduction of tropical tree species and viability of their populations. Most tropical trees are pollinated by animals, largely outcrossing and self-incompatible. For these species, the fecundity of the individuals is influenced by the availability of pollinators, their attractiveness to the pollinators and the availability of pollen donors in the neighborhood. The conspecific neighborhood can influence both quantity (number of seeds) and seed quality (genetic composition) of the trees. This is because the density of conspecific individuals impacts the amount of available pollen, while the genetic relatedness in the neighborhood could affect the quality of the pollination mediating the amount of biparental and shared-ancestry-related inbreeding. Despite the importance of individual and neighborhood for both seed quality and quantity, relatively little is known about density-dependent patterns in tropical tree reproduction. We hypothesize that bigger individuals in more dense neighbourhoods should present higher fecundity, but that this positive effect is negatively impacted by the level of genetic relatedness in the neighbourhood. We surveyed the 50 ha plot at Barro Colorado Island, Panama for all the reproductive adults of the common understory tropical tree Miconia affinis (N = 124) and estimated the seed set for 20 fruits within 20 randomly chosen individuals. We then used seven microsatellite markers to examine the interaction between neighbourhood density, genetic relatedness and seed set.  

Results/Conclusions

First, our results reveal that M. affinis reproductive trees exhibit significant spatial genetic structure at the 100 meters scale (r = 0.013, p < 0.01) with greater relatedness than expected at random. Beyond this spatial scale, the spatial genetic structure is not significantly different from the random expectation.

Second, our neighborhood analysis of reproduction in M. affinis provides evidence for positive density-dependent seed set. However, we find support for an interaction between the local density of conspecific trees and the genetic relatedness of these trees in determining levels of seed set across the study site. In addition, we find that the size of trees was negatively related to seed set, contrary to our expectations. We suggest that the combination of spatially explicit models and population genetic data constitute a powerful tool to evaluate the potential drivers of the reproductive success in tropical tree communities.