Analysis of a continent-wide seed dispersal network in Africa at the family level

Background/Question/Methods

We aggregated observations of frugivory and seed dispersal from the African continent into a database (www.africatreedatabase.com) to evaluate broad-scale trends in plant-animal relationships. Twenty-two journals have been examined, resulting in >4000 records from published literature linking plants and animals. Due to coverage of a large geographic area, we analyzed interactions at the level of plant families and animal families rather than individual species. Seed dispersal/frugivory network metrics (nestedness and modularity) of family-level interactions were calculated on a binary interaction matrix and compared to the values of random matrices generated from a null model constraining marginal matrix sums from a weighted matrix. We used multi-level modularity optimization to identify different modules of closely associated plant and animal families with the R package ‘igraph.’ We calculated nestedness using NODF in the R package ‘vegan.’ 

Results/Conclusions

Preliminary results show our continent-wide, family-level network was more nested (NODF= 52.4) and modular (M=0.285) than expected. High nestedness indicates that specialist plant (or animal) families interact with generalist animal (or plant) families, rather than specialist plants and animals interacting only with each other. Seven modules were identified with the following animal groups: 1) browsing & grazing ungulates and gulls, 2) monkeys and baboons, 3) elephants, bats, and some birds, 4) most bird families, 5) savanna-dwelling families (e.g. rhinos, ostriches, grasses), 6) rodents and hornbills, and 7) apes and turacos. Each module (except 2) contained at least one bird family, but most of the bird families were in one module. Many of the families (especially birds, fruit bats, and some plants) have only a few records in the database, so more data may further differentiate or connect modules. Analyses of frugivory and seed dispersal networks at a broader continental scale can highlight which animal and plant groups form important relationships, and offer insights in the ecological and evolutionary pathways that define these relationships. This approach could be applied to existing data sets at different taxonomic levels or other continents.