Plant-animal mutualisms for pollination and seed dispersal involve highly diversified interactions whose implications for coevolution have been difficult to track. Here we explore the ecological and phylogenetic correlates of interaction patterns in highly diversified, non-specific mutualisms between plants and frugivores that disperse their seeds. We use two highly resolved webs derived from an extended sampling of frugivory and fruit diet in two S Spanish communities. We use phylogenetic analyses by independent contrasts and phylogenetic variance partitioning of the number of interactions per species (i.e., degree), and centrality of interacting species, a variable that characterizes their position in the web of interaction. In addition, we explore how species-specific topological positions in the web correlate with local abundance, size, geographic range, and phenological spread. Finally we examine whether the interaction pattern itself is influenced by the phylogenetic history of the interacting plant and animal groups of species. 
Results/Conclusions

Phylogenetic effects are non-significant in explaining variation in degree and centrality of the plant species; however, they do explain a significant fraction of variance in degree across animal species. Animal taxa with high connectivity and/or centrality in the web often are evolutionarily related but this trend is not observable among plant species. Both the degree and centrality of frugivore species were positively correlated with variation in local abundance, and species-specific traits such as gape width, size of geographic range and phenological spread (length of period present in the community). For plants, both geographic range and phenological spread were positively correlated with degree and centrality, but not consistently in the two communities studied. The interaction pattern itself, as described by the interaction matrix among plant and animals, was significantly influenced by the plants phylogeny in the two communities, the animal phylogeny having barely a marginal influence. Our results indicate that species-specific traits of the frugivores have a large effect on how these species interact with fruiting plants, with closely related species showing similar interaction patterns. However, the overall interaction pattern depicted in the network is more influenced by the plant phylogeny, suggesting that the plant assemblage drives the interaction. Plant-frugivore coevolution is suggested to be a process of resource tracking by frugivores so that each plant species 'filters out' subsets of frugivore species given species-specific traits that constrain their interaction. The overall network of interaction thus has a distinct signal marked by the plants phylogenetic history.