Determinants of the microstructure of plant-pollinator networks

Abstract

Identifying the determinants of biological interactions in mutualistic networks is key to understanding the rules that govern the organization of biodiversity. We used structural equation modeling and dissimilarities in nine ecological variables to investigate community processes underlying the turnover of species and their interaction frequencies (interaction pattern) among highly resolved plant-pollinator networks. Floral and pollinator community composition, i.e., species identities and their abundances, were strong determinants of the microstructure of pairwise interactions among the networks, explaining almost 69% of their variation. Flower and pollinator traits were directly related to interaction patterns, but were partly masked in the model by shared variance with community composition. Time of year and geographic location, floral and pollinator abundances independent of species identity, and relative abundance of exotic flowers had indirect and relatively weak effects on interaction patterns. Our analyses lead to precise predictions about the processes behind the interaction patterns in mutualistic networks. Future understanding of these processes will be aided by studies that evaluate these predictions experimentally at the network level.