Mechanisms influencing network topology in plant–hummingbird pollination networks

Ecological communities result from complex species interactions, often summarized in interaction networks. The structure of these networks is described by metrics that provide insight into community assembly, ecosystem functioning and coevolutionary processes. Despite advances in measuring and mapping network structure, the mechanisms underlying its formation remain less explored. Network metrics may vary across communities owing to changes in species diversity and environmental conditions. However, network metrics may remain invariant if mechanisms influencing interactions (linkage rules) are independent of species composition and environmental conditions and instead influenced by traits. We investigated whether changes in taxonomic, phylogenetic and functional diversity along elevation gradients influence network modularity, nestedness, connectance and specialization across 32 sites in Brazil, Costa Rica and Ecuador. Despite elevation's impact on diversity, we found that it had no effect on network structure, which remained consistent across elevations. Instead, trait-based mechanisms, specifically the matching between hummingbird bills and flower corollas, emerged as a consistent driver of network structure. Species showing strong trait matching contributed more to modularity and specialization, but less to nestedness and connectance than expected by chance. These results suggest that trait matching influences the invariant structure of plant-hummingbird networks, persisting despite shifts in diversity across biogeographical regions and elevational gradients.