Phenotypic trait matching predicts the topology of an insular plant-bird pollination network.

Conceptualizing species interactions as networks has broadened our understanding of ecological communities. However, the factors shaping interaction patterns among species and, therefore, network structure remain unclear. One potentially important factor is the matching of phenotypic traits. Here, we tested for trait matching in a bird-flower visitation network from New Zealand. We first quantified the overall network structure and tested whether flower size could account for differences in the visitation rates of flowering plants. We then explored the relationship between the flower size and bill size. The results showed that the interaction network is nested. Plant species with large flowers received more visits from birds than plant species with small flowers. Moreover, plant species with large flowers were visited more frequently by birds with large bills, while species with smaller flowers were visited more frequently by birds with small bills. Overall, the interaction patterns between birds and flowering plants could be predicted by their morphology, suggesting that phenotypic trait matching is an important predictor of network structure.