Abstract
Pollination networks, which describe the aggregate outcome of pollinators’ foraging decisions among floral resources and the resulting facilitation of reproduction in flowering plants, exhibit highly conserved structural features across diverse habitats and species assemblages. The generality of these patterns strongly suggests general underlying mechanisms, and common outcomes of the coevolutionary process between mutualists. In an empirical study of a novel alpine pollination network I show that ecological models that accurately predict network structural metrics are not necessarily doing so by accurately predicting species interactions, which are best explained by species’ phenologies. Using theoretical models, I show that in the long term, commonly observed network structures naturally emerge from evolutionary feedback between pollinators’ foraging decisions, and the resulting selection on plants’ flowering phenologies. My results also suggest that common network structures are a plausible outcome of multiple theoretical explanations that can be generated by different, and widespread biological mechanisms.
