The structure of plant-pollinator interactions in montane meadow environments

The assumption that plant-pollinator relationships are ecologically specialized is central to the tradition view of angiosperm diversification. The pollination syndrome hypothesis maintains that specialized pollinators drive floral evolution, and thus it is possible to predict the pollinator based on stereotypical combination of floral characteristics. Recently however, biologists have suggested that mutualistic interactions are rather generalized and that even complex floral morphologies may not necessarily represent adaptations to specialized pollinators. To test these ideas, I conducted a community-level survey in a montane meadow system in both California and in Colorado. Analyses of plant-pollinator interaction networks suggest that these communities are highly nested, with most species involved in asymmetrical interactions with generalist partners, whereas reciprocally specialized plant-pollinator interactions were relatively rare.; To test the predictive utility of the pollination syndromes in a community context, I scored plant species for the presence/absence of 40 floral characteristics and attempted to classify each species into a syndrome based on its overall similarity to idealized syndrome descriptions derived from the literature. For most plant species the syndrome classifications failed to accurately predict the most frequently observed pollinator. Traits that influence access to floral rewards appeared to be more important in structuring plant-pollinator interactions, as opposed to floral signals of the presence of a reward, such as color or scent.; A corollary of the pollination syndrome hypothesis states that flowers will evolve adaptations to the most frequent and effective pollinator. To test this idea, I removed and analyzed the pollen carried by insect pollinators. I used MANOVAs to test for an association between the frequency of visitation and pollen load composition and found that the number of conspecific pollen grains and the proportion of conspecific pollen carried by insects was positively associated with the frequency of interaction. If pollen deposition is correlated with insect pollen load composition, then visitation patterns might reflect pollination patterns. However, the assumptions regarding ecological specialization of plant-pollinator interactions, and the predictive utility of syndromes are not supported in a community context, suggesting that some traditional views in pollination biology should be reevaluated.