Can you use floral traits to predict pollinators?

The pollination syndrome hypothesis suggests that specific floral traits, such as color and shape, can be used to predict the pollinator. However, recent community-level studies have shown that pollination syndromes may not be as accurate in predicting pollinators as previously thought. In this study we test the pollination syndrome hypothesis by compare the floral traits from several plant-pollinator communities from California to observed interactions. Floral traits were first scored using digital images and then analyzed using a non-metric multidimensional scaling to produce a floral phenotypic space. We then performed additional tests to determine if floral morphology is correlated with visits from specific pollinator taxa. Our results suggest that pollinator taxa do not restrict their visits to flowers with particular trait combinations, but are rather generalized. Thus, it appears that the pollination syndrome hypothesis may be limited in predicting plant-pollinator interactions in a community context.