Traditional pollination syndromes group angiosperms into categories based on how floral traits impact the functional group of pollinator most associated with those traits. The concept, while well supported for specialist-pollinated plants, is a poor predictor of pollinator identity in generalist systems, such as those common to the family Asteraceae. One potential avenue for future refinement of the concept is the combination of large floral trait datasets, quantitative pollinator data, and phylogenetic comparative methods. Helianthus is a well-studied genus of North American aster whose species include the agriculturally significant H. annuus, which represents the third largest oilseed crop globally. The genus is primarily bee pollinated and, while much is known about traits that are correlated with bee attraction at short ranges common to agricultural and horticultural settings, there has been little research on long range visual and chemical attraction traits within the genus. Using data on display size and shape, ray color, floral volatile composition, and floret depth collected from Helianthus species grown in a common garden, mixed models were constructed to predict pollinator visitation as a function of floral traits. For four of seven pollinator response variables, there was at least one model that outperformed null models, and three of the four best models were multivariate. This work will inform future research of pollination syndromes within generalist systems such as those common to Asteraceae.

Charles Pitsenberger

Advisor: Dr. Chase Mason