MACROEVOLUTIONARY TESTS OF POLLINATION SYNDROMES: A REPLY TO FENSTER ET AL.

Studies of floral ecology and evolution are often centered on the idea that particular floral trait combinations, or syndromes, represent adaptations for particular pollinators. Despite the conceptual importance of pollination syndromes, few macroevolutionary studies have statistically examined the relationship between pollinators and floral traits. Using 15 species of Iochroma, Smith et al. applied phylogenetically structured correlation analyses to test the relationship between floral variation and pollination system, quantified in terms of the importance of major pollinator groups. This study revealed that pollinator shifts are tied to changes in nectar reward and floral display but are not significantly correlated with changes in corolla length or color, contrary to what might be predicted from classical pollination syndromes. Fenster et al. question these findings because our pollinator importance estimates included recently introduced honey bees. To address this concern, we recalculated importance values excluding honey bees and repeated the analyses. We found the same patterns as in our original study with significant correlations between pollinators and nectar reward and display. We conclude that phylogenetic approaches provide essential tools for testing macroevolutionary predictions of pollination syndromes and, by applying these approaches to other radiations, we can refine our understanding of the role of pollinators in floral diversification.