Premise of the Study: Animal pollination is critical to plant reproduction and may cause convergent evolution of pollination syndromes. Pollination syndromes in Castilleja are hypothesized based on floral traits and historical observations of floral visitors. Here we address these questions: (i) Can pollination syndromes be distinguished using floral morphological traits or volatile organic compound emissions? (ii) Is there significant variation in floral traits within a pollination syndrome, at the level of populations or species? and (iii) Do pollination syndromes predict the most frequent floral visitor to Castilleja?

Methods: Floral traits and visitation were measured for five co-occurring Castilleja species (C. applegatei, C. linariifolia, C. miniata, C. nana, and C. peirsonii), representing three pollination syndromes (bee, fly, and hummingbird), at four sites in the Sierra Nevada Mountains. We used non-metric multidimensional scaling (NMDS) and multiple linear regressions to address key questions in the differentiation of Castilleja and floral visitors.

Key Results: Our analyses revealed that both morphological traits and floral VOCs distinguish between some pollination syndromes and Castilleja species. Morphological traits defined pollination syndromes reliably, but within the hummingbird syndrome, there was also significant variation among populations and species. Pollination syndrome was a poor predictor of visitors to Castilleja.

Conclusions: Floral trait differentiation among Castilleja individuals reflects both taxonomy and pollination syndromes. Differentiation was generally more evident in morphological traits compared to VOCs. Furthermore, a priori notions of pollination syndromes in this system are overly simplistic and fail to predict which animals most frequently visit Castilleja in natural populations.