Background/Question/Methods: Nectar volatiles may affect flower apparency and reward properties, yet their ecological functions remain poorly understood. Here we address the roles of nectar volatiles in pollinator attraction and antagonist deterrence, focusing on 2-phenylethanol (2PE), a major floral fragrance constituent in the alpine wildflower Polemonium viscosum.

Results/Conclusions: Floral fragrance profiles were measured for P. viscosum in the field to test for correlations of 2PE emission with traits involved in pollinator attraction (nectar production and corolla flare) and larcenist deterrence (corolla length). Emission was poorly correlated with nectar production suggesting 2PE is not an honest signal of reward level. Emission correlated positively with corolla flare and negatively with corolla length. Results suggest that efficacy of nectar volatile signals should vary with flower form.  Long narrow flowers should concentrate 2PE within the floral tube, increasing its impact on nectar-foraging ants at the base of the corolla.  Conversely, shorter more flared flowers should release greater amounts of 2PE into the atmosphere, potentially enhancing its delivery to flying pollinators.  We supplemented floral nectar of plants in natural populations with solutions of sucrose (S); sucrose with weakly (W, twice natural) or highly (H, 4X natural) concentrated 2PE, or sucrose with 4X natural 2PE in a blend of other trace fragrance constituents (B). Surprisingly, presence of trace volatiles more than tripled 2PE emission rate. Results show that trace volatiles may affect floral visitors indirectly by acting as ‘vectors’ to enhance movement of major scent components. To the best of our knowledge this is a novel hypothesis for the evolution of floral fragrance complexity.

Ants of the floral larcenist species, Formica neorufibarbus gelida were exposed to a wide range of 2PE concentration.  2PE was toxic to ants at the 4X concentration level, indicating that it is a potential deterrent to floral antagonists of P. viscosum. We also presented SC and H inflorescences to freely foraging bumblebee pollinators.  Bumblebees showed a strong aversion to flowers supplemented with 2PE, illustrating a potential ecological cost of 2PE mediated floral defense. Last, we studied mechanisms of selection on 2PE in plants with fragrance-supplemented flowers. Increasing 2PE concentration reduced ant damage and nectar consumption (but not pollination). Seed set was reduced in the B treatment compared to other groups.

Overall, results reveal new mechanisms for plant modulation of nectar volatile emission and suggest that volatiles may act to reduce the ecological costs of antagonists and reward production in  pollination mutualisms.