Reconstructing the pollinator community and predicting seed set from hydrocarbon footprints on flowers.

The measurement of insect visits to flowers is essential in basic and applied pollination ecology studies but often fraught with difficulty. Floral visitation is highly variable, and observational studies are limited in scope due to the considerable time necessary to acquire reliable data. The aim of our study was to investigate whether the analysis of hydrocarbon residues (footprints) deposited by insects during flower visits would allow reconstruction of the visitor community and prediction of seed set for large numbers of plants. In 3 consecutive years, we recorded bumblebee visitation to wild plants of comfrey, Symphytum officinale, and later used gas chromatography/mass spectrometry (GC/MS) to quantify bumblebee-derived unsaturated hydrocarbons (UHCs) extracted from flowers. We found that the UHCs washed from corollas were most similar to the tarsal UHC profile of the most abundant bumblebee species, Bombus pascuorum, in all 3 years. The species composition of the bumblebee communities estimated from UHCs on flowers were also similar to those actually observed. There was a significant positive correlation between the observed number of visits by each of three bumblebee species (contributing 3-68% of flower visits) and the estimated number of visits based on UHC profiles. Furthermore, significant correlations were obtained separately for workers and drones of two of the study species. Seed set of comfrey plants was positively correlated to overall bumblebee visitation and the total amount of UHCs on flowers, suggesting the potential for pollen limitation. We suggest that quantifying cumulative footprint hydrocarbons provides a novel way to assess floral visitation by insects and can be used to predict seed set in pollen-limited plants.