The intensity of nectar robbery (NR) – i.e. the proportion of flowers experiencing robbery – commonly varies both spatially and temporally. Multiple drivers of such variability have been postulated, yet no study has evaluated which are operating in specific instances of heterogeneous NR intensity. Without a mechanistic understanding of the ecological drivers of nectar robbery, it is difficult to predict the circumstances under which it will occur.
In this study, I investigated a pattern of spatial heterogeneity in NR by meliponine bees of the shrub Odontonema cuspidatum in a coffee agroecosystem in Mexico, where plants growing in forest fragments experience ~30% more NR than those in adjacent coffee farms. Through a combination of surveys and field manipulations, I investigated the following potential drivers of this pattern: 1) higher robber abundance in forest than coffee; 2) greater competition with legitimate pollinators in coffee; 3) differences in nectar quantity or quality between habitats; 4) differences in O. cuspidatum abundance; 5) differences in alternative floral resource availability; or 6) differences in abiotic conditions affecting bee foraging. I also investigated whether NR heterogeneity leads to differential reproductive output in O. cuspidatum.
Results/Conclusions
Alternate floral resource availability and/or nectar characteristics drive heterogeneity in NR. We found no difference in robber abundance, pollinator (hummingbird) abundance, or pollinator behavior between habitats. Nor was there difference in abundance of O. cuspidatum blooms. However, heterospecific floral resources (abundance and richness) were significantly greater in farms, due primarily to greater abundance of flowering weeds in the farm. Neighborhood floral abundance was inversely related to NR intensity suggesting that the availability of alternate floral resources relieves NR pressure on O. cuspidatum. However, O. cuspidatum nectar volume was significantly lower in blooms from forest patches, alternatively suggesting that lower nectar quantity spurs increased robbing in forest (more flowers must be robbed to extract equivalent resources). Further investigations to be conducted in June 2017 will determine which of these mechanisms is operating, as well as the role of nectar sugar concentration.
NR reduces fruit and seed set in O. cuspidatum across habitats; thus O. cuspidatum fitness likely is higher in farm than forest.
These results provide the first direct demonstration that heterogeneity in NR intensity can be generated by characteristics of the floral community. They also highlight the potential importance of weeds in mediating plant–nectar feeder interactions in agroecosystems.