Bumblebees are more efficient than honeybees to facilitate wind-blown pollen dispersal of alfalfa (Medicago sativa L.)

Abstract

An understanding of the visiting behavior of the pollinator on a crop and the affecting factors on gene flow would help to predict the potential gene flow risk and develop the strategies to limit gene flow. In alfalfa (Medicago sativa L.), pollination requires a tripping mechanism by a pollinator to release the pollen. This typical pollination method implies that the variations in tripping efficiency among bee species may directly or indirectly affect alfalfa pollen dispersal and gene flow. The objective of this study was to quantify and compare the difference in the formation of alfalfa pollen cloud density resulting from the two distinct bee species, honeybees (Apis mellifera L.) and bumblebees (Bombus terrestris L.), and their indirect effects on the promotion of wind-blown pollen dispersal under turbulent or non-turbulent weather conditions. In this study, pollen collection using rotorod pollen collectors under caged or uncaged alfalfa plots indicated no alfalfa pollen released from the source in the absence of insects. In contrast, pollens detected within, and beyond the alfalfa pollen source evidenced the necessity of tripping the flowers by bees for pollen release. Bee species greatly affected the visiting duration of a single alfalfa flower and the number of tripped flowers min−1. At the current experimental scale, although worker bees of bumblebees were less abundant than honeybees, they generated a significantly greater alfalfa pollen cloud density (mean across sampling times, dates, and weather conditions: 566 pollens m−3 h−1) than that of honeybees (416 pollens m−3 h−1), which was closely correlated with the higher number of tripped flowers by bumblebees (8.8 flowers min−1) relative to honeybees (3.1 flowers min−1). While the similar patterns of wind-blown alfalfa pollen density negatively correlated with the distance observed for the two bee species, the pollen density generated by bumblebees at the same distance from the pollen source was greater than the values by honeybees. Additionally, a directional wind effect was detected with the greater pollen density always observed at downwind sites. The developed model predicted that alfalfa pollen cloud density reached 95% reduction of maximum potential value at 31.6 m (10 pollens m−3 h−1) from the source edge, providing the reference value to understand the wind-blown alfalfa pollen dispersal and create isolation distance. Thus, the results will be helpful to understand the factors affecting the visiting behavior of pollinating bees associated with the flowers tripping, predict the gene flow risk, and develop management strategies to mitigate gene flow in alfalfa.