COS 249-2 - Why did the bee cross the road? Examining the influence of roads on solitary bee foraging movements.

Abstract: Roads are prominent features in human-dominated landscapes, yet little is known about how they influence the movement of foraging insects. In summer 2022, we studied whether roads posed barriers to movement of the solitary bee, Agapostemon virescens, in Medford, Massachusetts using a mark-recapture experiment. We established three study sites: 25x25m lawn squares bisected by an active road with a pot of flowering Echinacea purpurea at each corner. Over two weeks, we uniquely marked foraging bees and recorded the pot where initial capture and subsequent recaptures occurred. A bee recaptured at a different pot than the previous capture was recorded as either an along-road or an across-road movement. We estimated the proportion of across-road movements out of all orthogonal movements using a binomial generalized linear model (GLM). If roads were not barriers, we expected a null proportion of 50%. We also analyzed how the number of flowers at the destination pot and the distance moved influenced the probability of movement within and among squares using a binomial GLM. Lastly, we explored variation in the number of foragers at each square relative to the number of flowers, average traffic intensity, and the proportion impervious surface within 100m.



We marked 90 bees, 68 of the marked bees were recaptured at least once, and 54 were recaptured at least twice. On average, each bee was captured 5 times. We found that roads were partial barriers to movement. A. virescens was twice as likely to move along the same side of the road than across it, with 35% (CI95 = 0.25,0.46) of bees making road crosses. There were no significant differences in the proportion of road crosses between sites (χ² = 0.827; df = 2; P= 0.661). Distance had a significant negative effect on movement (estimate = -0.001; SE = 0.001; z value = -8.483; p < 0.001), and the number of flowers had a significant positive effect on movement (estimate = 0.077; SE = 0.0194; z value = 3.951; p < 0.001). The number of foragers visiting each square differed significantly from one another, and forager abundance was negatively correlated with impervious surface but not the number of flowers or traffic intensity. Our results provide compelling evidence that roads impede but do not completely block foraging bees. This study demonstrates that small patches of habitat serve as foraging resources for bees in cities and underscores the strength of population level experiments.