Fig. 3: Pesticide risk and composition and agricultural pollen in relation to bee species and landscape context.
From: Ecological traits interact with landscape context to determine bees’ pesticide risk

a, Results showed that pollen-based pesticide risk increased with the amount of agricultural land in the landscape for B. terrestris and O. bicornis, while A. mellifera pollen-based risk was independent of agricultural land extent. b, The proportion of agricultural land also influenced pollen use, with only O. bicornis using more agricultural pollen with increasing agricultural land. c, Risk for A. mellifera correlated with that of O. bicornis (grey) and B. terrestris (yellow). d, The composition of pesticide compounds in pollen differed between A. mellifera and O. bicornis, while B. terrestris overlapped the two based on PERMANOVA of Bray–Curtis dissimilarities. Dispersion varied between bee species (P = 0.03); therefore, these community differences should be interpreted cautiously. Predictions and 95% confidence intervals (a,b,c) come from linear models with risk log transformed and the proportion agricultural pollen logit transformed. NMDS points (d) are based on standardised Bray–Curtis distances.