Honey bee toxicological responses do not accurately predict environmental risk of imidacloprid to a solitary ground-nesting bee species.

Honey bees (Apis mellifera) are the current model species for pesticide risk assessments, but considering bee diversity, their life histories, and paucity of non-eusocial bee data, this approach could underestimate risk. We assessed whether honey bees were an adequate risk predictor to non-targets. We conducted oral and contact bioassays for Leioproctus paahaumaa, a solitary ground-nesting bee, and A. mellifera, using imidacloprid (neonicotinoid) and dimethoate (organophosphate). The bees responded inconsistently; L. paahaumaa were 36 and 194 times more susceptible to oral and topically applied imidacloprid than A. mellifera, but showed comparable sensitivity to dimethoate. Furthermore, the proposed safety factor of ten applied to honey bee endpoints did not cover the interspecific sensitivity difference. Our standard-setting study highlights the urgent need for more comparative inter-species toxicity studies and the development of standardized toxicity protocols to ensure regulatory pesticide risk assessment frameworks are protective of diverse pollinators.