Potential exposure of pollinators to neonicotinoid insecticides from the use of insecticide seed treatments in the mid-southern United States.

Research was done during 2012 to evaluate the potential exposure of pollinators to neonicotinoid insecticides used as seed treatments on corn, cotton, and soybean. Samples were collected from small plot evaluations of seed treatments and from commercial fields in agricultural production areas in Arkansas, Mississippi, and Tennessee. In total, 560 samples were analyzed for concentrations of clothianidin, imidacloprid, thiamethoxam, and their metabolites. These included pollen from corn and cotton, nectar from cotton, flowers from soybean, honey bees, Apis mellifera L., and pollen carried by foragers returning to hives, preplanting and in-season soil samples, and wild flowers adjacent to recently planted fields. Neonicotinoid insecticides were detected at a level of 1 ng/g or above in 23% of wild flower samples around recently planted fields, with an average detection level of about 10 ng/g. We detected neonicotinoid insecticides in the soil of production fields prior to planting at an average concentration of about 10 ng/g, and over 80% of the samples having some insecticide present. Only 5% of foraging honey bees tested positive for the presence of neonicotinoid insecticides, and there was only one trace detection (< 1 ng/g) in pollen being carried by those bees. Soybean flowers, cotton pollen, and cotton nectar contained little or no neonicotinoids resulting from insecticide seed treatments. Average levels of neonicotinoid insecticides in corn pollen ranged from less than 1 to 6 ng/g. The highest neonicotinoid concentrations were found in soil collected during early flowering from insecticide seed treatment trials. However, these levels were generally not well correlated with neonicotinoid concentrations in flowers, pollen, or nectar. Concentrations in flowering structures were well below defined levels of concern thought to cause acute mortality in honey bees. The potential implications of our findings are discussed.