[Review] Have suitable experimental designs been used to determine the effects of neonicotinoid insecticides on bee colony performance in the field?

Laboratory studies have shown than neonicotinoid insecticides can cause sub-lethal effects on bees. Field studies are needed to determine whether these effects also occur when bees forage on flowering crops grown from neonicotinoid-treated seeds. However, for many reasons the results of field experiments may be inconsistent. For example, neonicotinoid residues in crop nectar and pollen and the availability of alternative nectar and pollen sources vary. In addition to these practical difficulties, different field experiment designs, especially different controls, address different questions. Most field studies on neonicotinoids have compared the performance of colonies adjacent to a bee-attractive crop grown from treated seeds with control colonies adjacent to the same crop grown from untreated seeds. This makes it possible to determine the cost, C, of any insecticide residues but not the overall effect of the crop, B – C, which includes any benefits, B, of the additional pollen and nectar. An alternative and rarely used design, only 1 of the 12 field studies reviewed, determines the overall effect of the crop by siting control colonies out of foraging range of the crop but in the same landscape. This design may be more relevant because bee-attractive crops, such as oilseed rape, are normally grown as alternatives to arable crops that do not provide bee forage, such as wheat and barley. As a result, it is possible that B – C?>?0 (or B – C?=?0) even if C?>?0. The challenges and merits of alternative experimental designs are discussed in relation to practical considerations and policy making.