Current knowledge of the relationship between human consumption of pesticide residues and health risks is not adequate to reliably characterize the risks. Only a minority of the compounds used have been subjected to state-of-the-art toxicological testing, and even after a compound is tested in bacteria and laboratory mammals, substantial uncertainty remains about its effects on humans. In addition, the identities and quantities of pesticides remaining on produce at time of consumption are not well characterized. As in many areas of environmental and health policy making, substantial outcome uncertainty complicates the analysis.