Procedures used to select pesticides to be included in a NIOSH case control study examining the risk of brain cancer in farm workers in Iowa, Michigan, Minnesota, and Wisconsin associated with pesticide use and other environmental exposures was discussed. A procedure to identify and prioritize the pesticides most appropriate for the study was needed because of the large number and variety of pesticides available, approximately 24,000 commercially registered pesticides, therefore making it impossible to investigate them all. Lists of pesticides were created to document those used in the four states and in the US as a whole utilizing data from the US Department of Agriculture and the departments of agriculture and land grant colleges in each state. Pesticides were then prioritized according to the criteria: volume of use before 1985, ranking of use in the four states and the US as a whole according to pesticide type, and toxicologic evidence of carcinogenic, teratogenic, and mutagenic effects. The criterion 'pesticide use before 1985' was chosen to allow a latency period of at least 10 years for brain cancer. The study was to be based on brain cancer cases diagnosed in the four states from 1985 on. A total of 240 pesticides were used in the four states: 102 herbicides, 81 insecticides, 32 fumigants, and 35 fungicides. Using the criteria above, the list was reduced to 56 herbicides, 49 insecticides, 17 fumigants, and 12 fungicides. The final list represented more than 99% of the total pounds of herbicides and insecticides and more than 98% of the total pounds of fungicides and fumigants used in the states before 1985. Lists of the priority pesticides are to be sent to the study participants before the study questionnaire is administered to allow them time to recall details of their including specific years of use and crops they were used on. The authors conclude that the procedures used to select the pesticides to be examined in this study are useful for documenting past pesticide use.