We assessed the effects of 41 potential chemopreventive agents in the F344 rat using the inhibition of carcinogen-induced aberrant crypt foci (ACF) in the colon as the measure of efficacy. ACF were induced by the carcinogen azoxymethane in F344 rats by two sequential weekly injections at a dose of 15 mg/kg. Two weeks after the last azoxymethane injection, animals were evaluated for the number of aberrant crypts detected in methylene blue-stained whole mounts of rat colon. The 41 agents were derived from a priority listing that was based on reports of chemopreventive activity in the literature and/or efficacy data from in vitro models of carcinogenesis. The list of agents included representative examples of phytochemicals, vitamins, minerals, inhibitors of proliferation, inducers of Phase 1 and Phase 2 metabolism systems, nonsteroidal anti-inflammatory agents, and differentiation agents. Eighteen agents were positive in the assay, significantly reducing the incidence of ACF at least in one of two doses tested. As a chemical class, the nonsteroidal anti-inflammatory drugs, which included ibuprofen, ketoprofen, piroxicam, and indomethacin, were most active; other less potent agents were arginine, butylated hydroxyanisole, curcumin, diallyl sulfide, difluoromethylornithine, 18 beta-glycyrrhetinic acid, indole-3-carbinol, oltipraz, purpurin, rutin, and the sodium salts of butyrate, selenite, and thiosulfate. Twenty-three agents did not inhibit ACF; included among these were several agents that promoted the development of ACF at one or both doses tested: benzyl isothiocyanate,calcium glucarate, catechin, dihydroepiandosterone, fluocinolone acetonide,folic acid, levamisole, 2-mercaptoethanesulfonic acid, nordihydroguiaretic acid, potassium glucarate, propyl gallate, beta-sitosterol, sodium cromolyn, sodium molybdate, and sulfasalazine. The aberrant crypt assay demonstrates reasonable specificity and sensitivity in predicting which agents are likely to prevent colon cancer.