Although both the epidemiologic and experimental studies have led to the identification of chemical carcinogens, the limitations in epidemiologic approaches and the need for primary prevention of cancer require a greater reliance on experimental studies. Long-term carcinogenicity studies in experimental animals have been instrumental in identifying chemicals with carcinogenic activity, and, in some cases, the experimental evidence has preceded the epidemiologic evidence (for 4-aminobiphenyl, aflatoxin B1, diethylstilbestrol, melphalan, mustard gas, and vinyl chloride). A better understanding of the multistage process of carcinogenesis and the findings from various short-term tests available more recently may provide a more solid basis for extrapolating experimental findings to man.