Various long-term bioassay methods have been used to determine the carcinogenicity of chemical substances. Among them, a long-term method developed by Toth (1968), which scores the incidence of pulmonary adenoma formation at 28 to 56 weeks following subcutaneous administration of chemical carcinogens to new-born mice is used widely. This particular long-term bioassay method, while it takes long periods, has proved useful in determining the anticarcinogenic effect of ginseng extracts against the pulmonary adenoma formation inducible with dimethylbenzanthracene, urethane, and aflatoxin B1. More recently, in order to shorten the assay durations to 9 weeks (medium-term), we have modified the existing method by adjusting the doses of carcinogens to be administered to the new-born mice. We have established a modified method in which 40-50% of mice were found to develop pulmonary adenoma 9 weeks after a subcutaneous injection of 0.5 mg/kg to new-born mice of NIH(GP) strain and this modified medium-term bioassay system was found useful in the screening of cancer preventive agents among natural products, such as ginseng and caffeine.