Routine carcinogenicity testings have been designed and conducted primarily for providing data to show whether or not the test compounds have a potential of inducing tumors in animals. Therefore, in the case of safety assessment or risk assessment of test compounds in humans, additional data are needed, at times, to learn how intensely the test compound can induce tumors in animals or by which mechanism the test compound can induce tumors in animals. The initiation/promotion experiment is performed as a proceeding for such requests. This paper describes the updating principle and procedures to evaluate initiation effects and promotion effects separately. However, it must be realized that our present knowledge about the initiation/promotion is still limited to some qualitative evidences. Actually, we know very little about mechanistic background and quantitative aspect of the initiation/promotion such as the site or mode of action of promoter action, the organ-specificity of promoter action, or threshold of the initiator action and promoter action. All these problems are necessary to be studied systematically in order that the initiation/promotion design can make a more important contribution to the evaluation of carcinogenicity of chemicals.