There is now a compelling body of evidence supporting the important role of tumor suppressor genes in carcinogenesis. Evidence for the existence of these genes emerged gradually from somatic cell genetic and epidemiologic studies, as well as from studies of chromosome losses in tumor cells by cytogenetic and basic molecular genetic techniques. Within the last decade more than a dozen tumor suppressor genes, both the well-documented and candidate variety, have been identified. When these genes are inactivated in the germline of some individuals, there may be a strong predisposition to cancer. More often, tumor suppressor genes are inactivated by somatic mutations arising during tumor development. For the future, a more complete, description and understanding of carcinogenesis will emerge with the identification of additional tumor suppressor genes, the detailed characterization of their normal cellular function, and the elucidation of the germline and somatic mutations that lead to the inactivation of these genes in human tumors. Although we are not there yet, a better understanding of tumor suppressor genes, coupled with our ability to manipulate the genome with gene therapy, has the potential to open broad vistas in the treatment of a variety of human malignancies.