We have reviewed some of the data that link the reactive oxygen species produced by inflammatory phagocytes to cancer development. While it is clear that these substances induce phenotypic changes characteristic of those produced by known carcinogens, the precise mechanisms by which these effects are produced require much further study. In vitro, it would appear that phagocyte-generated oxidants could be complete carcinogens, ie, could cause both tumor initiation and promotion. In vivo, however, these substances appear usually to function as tumor promoters or cocarcinogens perhaps because of high levels of endogenous antioxidant defenses. This suggests that there may be even more reason to be optimistic about the potential for positive results in cancer chemoprevention trials in humans, and provides further rationale for the continuing interest in the use of antioxidants and anti- inflammatory drugs in current and future trials. For example, the Chemoprevention Branch of the National Cancer Institute is currently sponsoring seven extramural human efficacy intervention trials testing whether the antioxidant beta carotene can prevent cancer.