So-called antioxidants have yet to make a clinical impact on the treatment of human cancer. The reasons for this failure are several. First, many agents that are called antioxidants are truly antioxidants at a given dose, but this dose may not have been given in clinical trials. Second, many agents are not antioxidants at all. Third, not all tumors use reactive oxygen as a signaling mechanism. Finally, reactive oxygen inhibition is often insufficient to kill or regress a tumor cell by itself, but requires sequential introduction of a therapeutic agent for maximal effect. We hope to provide a framework for the logical use of these agents in cancer.