The validity of melatonin as a prominent, naturally occurring oncostatic agent is examined in terms of its putative oncostatic mechanism of action, the correlation between melatonin levels and neoplastic activity, and the outcome of therapeutically administered melatonin in clinical trials. Melatonin's mechanism of action is summarized in a brief analysis of its actions at the cellular level, its antioxidative functions, and its indirect immunostimulatory effects. The difficulties of interpreting melatonin levels as a diagnostic or prognostic aid in cancer is illustrated by referral to breast cancer, the most frequently studied neoplasm in trials regarding melatonin. Trials in which melatonin was used therapeutically are reviewed, i.e., early studies using melatonin alone, trials of melatonin in combination with interleukin-2, and controlled studies comparing routine therapy to therapy in combination with melatonin. A table compiling the studies in which melatonin was used in the treatment of cancer in humans is presented according to the type of neoplasm. Melatonin's suitability in combination chemotherapy, where it augments the anticancer effect of other chemotherapeutic drugs while decreasing some of the toxic side effects, is described. Based on the evidence derived from melatonin's antiproliferative, antioxidative, and immunostimulatory mechanisms of action, from its abnormal levels in cancer patients and from clinical trials in which melatonin was administered, it is concluded that melatonin could indeed be considered a physiological anticancer substance. Further well-controlled trials should, however, be performed in order to find the link between its observed effects and the underlying mechanisms of action and to define its significance as a therapeutic oncostatic agent.