Immunocyte and germ cell neoplasms, often curable by chemotherapy, arise from normal tissues most vulnerable to the effects of cytotoxic drugs; generalizing from these results to treating other tumors with such agents may not be entirely valid. Our limited success in treating epithelial neoplasms may be due to insensitivity rather than to drug resistance. Well-designed attempts to overcome resistance have been unsuccessful. The acquired immunodeficiency syndrome has not confirmed the putative role of immune surveillance in the pathogenesis of most neoplasms. The limited success of the most elaborate immunotherapies suggests that they, too, are nonspecific cell-killing techniques. Immunologic and cytotoxic drug therapies deserve further investigation but on a smaller scale. Neoplastic cell molecular biology, unknown when these therapies were developed, is being rapidly elucidated and may make it possible to treat malignancies by modulating cell physiology. Success of therapies based on the advances, in molecular biology, is not more uncertain than that of traditional treatments. Differentiation-induction techniques have already induced remissions in patients with acute promyelocytic leukemia and squamous cell carcinomas.