1. Nine patients in whom acute non-lymphoblastic leukemia (ANLL) developed following prolonged alkylating agent therapy are described. Five of the patients received no radiotherapy. The conditions treated were: Hodgkin's disease (four patients), primary amyloidosis, primary macroglobulinemia, malignant lymphoma, multiple myeloma, and carcinoma of the tonsil. 2. Prior to the advent of chemotherapy, this complication was not observed in large series of patients with lymphoproliferative disorders and multiple myeloma. However, the medical literature now contains at least 125 other detailed reports of ANLL developing after prolonged cytotoxic agent therapy. 3. Multiple myeloma and Hodgkin's disease, both of which commonly have good responses to chemotherapy, predominate as the underlying diseases. However, 35% of the case reports involve patients with other illnesses, including 12 patients who did not have neoplasms. 4. More than half of the patients developing ANLL have received chemotherapy alone without radiotherapy. 5. At least half of the patients developing ANLL experienced long periods of significant cytopenia during therapy, often with documentation of bone marrow dysplasia. 6. The wide variety of drugs associated with this complication suggests that any cytotoxic agent may be leukemogenic. However, alkylating agents overwhelmingly predominate as the class of compounds which are most often associated with terminal ANLL. 7. The vast majority of patients reported in the literature with ANLL complicating underlying malignancies have received cytotoxic drugs for prolonged periods (median 3 1/2 years) and leukemia developed most commonly 3 to 5 years after the diagnosis of the underlying disease. Most of these patients benefited from therapy and survived longer (median 5 years) than historical control of untreated patients. 8. The leukemogenic potential in man of prolonged cytotoxic agents therapy, especially with alkylating agents, seems to be well established. This evidence admonishes against the prolonged use of these drugs in non-fatal disorders. 9. More accurate assessment of risk: benefit ratios awaits the results of prospective controlled studies. The results of these studies could also lead to significant modifications in recommendations for long-term maintenance therapy with cytotoxic agents.