From 1979-1983, 299 patients with stage III or IV Hodgkin's disease (HD) were randomised to receive cyclical chemotherapy with MOPP (mustine, Oncovin, procarbazine, prednisone) or LOPP (Leukeran substituted for mustine). Two hundred and ninety patients were evaluable. There was no statistically significant difference between the complete remission (CR) rates (63% for MOPP, 57% for LOPP), percentage of patients remaining disease free at 5 years (38% for MOPP, 35% for LOPP) and overall survival at 5 years (65% for MOPP, 64% for LOPP). On multivariate analysis younger age, grade I histopathology, absence of systemic symptoms, and normal albumin level were favourable prognostic factors for survival. Acute toxicity in the form of nausea/vomiting, myelosuppression, and phlebitis were less with LOPP than MOPP. Deaths in both groups were usually due to disseminated Hodgkin's disease; there were no infective deaths in the absence of Hodgkin's disease. Second malignancies occurred in six patients treated with MOPP--three acute myeloid leukaemia (AML), one non-Hodgkin's lymphoma (NHL), two carcinomas (Ca); with LOPP, four second malignancies occurred (one AML, one NHL, two Ca). These long term results confirm that LOPP is as effective as MOPP, and less toxic, in the treatment of advanced Hodgkin's disease.