Two hundred patients treated with curative intent for Hodgkin's disease between October 1964 and April 1984 at a single institution were studied retrospectively for development of second malignancies. The minimum follow-up was 2 years (median, 11 years). The staging distribution was IA-B, 61; IIA, 54; IIB, 20; IIIA, 46; and IIIB, 19. Sixty-one percent of the patients had laparotomy. Initial management was irradiation alone (RA) in 143 patients and a combination of chemotherapy and irradiation (CB) in 57 patients. Actuarial 10-year survival rates were 82%, IA-B; 78%, IIA; 66%, IIB; 66%, IIIA; and 24%, IIIB. Cause-specific deaths due to Hodgkin's disease or complications of initial or salvage therapy occurred in 3% of IA-B patients, 18% of IIA-B patients, and 35% of IIIA-B patients. One patient had a prior T3N1 squamous cell carcinoma of the retromolar trigone, and a second was diagnosed with concurrent Hodgkin's disease and granulocytic sarcoma. Subsequent solid tumors have occurred in six patients from 5 to 21 years after treatment, including papillary carcinoma of the thyroid, renal cell carcinoma, unilateral breast carcinoma, cervix carcinoma in situ, and lung carcinoma after RA, and bilateral breast carcinoma after CB. Seven fatal hematopoietic disorders (HPDs) were observed, including four acute leukemias, one dysmyeloproliferative syndrome (DMPS), one autoimmune hemolytic anemia, and one aplastic anemia. Two occurred in patients initially managed with RA who subsequently required chemotherapy for relapse. Five HPDs occurred in patients initially managed with CB who never relapsed. All HPDs were observed between 2 and 7.5 years after administration of chemotherapy. Statistical analysis of the data using a rerandomization test on Gehan ranks of treatment and clinical variables showed significant correlations between development of a secondary HPD and (1) initial management with CB; (2) higher doses of chemotherapy; and (3) more advanced disease, particularly IIIB. When only the five events generally associated with treatment (i.e. the four leukemias and one DMPS) were considered, there was a significant correlation with exposure to chemotherapy and presentation with advanced disease. The patient population was small so that interdependence between treatment factors and initial extent of disease in affecting the risk of a secondary HPD cannot be discounted but should be further investigated with larger patient populations.