Patients who have received radiation to localized areas of marrow eventually regenerate marrow in the irradiated area, if the dose is 2,400 centigrays (cGy) or less. This trial was designed to deliver a radiation dose of 1500 cGy to all marrow containing sites in patients with multiple myeloma, a technique we refer to as total bone marrow irradiation, or TBMI. Patients with previously untreated myeloma received 12 weeks of melphalan (L-PAM) and prednisone (pred) therapy. Four weeks later, sequential irradiation was administered using the 3-2 technique with rest periods to permit recovery from radiation-induced cytopenia. This was followed by electron beam irradiation of the rib and skull fields. Following completion of TBMI, patients were untreated until relapse. Twenty patients were entered. At entry 5, 8, and 7 patients had low, intermediate and high tumor cell loads, respectively. Two patients had a serum Ca in excess of 12 mg/dl; 3 had an increased creatinine. The median performance (ECOG) was 1. At week 16, immediately prior to TBMI, 5 of the 20 patients fulfilled the Myeloma Task Force criteria for response and 5 others had improved. Six patients did not begin the radiation therapy portion of the protocol. Three had rapidly progressive disease, one persistent leukopenia, one refused radiation therapy and one was withdrawn by his physician. Only 6 of the fourteen patients receiving the radiation treatment phase of the protocol were able to tolerate the intended course of 1500 cGy to all areas. Eight other patients received lower doses. Patients completing the radiation phase of the protocol failed to have further reductions in M-protein or improvement in other parameters beyond those obtained on the chemotherapy phase of the protocol. The median duration of response and survival was 12.0 and 42 months, respectively. We suggest possible reasons for the disappointing results of this trial and conclude that this approach to the primary treatment of myeloma holds little promise.