A total of 361 evaluable patients with previously untreated multiple myeloma were randomized to receive oral melphalan (0.15 mg/kg/day for seven days, followed by 0.05 mg/kg/day after recovery from the nadir of the leukocytes), BCNU (150 mg/m2 intravenously every six weeks) or CCNU (100 mg/m2 orally every six weeks). All patients received a tapering six-weeks) or CCNU (100 mg/m2 orally every six weeks). All patients received a tapering six-week course of prednisone starting at 0.8 mg/kg for the first two weeks. At week 22, one-half of the patients were randomized to receive vincristine (1 mg/m2) and prednisone (0.6 mg/kg for seven days) every two months in addition to previous therapy. The melphalan treated patients showed a significantly higher overall objective response frequency (59%), according to Myeloma Task Force criteria, when compared to those treated with BCNU (40%) or CCNU (42%). The survivals for all patients were not statistically different for the three treatment programs. However, the good-risk patients treated with melphalan had significantly longer survival (P = 0.02) than the equivalent patients who received BCNU or CCNU. The addition of vincristine and prednisone at week 2 did not significantly increase the percentage of subsequent objective responses or prolong the subsequent survival of any treatment group. It is concluded that oral melphalan is superior to BCNU and CCNU in producing objective responses and in prolonging survival in good risk patients.