Twenty-six evaluable patients were entered into two fractionated total body irradiation (TBI) programs; 11 patients received a course of 150 rad TBI (× 3 if tolerated) and 15 patients received a lower dose course of 50 rad (× 3 if tolerated). Complete remissions (CR) were not produced by either course; however, the higher dose course (Plan I) yielded a partial response (PR) rate of 73 %, while the lower dose course yielded a PR of 47 %. Although fraction size seemed trivial in both TBI plans, an unexpected high degree of hematologic toxicity was encountered, and was parallel to the response rates: in Plan I 173 % of patients experienced severe to life-threatening depression of platelets or granulocytes, whereas in Plan II this rate was 47 %. This was of short duration with rapid return of blood counts to normal levels. One death can be attributed to TBI. The chemotherapy arm of the study demonstrated superiority in terms of complete responses. Twenty-three percent of patients treated by cholrambucil and prednisone attained CR, in contrast to 0% of TBI patients. PR for chemotherapy was similar to that obtained with TBI. Chemotherapy also proved superior in terms of overall response rate, number of patients in remission, and in the median duration of response, but not in the median duration of survival. Fractional TBI techniques for active chronic lymphocytic leukemia (CLL) should be interrupted when the platelet count dips below 100,000 and the granulocyte count is lower than 2,000. Future studies should combine TBI radiation therapy and chemotherapy.