One hundred and fifty-one consecutive patients underwent allogeneic bone marrow transplantation (B.M.T.) following high-dose chemotherapy and single dose total body irradiation (T.B.I.) for hematologic malignancies between September 1980 and December 1985. All patients included in this study were treated using a 60 Co beam to deliver a prescribed dose of 10 Gy to the mid-plane of the abdomen. Total body irradiation was performed the day before B.M.T. The mean instantaneous dose-rate was 3.5 cGy/min (range: 2.6 to 4.7). The real dose received was measured using thermoluminescent dosimeters (lithium borate). The difference between the doses delivered to the liver and to the mid-plane of the abdomen did not exceed 5%. The mean real dose delivered to the reference point was 10 Gy (range 8.3 to 11.7). Ninety five per cent of the patients received a dose ranging from 9.1 Gy to 10.9 Gy. High-dose cyclophosphamide was given to 126 patients with a "standard-risk" of relapse (60 mg/kg on day 5 and 4 before B.M.T.). Chemotherapy was intensified by the addition of other drugs in 25 patients with "higher-risk" of relapse. We analyzed the effect of the following pretransplant characteristics on the subsequent posttransplant development of V.O.D.: age, sex, ASAT and/or ALAT before conditioning regimen, diagnosis and status of malignant disease, history of liver disease, interval between diagnosis of hematologic malignancy and B.M.T., conditioning regimen (i.e., classical or intensified) and dose delivered to the liver during T.B.I. Seventeen patients were classified as having clinical V.O.D. giving a prevalence of 11.2%. In the first 2 months following B.M.T., death occurred respectively in 9 of 17 (53%) and 23 of 134 (17%) patients with and without clinical V.O.D. Univariate analysis showed that four characteristics were significantly related to an increased prevalence of V.O.D.: sex (11/62 females vs 6/89 males; p less than 0.05); history of liver disease (7/28 vs 10/117 patients without antecedent; p less than 0.01); ASAT and/or ALAT levels greater than 1.5 upper normal limit (11/49 vs 6/102 patients with levels less than 1.5; p less than 0.01) and intensified conditioning regimen (6/25 vs 11/126 patients with classical regimen; p less than 0.05). The conditioning regimen and history of liver disease were highly correlated to transaminases levels. Only two factors, transaminases levels and female sex, remained significantly associated with V.O.D. after multivariate analysis.