A prospective randomized trial is described in 119 patients undergoing major abdominal general surgical procedures. Half of the patients were treated prophylactically with intermittent pneumatic calf compression, begun after the induction of anesthesia and continued until the patient was walking; the other half acted as controls. Deep venous thrombosis was detected by iodine-125 fibrinogen scanning and confirmed by venography, and did not differ significantly in the control and treated groups. One fatal pulmonary embolism occurred in each group. The presence of malignancy of the gastrointestinal tract did not influence the results. The findings suggest that pneumatic compression delayed the development of deep venous thrombosis postoperatively and that perhaps it should be continued until discharge from hospital.