Polytetrafluoroethylene (PTFE) (Gore-Tex) and human umbilical vein (Biograft) arterial grafts were compared for below-knee femoropopliteal bypass grafting in a prospective randomized clinical trial. One hundred five patients (105 limbs) entered the trial. Seventy-six percent suffered from rest pain, ulceration, or gangrene. The median postoperative ankle-arm blood pressure index was 0.36. Twenty-three limbs had three patent tibial arteries, 46 limbs had two tibial arteries, 31 limbs had one patent artery, and five limbs had isolated popliteal segments. Thirty-four percent were repeat operations. Fifty-five patients were allocated to receive PTFE grafts and 50 to receive human umbilical vein grafts. The two groups were comparable as to preoperative risk factors and operative and postoperative treatment. During the first 4 years (maximum 1609 days) 40 PTFE grafts and 24 umbilical veins occluded. At 1 year the PTFE patency rate was 53% and at 4 years was 22%. For umbilical vein the corresponding figures were 74% and 42% (p = 0.005, Gehan test). During follow-up the incidence of PTFE failure was on the average 2.1 times higher than that of umbilical vein failure (95% confidence limits 1.2 to 3.4).