The records of sixty-nine consecutive elderly diabetic patients over age 70 years who underwent femoropopliteal bypass at the New England Deaconess Hospital from 1966 to 1975 were reviewed and analyzed statistically. Grafts were classified as patent only if the patient had a palpable distal pulse postoperatively. No hospital deaths occurred. The patency rate declined with decreasing runoff below the knee. Patients receiving insulin had a patency rate of 7 per cent, compared with a rate of 35 per cent in patients treated with diet or oral agents (p < 0.05). Cross analysis disclosed no other significant difference between these two groups. No benefit from Dextran was demonstrable in our series. Graft closure carried a 36 per cent risk of above knee amputation. Life table analysis of patients who underwent autogenous vein graft (fifty-seven patients) showed one and three year patency rates of 72 and 56 per cent, respectively. Conversely, grafts of Dacron had one and three year patency rates of 48 per cent and less than 10 per cent, respectively. In a favorable subset of twenty-six patients who (1) were receiving insulin, (2) had either previously successful bypass or no previous vascular surgery, (3) had arteriographic evidence of one or more vessel runoff, and (4) had autogenous vein grafts, the one and three year patency rates were 91 and 82 per cent, respectively. Our experience shows that femoropopliteal bypass is safe and effective when patients are carefully selected. We strongly discourage the use of Dacron materials in elderly diabetic patients. © 1979.