One hundred nine primary bypass operations were performed on 97 patients who had limb ischemia caused by arteriosclerosis obliterans, over a 32-month period starting in January, 1978. The majority of those procedures were done for rest pain or gangrenous changes. The greater saphenous vein was used whenever possible (44 of 109 procedures), and a polytetrafluoroethylene (Gore-Tex) prosthesis was used in 65 of 109 procedures. Average follow-up was 9.3 months. The cumulative patency rate in all vein grafts was 93.1% at 1 month and 73.4% after 31 months, whereas the cumulative patency rate for polytetrafluoroethylene was 83.3% at 1 month and 51% at 31 months. The difference is not statistically significant (0.10 less than P less than 0.25). In above-knee procedures for claudication, polytetrafluoroethylene performed identically to vein (P greater than 0.50), although vein appeared to perform better in above-knee procedures done for rest pain or necrosis. A limited number of below-knee revascularizations were performed, and although no conclusions of statistical significance can be drawn, cumulative patency of vein appeared to be better than polytetrafluoroethylene. For additional comparative analysis, this study was compared to a previous study in which Dacron was used in femoropopliteal arterial reconstructions. Patency rates in above-knee revascularizations appeared to be comparable in Dacron and polytetrafluoroethylene. Diabetes did not appear to alter age of onset of disease or patency rates, although smokers became symptomatic 10 years earlier than nonsmokers. Although the early and long-term results of arterial reconstructions in the ischemic extremity are known, the availability of new alternate arterial prostheses raises the question of the optimum material depending on the specific clinical circumstances.