In a prospective, randomized trial, 226 patients undergoing femoropopliteal bypass for lower limb ischaemia were allocated to reversed (123 patients) or in situ (103) techniques. The groups were comparable for age, sex, incidence of diabetes, and indications for surgery. Eleven veins were rejected at operation, nine in the reversed group and two in the in situ group, leaving 114 reversed and 101 in situ grafts for study. Cumulative patency rates were not significantly different between reversed and in situ grafts at any time up to 6 years after operation, with primary patency rates at 1, 3 and 5 years of 84.8, 69.5 and 62.4 per cent for reversed grafts and 79.9, 71.2 and 63.5 per cent for in situ grafts. Small vein grafts (< 4 mm in diameter) were associated with patency rates at 1, 3 and 5 years of 63.5, 46.7 and 36.0 per cent compared with 93.9, 82.5 and 75.9 per cent for vein grafts > or = 4 mm in diameter (P < 0.002, log rank test). The patency rates of small veins employed in situ and reversed were similar. The in situ technique confers neither short- nor long-term advantage over reversed vein grafting for femoropopliteal bypass.