Of 152 patients who were scheduled for an amputation for ischemia, seventy-seven were randomly assigned to perioperative prophylaxis with cefoxitin (Mefoxin) and seventy-five patients, to injections of a placebo. The patients were followed for twenty-one days or, in the case of wound complications, to the end of treatment. An infected wound occurred in 38.7 per cent of the patients in the placebo group and 16.9 per cent of those in the antibiotic group (p less than 0.005). Clostridial infection occurred in eight patients in the placebo group and in none in the antibiotic group (p = 0.003). Three of the patients with clostridial infection died of gas gangrene. A multivariate analysis showed that the absence of antibiotic prophylaxis increased the risk of infection by a factor of 3.3 (p = 0.004) and increased the need for reamputation by a factor of 4.5 (p = 0.003). We concluded that amputation patients should have prophylaxis with a broad-spectrum antibiotic given perioperatively.