A total of 120 patients who were to be delivered by cesarean section and who were at high risk of postoperative infection received three doses of either cefamandole, cephalothin or placebo perioperatively. Maternal serum levels for both antibiotics were in the therapeutic range. Although both drugs reduced the incidence of febrile morbidity and endometritis, only cefamandole significantly reduced the fever index. Risk factors for postoperative infections were the presence of ruptured membranes, labor, and internal fetal monitoring. Cefamandole beneficially influenced all risk factors while cephalothin was able to reduce only the risk of ruptured membranes. When a new method for obtaining endometrial tissue was utilized, 50% of cultures were negative. There was no difference in the organisms isolated from patients with and without endometritis.