The route of administration of prophylactic antibiotics was studied in a randomized prospective trial. Cefoxitin was administered to high-risk patients at cesarean section by three treatment regimens: intravenous antibiotic (2 g) for eight doses, irrigation of uterus and peritoneum with 2 g of antibiotic, and a combination of intravenous and irrigation as described. A control group received no antibiotic prophylaxis. The incidence of febrile morbidity was similar in each treatment group: intravenous, two of 39 (5%); irrigation, three of 42 (7%); intravenous and irrigation, two of 38 (5%), and were all significantly lower than the control group 14 of 39 (36%) (P less than .05). Similar results were found when prevention of endometritis was the end point: intravenous, two of 39 (5%); irrigation, two of 42 (5%); intravenous and irrigation, two of 38 (5%) compared with 13 of 39 (33%) in the control group (P less than .05). Administration of antibiotics by irrigation is equally effective in preventing postoperative febrile morbidity and endomyometritis as intravenous dosing and a combination of intravenous and irrigation. This affords a potential cost savings.