We determined the intraoperative serum and wound-muscle concentrations of cefazolin and cefoxitin in 40 patients who were undergoing cholecystectomies. The study employed an open-label design in which all of the patients randomly received cefazolin sodium (20 mg/kg) or cefoxitin sodium (30 mg/kg) intravenously while the patient was in the ward ("on call") or with the induction of anesthesia. Multiple blood and wound-muscle samples were collected intraoperatively and assayed for their cephalosporin concentrations. Considerable differences in intraoperative serum and tissue concentrations between antibiotics were apparent; there were usually higher levels of cefazolin. In all of the patients who received cefazolin sodium, the antimicrobial was detectable in wound tissue at wound closure, while it was detectable in 86% and 38% of patients who received cefoxitin sodium with anesthesia and on call, respectively. Because cefoxitin has a much shorter elimination half-life than cefazolin it seems prudent to administer the agent as close to the start of the operation as possible, and readminister the agent every two to three hours until the wound is closed. For cefazolin, on-call administration appears to be acceptable, with readministration not required for at least four hours.