A prospective randomized blinded study was conducted comparing a single 2-g preoperative dose of cefoxitin with three 2-g doses of cefoxitin over 12 hours given to premenopausal women scheduled for vaginal hysterectomy in Parkland Memorial Hospital. The incidence of major pelvic infection in 58 women given one dose was 1.7%, and it was 3.7% for 54 women given three doses. The mean hospital stay (4.5 days) was similar for both regimens; for those who developed major infection it was eight days. More aerobic bacteria with altered species dominance and fewer anaerobic bacteria were recovered from the vaginal cuff at discharge from the hospital when compared with those recovered from the endocervix preoperatively. There was a trend toward increasing minimal inhibitory concentrations to cefoxitin in bacteria isolated after surgery, which was more frequent in women given three doses. There was not a statistically significant intergroup difference in the recovery of bacteria resistant to cefoxitin in vitro after surgery. A single preoperative dose of cefoxitin was as effective in preventing major infection as were three perioperative doses, while providing other real and theoretic benefits.