One hundred and two patients undergoing elective colorectal surgery were randomized in a prospective but "non-blind" way into two groups. The first group received 0.4 g of doxycycline preoperatively and the second group 0.2 g of doxycycline preoperatively and 0.1 g of doxycycline on the first, second and third postoperative day. All doxycycline was administered by intravenous infusion. The infection rate was 10% in the "one-dose" treatment group and 10% in the "four-dose" treatment group. It is thus concluded that 0.4 g of doxycycline preoperatively results in an acceptable rate of postoperative infections, which may be explained by the satisfactory concentration of doxycycline up till three days after administration. The advantage of "one-dose" treatment compared to "four-dose" treatment from practical and economical standpoints are obvious.