During the period May 1986 to July 1989, a prospective, double blind, randomized trial of antibiotic prophylaxis in colorectal surgery was undertaken at the Royal Brisbane Hospital. Three hundred and thirty patients were considered eligible for the trial. Three regimens were compared: a combination of 2 g ceftriaxone and 1 g metronidazole; a single dose of 2 g ceftriaxone; or 1 g cefazolin and 1 g metronidazole, as antibacterial prophylaxis in colorectal surgery. Fifty patients were excluded from analysis. The overall incidence of wound sepsis was 7.9% (22 patients). There was no statistical difference in the incidence of wound infections between the three groups. The presence of drains and the non-performance of a bowel anastomosis at the time of surgery predisposed patients to wound infection. Staphylococcus aureus or Sfaphylococcus epidermidis were the cause of wound infection in 16 cases. Patients in the cefazolin and metronidazole group had a significantly higher number of postoperative urinary tract and respiratory tract infections than the other two groups combined (P < 0.01). There did not appear to be any change in sensitivity patterns to ceftriaxone during the 3 year trial. During the 3 year period of the study, ceftriaxone was found to be a safe and effective drug in antibacterial prophylaxis in colorectal surgery.