A double blind, randomized, controlled trial was carried out at two hospitals to try the effect of metronidazole, which is effective against most anaerobes, and cefazolin, which is effective against many aerobic organisms, singly and in combination, on the incidence of wound sepsis following appendicectomy. Patients were placed at random into one of four groups; to receive metronidazole and placebo, cefazolin and placebo, metronidazole and cefazolin, or double placebo. Patients with generalized peritonitis were excluded for ethical reasons. Treatment was started before operation and continued eight-hourly for twenty-four hours. All patients in the trial were followed up at about two weeks after discharge from hospital and their wounds inspected. Two hundred and seventy-one patients were assessed. Sepsis rates at the two hospitals were similar. On the untreated controls, 30% discharged pus from their wounds, and in the groups receiving one drug only, about 20%. In those receiving both drugs the infection rate was 3%, a highly significant difference from that in the other three groups.