Two hundred and seventeen patients, undergoing abdominal colonic and rectal surgery, received after randomization, the following regimen: group A (74 patients): cefotaxime 1 g intravenous at the induction of anaesthesia, the beginning of the resection, 4 and 8 h later; group B (72 patients): cefotaxime in the same regimen associated with ornidazole or metronidazole 0.5 g intravenous at the induction of anaesthesia and 0.5 g intravenous with the last injection of cefotaxime; group C (71 patients):cefotaxime following the same regimen as groups A and B and metronidazole orally 0.5 tds 3 days before surgery. All wounds were assessed daily, until discharge from hospital. Severe sepsis included: septicaemia, peritonitis, intra-abdominal abscess and extra-abdominal infections with death. Non-severe sepsis included all others. All the patients having a history of allergy to beta-lactam antibiotics and those with pre-operative infection were excluded. Mean age of the population was: 64.5 years. Seventy-seven patients had rectal cancer and 82 patients cancer of the colon; Twenty-five patients had inflammatory bowel disease, and in 33 others disease such as polyposis was present. Risk factors of post-operative infection were present in 115 cases (A, 36 patients; B, 37 patients; C, 42 patients). All three groups were very well matched for age, sex, type of intervention and diagnosis. Non-infectious complications appeared in 56 patients. Sepsis developed in 76 patients (A, 27 patients; B, 27 patients; C, 22 patients, no significant difference). Severe sepsis occurred in 14 patients (A, 6 patients; B, 4 patients; C, 4 patients, no significant difference) and in 62 patients non severe sepsis (A, 21 patients; B, 23 patients; C, 18 patients, no significant difference). Post-operative peritonitis was not seen. This study suggests that cefotaxime alone 4 g peri-operatively is useful in prophylaxis during rectal and colonic surgery.(ABSTRACT TRUNCATED AT 250 WORDS)