The impact of three current types of preoperative large bowel preparation on the microbial flora and the colonization resistance (CR) was investigated in 15 volunteers. In the first group a whole gut irrigation was performed without administration of antibiotics (group WGI). In the second group 0·5 g/1 metronidazole and 1 g/1 neomycin was added to the irrigation fluid (group WGI + AB). A whole gut irrigation with prior oral administration of 1 l mannitol 10% was performed in the third group. The antibiotic prophylaxis in this group consisted of two doses of 80 mg gentamiein i.v. and 500 mg metronidazole orally 24 h after lavage (group Mann + AB). One hour after the mechanical cleansing procedure was finished all volunteers were orally contaminated with one dose of an Escherichia coli test strain. The aerobic faccal reduction due to the cleansing procedure was 2–3 logs, while for the anaerobes it was 4–5 logs. The anaerobic flora in group WGI recovered within 24 h, while the aerobes showed a transient ‘overgrowth‘ for the period of 2 days. The overgrowth of aerobes in group WGI + AB was observed for more than a week and the total numbers of aerobes started gradually to decline after the anaerobic flora had reached pretreatment levels at day three or four. Despite the normal numbers of anaerobes present 24 h after treatment, overgrowth of E. coli was seen in the group Mann + AB, probably due to residual mannitol left in the intestinal tract. The test strain of E. coli was excreted for a period of 1 week by the volunteers in the groups WGI and Mann + AB, but it was isolated for more than 10 weeks in the group WGI + AB. It is thought that all three methods of preoperative large bowel preparation decreased the CR of the gastrointestinal tract because of a disturbance of the interaction between aerobic and anaerobic microorganisms and alterations of the colonic wall. The anaerobic microflora. however, appeared to be primarily responsible for the maintenance of the CR. Antimicrobial prophylaxis should consist of a high dose, short term. systemic antibiotic regimen, not only because an adequate serum level of an appropriate drug at the time of operation substantially decreases the incidence of postoperative septic complications but also because a systemic regimen scarcely influences the CR of the gastrointestinal tract. β-Aspartylglycine appeared to be a specific but not very sensitive marker for decreased CR.