The risk of postsurgical infection is closely related to the type of operation performed. Wound sampling by Raahave's velvet pad method shows that operation wounds harbour ordinary aerobic and anaerobic commensals after incision (exogenous flora). When bacteria-containing viscera are opened or found perforated, bacterial densities increase significantly before wound closure (endogenous flora). Contamination level justifies the criteria for use of wound class, i.e. clean, potentially contminated, contaminated and dirty. There is a dose-response relationship between the density of bacteria during operation and postsurgical wound sepsis. The median infective dose of aerobic and anaerobic bacteria is 4.6 x 10(5) CFU x cm-2 wound. The surgeon opens up tissue planes which are primarily sterile and breaks natural barriers. However, using a trained technique, the surgeon should be able to control wound contamination safely throughout the operation. Although hos immunity and tissue resistance intervene between contamination and infection, the magnitude of contamination seems crucial. It ought to be axiomatic by now that reducing bacterial contamination reduces infection.