Microbiologic surveillance of both the patients and the ambient environment of a surgical intensive care unit allowed us to relate the incidence of nosocomial respiratory tract infection to levels of airborne bacteria. Over the study period respiratory tract nosocomial infection rates varied from 0.7% to 17.0%, and nonrespiratory infection rates varied from 1.0% to 25.0%. Airborne bacteria counts during that time varied from 1.0 ± 0.8 S.E. CFU/ft3 to 96.0 ± 6.8 S.E. CFU/ft3. There was a reasonably close correlation between airborne bacteria levels and the incidence of nosocomial pneumonia (r=0.81, p<0.05). Furthermore, there was a close correlation between a specific organism nosocomial infection rate and the number of colonies of that organism present in the air (r=0.88, p<0.05). In contrast, there was no significant relationship between airborne bacteria counts and nonrespiratory attack rates (r=0.60, p<0.05). If the bacteria traveled from the air to the patients, there appear to be at least three possible explanations for this significant relationship: (1) direct inoculation of the airway by the airborne bacteria; (2) inoculation of the airway by direct contact, which is related to the degree of “cleanliness” of the environment; and (3) an increased incidence of contaminated respiratory equipment and airway inoculation because of high counts of airborne bacteria. It is also possible that high bacterial air counts represent contamination of the air from patients with respiratory infections. Regardless of the pathway(s), surveillance of the ambient environment may prove to be a useful epidemiologic tool in the study and control of nosocomial respiratory tract infections in certain high-risk patient care areas.