This study evaluated the efficacy of an upper-room air ultraviolet germicidal irradiation (UVGI) system for inactivating airborne bacteria, which irradiates the upper part of a room while minimizing radiation exposure to persons in the lower part of the room. A full-scale test room (87 m3), fitted with a UVGI system consisting of 9 louvered wall and ceiling fixtures (504 W all lamps operating) was operated at 24 and 34 degrees C, between 25 and 90% relative humidity, and at three ventilation rates. Mycobacterium parafortuitum cells were aerosolized into the room such that their numbers and physiologic state were comparable both with and without the UVGI system operating. Airborne bacteria were collected in duplicate using liquid impingers and quantified with direct epifluorescent microscopy and standard culturing assay. Performance of the UVGI system degraded significantly when the relative humidity was increased from 50% to 75-90% RH, the horizontal UV fluence rate distribution was skewed to one side compared to being evenly dispersed, and the room air temperature was stratified from hot at the ceiling to cold at the floor. The inactivation rate increased linearly with effective UV fluence rate up to 5 microW cm(-2); an increase in the fluence rate above this level did not yield a proportional increase in inactivation rate.