Currently it is standard practice in many acute care hospitals to use only sterile water in humidifier reservoirs for the delivery of low-flow oxygen therapy. This practice is based primarily on the Centers for Disease Control's 1982 "Guideline for the Prevention of Nosocomial Pneumonia." The present study was designed to investigate and compare the bacterial contamination of tap water and sterile water used to fill clean (nonsterile) disposable oxygen humidifier reservoirs. Disposable oxygen humidification reservoirs were assembled weekly according to standard protocol and regulated to deliver oxygen, 4 to 6 L/min, continuously for 5 consecutive days. Each of 48 reservoirs was filled with either tap water (24) or sterile water (24) and cultured daily. All disposable equipment was discarded after the final culture on the fifth day. An aliquot of water was obtained daily from each reservoir and plated to trypticase soy agar pour plates and to a matched trypticase soy broth. Bacterial growth was observed on trypticase soy agar pour plates from 54 (45.0%) of 120 sterile water reservoir cultures and from 38 (31.7%) of 120 tap water reservoir cultures (p greater than 0.05). Bacterial growth from the matched trypticase soy broth cultures was observed only from two sterile water reservoirs. The microorganisms identified from the sterile water reservoirs included Enterobacter agglomerans and Serratia and Bacillus spp. The use of tap water for low-flow oxygen humidification was determined to be safe at our hospital by the infection control committee. This procedural change contributed approximately $6000 to the cost reduction efforts of the respiratory therapy department.