Humification of oxygen prior to administration by nasal cannula is an expensive practice which has been justified on the basis that it improves the comfort of patients receiving supplemental oxygen therapy. Routine humidification of low-flow oxygen (less than or equal to 4 L/min) delivered by nasal cannula has recently been challenged based on theoretic grounds and on the results of a clinical study. Nevertheless, we found, in a telephone survey of medium-sized American hospitals, that routine humidification of nasal cannula oxygen remains a common practice. To further evaluate the necessity of oxygen humidification, we prospectively evaluated, on a daily basis, the subjective complaints of consecutive patients in our institution who were ordered to receive nasal oxygen at relatively high flow rates (5 L/min). Of the 185 patients evaluated over a period of three wintertime months, 99 received humidified oxygen and 86 received dry oxygen. Complaints, especially dry nose and dry throat (42.9 percent and 43.9 percent of the daily interviews, respectively) were common in both groups, but the symptoms were relatively mild and did not increase significantly when oxygen was administered without prior humidification. We conclude that routine humidification of oxygen for administration by nasal cannula is not justifiable, and that cessation of this practice would result in significant reductions in both time and material costs in respiratory care.