Pulse oximetry was used to determine the incidence of intraoperative hypoxemia in 108 patients undergoing ambulatory gynecologic operation. Eleven (10%) experienced moderate desaturation (arterial oxygen saturation less than 90%), and 5 (5%) suffered severe hypoxemic episodes (arterial oxygen saturation less than 85%). Among patient risk factors--including operation, body habitus, smoking habits, history of asthma, age, and airway characteristics--an association with moderate hypoxemia was found only with nonlaparoscopic gynecologic operation, obesity, and age over 35 years, and an association with severe hypoxemia was found only with obesity and age over 35. Among operative events--including inspired oxygen concentration, position, mode of ventilation, and anesthesia phase--an association with moderate hypoxemia was found only with the lithotomy position, manual ventilation, and arousal. The cost per patient of monitoring with a pulse oximeter is about +1.35. A cost-benefit analysis reveals that a mortality rate of 1 in 40,000 among patients who actually become moderately hypoxemic would justify the cost of monitoring arterial oxygen saturation. We conclude that pulse oximetry should be part of routine anesthetic monitoring.