In human volunteers, the response times of 11 pulse oximeters to a 10% step reduction in arterial oxygen saturation were measured using an acute decompression technique. When finger probes were used, nine oximeters had similar response times and two were significantly slower (P less than 0.05). The ear probe response time was similar on six oximeters assessed, and faster than the finger probes. The response times of the oximeters to an acute increase in arterial saturation were tested by suddenly changing the inspired gas from air to 100% oxygen at an ambient pressure of 380 mm Hg. For ear probes, the response times were similar for all oximeters; for finger probes, three fast-responding and three slow-responding oximeters were identified (P less than 0.05). A faster response could be elicited by placing the probes on the thumb (P less than 0.05). We conclude that if a rapid indication of changes in arterial saturation is required, pulse oximeters with ear probes should be used. If finger probes are used, they should be placed on the thumb. The oximeter used will influence the response time if finger probes are used, but it will have little effect if ear probes are used.