Trachea is an alternative site for pulse oxygen saturation monitoring. The response time of the oximetry probe has been reported more rapid when placed in central than in peripheral. The purpose of this study is to compare the performance of the oximetry probes placed in trachea and peripheral site during rapid desaturation. Endotracheal tubes with an oximetry sensor were intubated in ten anesthesia dogs. Both the oxygen saturation signals from trachea (S(t)O(2)) and tail (S(p)O(2)) were shown on the same monitoring screen. The mechanical ventilation was disconnected to produce a rapid desaturation when both S(t)O(2) and S(p)O(2) were 100%, and the mechanical ventilation was reconnected when S(p)O(2) decreased to 80%. The time of S(t)O(2) and S(p)O(2) dropped to 95, 90, 85, and 80% was recorded, respectively during the mechanical ventilation disconnection, and the arterial blood was sampled for arterial oxygen saturation (S(a)O(2)) measurement simultaneously. The levels of measurement agreement between two oximetry readings (S(p)O(2), S(t)O(2)) and S(a)O(2) were analyzed, respectively with the Bland and Altman method. The mean response time of S(t)O(2) was significantly shorter than S(p)O(2) when both of them decreased from 100 to 80% (172.6+/-68.9 vs 220.7+/-72.3 s) during rapid desaturation. The 95% confidence interval for absolute difference between S(p)O(2) and S(a)O(2) was 4.12+/-6.47%, and between S(t)O(2) and S(a)O(2) was 3.33-3.46%. Oxymetry placed in trachea provides a better monitoring for detecting rapid desaturation than in peripheral.