The sensitivities and durations of positive responses of routinely used methods for detecting venous air embolism and physiologic variables reflecting changes accompanying air embolism were compared in 10 anesthetized dogs receiving incremental venous injections of 0.1, 0.25, 0.5, 0.75, 1.0, 1.5 and 2.0 ml/kg air. These doses of air produced a spectrum of changes ranging from subtle alterations in Doppler frequency to dramatic, life-threatening cardiovascular depression. Precordial Doppler ultrasound monitoring was the most sensitive detection method, but failed to reflect embolus size. Mean pulmonary arterial pressure .**GRAPHIC**. and end-tidal CO2 concentration (FETCO2), while slightly less sensitive than the Doppler device, reliably detected 0.25-0.5 ml/kg of injected air and quantitatively reflected the size of the embolus. Changes in Doppler, .**GRAPHIC**. and FETCO2 were predictably positive at air volumes 1/4 of those producing decreases in arterial blood pressure and cardiac output. Other frequently monitored signs of air embolism, such as a millwheel murmur and cardiac arrhythmias, were late manifestations and were seen only with injection of 1.5-2.0 ml/kg air. .**GRAPHIC**. and FETCO2 remained altered the longest after air injections (20-30 min at 2.0 ml/kg), while changes in Doppler-transmitted signals at the same dose lasted 15 min. Positive responses with other methods of air embolus detection, although dose-related, were significantly shorter-lasting than .**GRAPHIC**. and FETCO2 alterations. Cardiac murmurs and arrhythmias were transient after air injection, lasting less than 1 min, even with emboli of 2.0 ml/kg. The combination of application of a Doppler device with either end-tidal CO2 measurement or a flow-directed pulmonary-artery catheter with .**GRAPHIC**. display provided the desired combination of sustained, sensitive and reliable detection of air embolism and quantitative evaluation of embolus size.