The authors assessed the effect of cardiac output on the rate of clearance of air emboli from the superior vena cava (SVC) in 11 dogs placed in a supine 30° head-up inclined position. Venous air emboli were produced by infusing 4 ml of air into the dorsal sagittal sinus. Air emboli appearance and clearance from the SVC were determined from the analog record produced by a transesophageal Doppler ultrasound monitor. Cardiac output was altered by changing the level of anesthesia and the administration of intravenous fluid. Eighty-eight per cent of air infusions resulted in an analog response of consistent amplitude indicating that all or a similar fraction of the infused air arrived consistently at the ensonified region of the SVC. In this group, air emboli passed from the SVC with a clearance time (CT) that was related inversely to the cardiac index (CI). CT = 24 CI-1.3with a correlation coefficient = 0.57. Following 12% of the air infusions, there was evidence that a part of the air became detained in vascular sites cranial to the ensonified portion of the SVC. This evidence consisted of the following observations: analog responses were diminished in amplitude; subsequent air injections evoked responses that sometimes were exaggerated; and/or subsequent increases in cardiac output alone, produced spontaneous analog responses characteristic of air emboli. The authors postulate that air similarly may be detained in clinical settings and may explain such phenomenon as persistent precordial Doppler sounds of air emboli after pulmonary artery pressure has returned to normal following symptomatic air embolism and occurrence of air emboli after scalp closure.