T he cause of hypoxemia was studied in 21 patients with no previous heart or lung disease shortly after an episode of acute pulmonary embolism. The diagnosis was based on pulmonary angiography demonstrating distinct vascular filling defects or “cutoffs.” It was found that virtually all of the hypoxemia in patients with previously normal heart and lungs could be accounted for on the basis of shunt-like effect. The magnitude of the shunting did not correlate with the percent of the pulmonary vascular bed occluded nor with the mean pulmonary artery pressure. The shunts tended to gradually recede over about a month after embolism. Patients without pulmonary infarction were able to inspire 80-111% of their predicted inspiratory capacities, and this maneuver temporarily diminished the observed shunt. Patients with pulmonary infarcts were able to inhale only to 60-69% of predicted inspiratory capacity, and this did not reverse shunting. These data suggest that the cause of right-to-left shunting in patients with pulmonary emboli is predominantly atelectasis.When the elevation of mean pulmonary artery pressure was compared to cardiac index per unit of unoccluded lung, it fell within the range of pulmonary hypertension predicted from published data obtained in patients with exercise in all except one case. This observation suggests that pulmonary vasoconstriction following embolism is not important in humans, although these data are applicable only during the time interval in which our patients were studied and in patients receiving heparin.