Résumé : To further explore the mechanism of hypoxic pulmonary vasoconstriction, we studied the mean pulmonary arterial pressure (Ppa)/left atrial pressure (Pla) relationship at fixed cardiac index (Q) and the Ppa/Q relationship at several levels of fixed Pla in pentobarbital sodium-anesthetized dogs ventilated alternately in hyperoxia [fraction of inspired O2 (FIO2) 0.4 or 1.0] and in hypoxia (FIO2 0.1). In all experimental conditions, Ppa/Q plots were linear with extrapolated pressure intercepts (Pi) not significantly different from Pla. Hypoxia increased the slope of Ppa/Q plots and did not affect Pi. In hyperoxia, increasing Pla (3 to 26 mmHg) induced approximately equal increases in Ppa at fixed Q and shifted Ppa/Q plots toward higher pressures in a parallel manner. In hypoxia, increasing Pla (4 to 25 mmHg) did not affect Ppa at fixed Q until Pla exceeded 16 mmHg and shifted Ppa/Q plots toward higher pressures with a decrease in slope. Consequently, the hypoxia-induced increases in Ppa at constant Q and constant Pla were attenuated at higher Pla. Thus, in anesthetized dogs, hypoxia increases the slope of Ppa/Q plots without affecting Pi at fixed Pla, and an increase in Pla inhibits hypoxic pulmonary vasoconstriction. These results can be explained without invoking a hypoxia-induced Starling resistor mechanism in the pulmonary circulation.