1. Three preparations were used to study the actions of CO(2) on the pulmonary vasculature: isolated rat and cat lungs perfused at a constant flow rate with homologous blood and a lobe of cat lung perfused at a constant flow rate in vivo. In all three changes in pulmonary artery pressure (P(Pa)) reflected changes in pulmonary vascular resistance.2. In the isolated rat lung CO(2) caused vasodilatation when vascular tone was high. The lung was first ventilated with N(2) causing a rise in P(Pa). CO(2) caused vasodilatation during hypoxia whether the initial blood CO(2) level was low or normal and in spite of a fall in blood pH which usually augments the constrictor action of N(2).3. In the rat lung ventilation with CO(2) from the control state usually caused weak vasoconstriction. Reducing blood pH with acids also caused weak vasoconstriction while alkali caused vasodilatation.4. The over-all effect of CO(2) on the pulmonary vasculature depends on a balance between a vasoconstrictor action probably caused by carbonic acid and vasodilatation caused by some other property of the molecule. The dilator action is powerful in the isolated rat lung.5. By contrast, in both cat lung preparations, no direct evidence for a vasodilator action of CO(2) was obtained. Ventilation with CO(2) when vascular tone was raised by hypoxia, drugs or acids caused further vasoconstriction. From the control state CO(2) caused strong vasoconstriction.6. Indirect evidence from other work suggests that a pulmonary vasodilator action of CO(2) exists in the cat but is usually masked by the strong vasoconstrictor action of carbonic acid. In life the dilator mechanism may be important when pH changes caused by CO(2) are minimized by renal compensation.