Mongrel dogs weighing 15-25 kg and anesthetized with thiopental-gamma-hydroxybutyric acid were used to investigate the effects of pulmonary gas embolism on pulmonary arterial pressure (Pap), systemic arterial pressure (Pa) and cardiac output (Q). Pulmonary gas embolism was produced either by venous injecton or by venous infusion. The most marked effect of pulmonary gas embolism on circulation was an increase in Pap which returned to the original level after stopping the gas administration. 1. After gas injection Pap rose to a maximum within 30--60 s. The extent of this rise in Pap showed a positive correlation with the volume of the injected gas. The kind of gas (oxygen, helium, neon, nitrogen, air), however, did not influence the extent of the rise in Pap, but did influence the time of return of Pap to the original level. Carbon dioxide showed an exceptional behavior in that it had almost no effect on Pap at all. P a hardly changed with the volume of the gas injections (20--60 ml injected within 1 s); Q was not measured after gas injection (the direct Fick method is not usable in this situation). 2. Gas infusion caused a slow rise of Pap, its steepness and extent depending on the rate of infusion and on the physical properties of the infused gas. When the right ventricle was able to maintain its output, a constant level of Pap was reached after 10--15 min. In this circulatory steady state Pap appeared to be a measure of the degree of embolization. However, this relationship no longer held when the right ventricle failed as evidenced by a fall in Pap, Pa and Q. It may be concluded that pulmonary gas embolism produces a transient partial obstruction in the pulmonary circulation and that the performance of the right ventricle determines the maximum degree of embolization compatible with a sufficient circulation.