Dogs were ventilated with a high-frequency oscillation device varying the frequency (5-15 Hz), the tidal volume (25-100 ml), and the resident gas (He, N2, SF6). Tidal volume was measured with a body plethysmograph. Blood gases were measured after a quasi-steady state was established. The kinematic viscosity of the breathing gas mixture, which changed by 1,700%, was found to have little effect on arterial PO2 and PCO2. The results are consistent with findings in a branched model that consisted of tubes with a diameter of 1 cm and with the theory of Taylor-type diffusion in turbulent flow. In addition, experiments were performed reducing and increasing the equipment dead space. This resulted in changes of PO2 and PCO2 that were appreciably less than those resulting from variations of tidal volume of the same magnitude.