The pattern of intrapulmonary pressure distribution was studied during high-frequency ventilation in order to explain the inconsistent results reported in the literature. Methods. Pressure and flow velocity (hot-wire anemometry) were measured in different lung compartments: 1. In transalveolar chambers sealed to the perforated pleural surfaces of dried pig lungs; 2. In emphysema-simulating airbags sealed to the isolated bronchial trees of dried pig lungs; and 3. In transalveolar chambers sealed to the perforated pleural surfaces of freshly excised pig lungs. Results. 1. The pressure amplitudes change from one area to another and depending on the exciting frequency. 2. High-frequency oscillation is associated with an increase in pressure amplitude when the exciting frequency rises, whereas with conventional high-frequency jet ventilation the pressure amplitude is more likely to decrease with frequency. 3. During high-frequency jet ventilation the local pressure amplitude changes with the position of the tube in the trachea rather than with the exciting frequency. 4. When the volume of the measuring chamber is doubled the resulting pressure amplitude falls to half the control value. 5. The pressure amplitude and mean pressure measured in the transalveolar chamber vary more or less independently from the peak flow velocity. High-frequency ventilation is thus seen to be a frequency-dependant, inhomogeneous mode of ventilation that can essentially be homogenized by systematically changing the exciting frequency. The frequency-dependant response to different lung areas to excitation is likely to result from an intrabronchially-localized aerodynamic effect rather than the mechanical properties of the lung parenchyma.