High-frequency external body vibration, combined with constant gas flow at the tracheal carina, was previously shown to be an effective method of ventilation in normal dogs. The effects of frequency (f) and amplitude of the vibration were investigated in the present study. Eleven anesthetized and paralyzed dogs were placed on a vibrating table (4-32 Hz). O2 was delivered near the tracheal carina at 0.51.kg-1.min-1, while mean airway pressure was kept at 2.4 +/- 0.9 cmH2O. Table vertical displacement (D) and acceleration (a), esophageal (Pes), and tracheal (Ptr) peak-to-peak pressures, and tidal volume (VT) were measured as estimates of the input amplitude applied to the animal. Steady-state arterial PCO2 (PaCO2) and arterial PO2 (PaO2) values were used to monitor overall gas exchange. Typically, eucapnia was achieved with f greater than 16 Hz, D = 1 mm, a = 1 G, Pes = Ptr = 4 +/- 2 cmH2O, and VT less than 2 ml. Inverse exponential relationships were found between PaCO2 and f, a, Pes, and Ptr (exponents: -0.69, -0.38, -0.48, and -0.54, respectively); PaCO2 decreased linearly with increased displacement or VT at a fixed frequency (17 +/- 1 Hz). PaO2 was independent of both f and D (393 +/- 78 Torr, mean +/- SD). These data demonstrate the very small VT, Ptr, and Pes associated with vibration ventilation. It is clear, however, that mechanisms other then those described for conventional ventilation and high-frequency ventilation must be evoked to explain our data. One such possible mechanism is forcing of flow oscillation between lung regions (i.e., forced pendelluft).