The avian lung: is there an aerodynamic expiratory valve?

The unidirectional gas-flow pattern through the avian lung is thought to result from 'aerodynamic valves'; support for this hypothesis lies mainly in the failure to find any evidence for anatomical valves. During expiration, air flows from the caudal air sacs through the major exchange area of the lung, the paleopulmonic parabronchi, instead of bypassing the lungs via the intrapulmonary bronchus. We tested whether the effectiveness of this expiratory flow control mechanism depends on aerodynamic factors, especially convective inertial forces that depend on gas density and flow velocity. In pump-ventilated, anaesthetized geese, a bolus of tracer gas was introduced into both the right and left caudal thoracic air sacs during an end-inspiratory pause. During the first expiration, the rise of tracer levels within the caudal trachea was measured. Valve efficacy was positively correlated with the rate of expiratory gas flow, V·ao (range 8­200 ml s-1). At flows assumed to occur during exercise in geese (V·ao>100 ml s-1), the expiratory valve efficacy was approximately 95 %; it was less effective at lower flows. Surprisingly, the density (rho) of the background gas (rho of He/O2=0.43 g l-1, Ar/O2=1.72 g l-1 or SF6/O2=5.50 g l-1) had no effect on expiratory valving. We suggest two possible mechanisms that might explain this unusual combination of flow dependence without density dependence. (1) If airway geometry changes occurred between experiments with different gases, flow in the vicinity of the expiratory valve may have varied independently from flow measured at the airway opening. (2) Alternatively, valving may depend on dynamic compression of the intrapulmonary bronchus, which could depend mainly on viscous resistance and thus on flow velocity but not gas density.