The presence of unidirectional flow in the avian lung is thought to be effected by aerodynamic 'valves'. First we review the history of this hypothesis and summarize existing evidence. Second, we present a semiquantitative treatment of the various fluid dynamic factors that may be involved in directing fluid flow. The resulting calculations show in some detail how the inspiratory valve may work, and upon what mechanisms it may depend. Our calculations suggest that gas convective inertial forces are sufficient to effect inspiratory valving. Finally, we give some heuristic arguments regarding the mechanisms of expiratory valving.