We evaluated the ability of air and helium-O2 maximal expiratory flow-volume curves to distinguish upper airway obstruction from the diffuse, peripheral airway obstruction of chronic obstructive pulmonary disease. The increase in expiratory flows at peak, 75, 50, and 25 per cent of the vital capacity during helium-O2 breathing compared to air breathing was determined in 5 normal subjects and 3 patients with chronic obstructive pulmonary disease while breathing through fixed resistances, and in 6 patients with documented tracheal obstruction. In the normal subjects, the helium response at all 4 points remained normal and was unchanged from baseline until the simulated obstruction was severe (6-mm orifice), at which point all ofthe helium responses increased by 50 per cent. The patients with chronic obstructive pulmonary disease maintained their low baseline helium responses until the obstruction was severe (6-mm orifice), when only the expiratory flows at peak, 75, and 50 per cent of the vital capacity increased by at least 50 per cent. Five of the 6 patients with upper airway obstruction had helium responses very similar to those of the normal subjects with similar degrees of simulated obstruction, but the one patient with concomitant airway obstruction extending well below the carina had very small helium responses at each point. We conclude that upper airway obstruction can usually be identified by high helium responses and that upper airway obstruction, if severe, can be identified even in the presence of more peripheral airway obstruction by a normal helium response at high lung volumes.