We measured the separation of the vocal folds, the glottis chink, on inspiration and expiration during quiet breathing in 34 patients with varying degrees of airflow obstruction as measured by forced expired volume in one second, Width of the glottis chink was calculated from photographs of the vocal folds taken via a fiberoptic bronchoscope. To adjust the differences in photographic image size caused by variations in instrument position, we predicted the internal anteroposterior (A.P) diameter of the glottis in each patient. This was made possible by the observed high correlation (r = 0.79) between A.P diameter of the glottis and height demonstrated in a separate study in 49 adult cadavers. In patients with airflow obstruction, the glottis chink was narrowed during quiet breathing, particularly on expiration. Such differences in glottis width between patients could not be attributed to frequency of breathing or tidal volume. Indeed, during high frequency breathing (panting) at 1 to 3 Hz studied in 18 patients, further glottis narrowing was commonly observed in those with airflow obstruction. Furthermore, when 13 patients performed a maximal exhalation, in those with airflow obstruction, the glottis remained narrowed, whereas in patients with FEV1 greater than 80% predicted, the glottis opened to an inspiratory width.l These observations suggest that narrowing of the glottis potentially plays an important part in controlling airflow in patients with airway obstruction.