Endotracheal tube "leak" is often estimated in children to judge the fit of uncuffed endotracheal tubes within the trachea. Twenty-five swine were intubated with uncuffed tracheal tubes to determine whether a more sensitive measurement of leaks could be devised and whether leak pressure estimates fit between tracheal tube and trachea. We compared leak pressure measurement using a stethoscope and aneroid manometer with a technique using a microphone, pressure transducer, and recorder, and found no differences between the two methods. The tracheas were then removed and slides prepared of tracheal cross-sectional specimens. Regression analysis revealed a linear relationship between tracheal lumen size and tracheal tube size for both low leak pressure (y = -0.4 + 0.79x, r = 0.88, P less than 0.05) and high leak pressure (y = -2.9 + 0.71x, r = 0.92, P less than 0.05) groups. We conclude that leak testing with a stethoscope and aneroid manometer is sensitive and accurate, and that tracheal tube leak pressure accurately portrays fit between tube and trachea.