Sixty-two patients with isolated aortic valvular stenosis were analyzed by a series of common noninvasive procedures and by cardiac catheterization. The data from 50 of these were evaluated in a retrospective fashion by multiple regression methods to determine significant objectively obtained predictors of aortic-left ventricular gradient and valvular area. Formulae were derived from these analyses and an additional 12 patients were then studied prospectively to evaluate the validity of the predictive formulae. Forty-three of 50 patients (86 per cent) were correctly identified as to a gradient of greater or less than 50 mm. Hg in the initial group, and all those in the prospectively studied sample were correctly classified. Thiry-five of 43 patients (82 per cent) of those with valve area data in the first application were correctly classified as to valve area or greater or less than 0.8 cm.2, and all patients in the prospectively studied group were appropriately identified as to the same area. The combined application of the observations of calcification of the aortic valve, shudder waves on the anacrotic limb, prolonged time to peak of the percussion wave and alteration of the dicrotic notch of the carotid pulse tracing, left ventricular hypertrophy by electrocardiogram, and the altered duration of ventricular ejection time were reliable predictors of elevated aortic-left ventricular gradient and decreased aortic valvular size.