Diminished echocardiographic aortic cusp separation is used as one indicator of the severity of aortic stenosis. To test the validity of this index, 22 patients--12 (55 percent) with isolated aortic valve disease and 10 (45 percent) with aortic stenosis associated with mitral or coronary artery disease--underwent M mode echocardiographic examination before aortic valve replacement. Tracings of diagnostic quality were obtained without difficulty from all 22 patients. Cardiac catheterization was performed in 21 patients. Echocardiographic cusp separation was measured from the apparent mid-systolic orifice and from the outer periphery of the anterior cusp to the outer periphery of the posterior cusp (maximal peripheral cusp separation). Mid-systolic cusp separation varied in nearly every patient, depending on the angle of leaflet presentation to the ultrasonic beam. Maximal peripheral cusp separation measured 16 mm in 18 of 22 patients (82 percent); it indicated neither the severity of the aortic stenosis as documented with cardiac catheterization nor the mobility of the cusps seen at operation. The surgical and echocardiographic descriptions of leaflet and aortic root calcification were similar. In situ examination of aortic cusp separation indicated that diseased aortic valves are not comparable with normal valves or valves with uncomplicated congenital obstruction. The aortic leaflets were curled, fused, calcified and deformed from their natural state of coaptation. Abnormal thickening and limited or eccentric mobility of the aortic leaflets were useful indicators of the cause of valve disease, but cusp deformity secondary to aortic stenosis invalidated mid-systolic cusp separation and maximal peripheral cusp separation as indicators of the severity of aortic stenosis.