Measurements made from carotid pulse tracings and phonocardiograms are helpful in the prediction of severe aortic stenosis in young patients. In the elderly, however, analysis of the carotid pulse tracing and, in particular, its maximum rate of rise has proved unreliable because the range of normal values for this age group has not been established. Consequently, we studied 44 young normal and 44 older normal subjects, 15 older patients with systolic hypertension, and 36 older patients with significant aortic stenosis. Measurements obtained from the carotid pulse included the left ventricular ejection time, upstroke time, half-rise time, and the maximum rate of rise. The peak of the systolic murmur in aortic stenosis was measured phonocardiographically from the Q wave of the simultaneously recorded electrocardiogram (QP interval). The results confirm that the maximum rate of rise of the carotid pulse of normal older subjects is much more rapid than that of young subjects. Among elderly patients with aortic stenosis, the maximum rate of rise falls within the range of the normal young. Measurement of the half-rise time is imprecise, and there is considerable overlap with the normal. This is also true of left ventricular ejection time. The upstroke time is unreliable because of wave distortion produced by the arteriosclerotic vascular bed. The two most valuable and easily recorded measurements are the maximum rate of rise and the QP interval. A maximum rate of rise of more than 1000 mmHg/s and a QP interval of less than 220 ms argue strongly against the diagnosis of severe aortic stenosis. These measurements provide a simple, non-invasive means of avoiding left heart catheterisation in older subjects suspected of having severe aortic stenosis.