The quantitative analysis of left ventricular shape in ischaemic heart disease has seldom been performed because of the lack of a reliable and reproducible method of analysis. Diastolic and systolic left ventricular outlines in two groups of patients with previous myocardial infarction (anterior in 13 cases and inferior in 14 cases) were studied by analyzing the regional curvature and the power spectrum. This method allowed us to obtain left ventricular curvatures from the mitral to the aortic corner and the power spectrum of the first 12 harmonics. The results obtained in these two groups were compared with those obtained in 16 normal subjects. The diastolic power spectrum in both infarcted groups was similar. It was characterized by a double peak which was different from that of the normal subjects. Slight differences between the two infarcted groups were due to the diverse amplitude of the highest harmonics. The regional analysis of the curvature in both groups showed similar abnormalities of the posterobasal, inferior and antero-basal regions. The posterior wall showed a uniform curvature with the point of minimum shifted towards the mitral corner; the anterior wall showed a rounded profile with a regular curvature. In the group with anterior myocardial infarction the curvature of the inferior wall was negative, i.e., convex towards the left ventricular cavity. The systolic power spectrum showed a double peak profile which was different from that of the normal subjects. There were some differences between the two groups as regards the first and the highest degree harmonics. In inferior myocardial infarction the apex was rounded whereas in the anterior one the most important abnormalities were the convexity of the inferior wall towards the inside and the presence of a region with minor curvature between two regions with greater curvature of the anterior wall. Some of the systolic abnormalities involved the probable site of the infarct while others were in remote regions. The meaning of remote abnormalities is not clear. However, we did not verify a correlation between wall motion, at least as shortening of radii, and regional curvature. The abnormalities of the diastolic outline were independent of the site of the infarct and did not appear to be correlated to end-diastolic pressure or to the ejection fraction. They seemed to be the morphological counterpart of the filling abnormalities reported in ischaemic cardiac disease and they may depend on the regional distribution of stresses.