The relation of premorbid cardiovascular risk attributes routinely measures at regular biennial examinations to cardiac necropsy findings using a special autopsy protocol was examined among 127 decedents of the Framingham cohort. Necropsy findings analyzed were: heart weight, left ventricular (LV) muscle thickness, percent luminal insufficiency of the coronary arteries, and percent intimal involvement with atherosclerosis. Clinical data analyzed included weight, height, blood pressure and serum cholesterol measured 1, 5, and 9 years prior to death. Systolic blood pressure, but not diastolic, correlated strongly with heart weight and LV muscle thickness in both sexes, and with atherosclerotic involvement in women, but not men. Serum cholesterol 1, 5, and 9 years antemortem all correlated positively with the degree of luminal insufficiency in men, while in women only cholesterol 9 years before death correlated significantly. In multivariate analysis only systolic pressure in women correlated independently with left ventricular muscle thickness and relative weight was the only independent correlate of heart weight. Only age was an independent significant correlate of the extent of coronary atherosclerosis in women. For men, coronary atherosclerotic involvement was independently correlated with only the serum cholesterol and measures of obesity were the major predictors of heart size. The degree of coronary pathology also correlated positively with heart weight in men and LV muscle thickness in women. In general clinical data measured just prior to death did not correlate as well with pathological findings as did earlier measurements.