hypertension.5 Measures to prevent the development of high blood pressure would presumably be even more effective. It is generally accepted that weight, heart rate, blood glucose, salt and alcohol consumption, hematocrit and smoking status correlate with blood pressure levels.68" However, with few exceptions, these studies have been cross-sectional, so it is not clear whether these correlated conditions predate, accompany, or occur after elevation in blood pressure. Data collected by the Normative Aging Study allow prospective examination of the effects of these and other factors on the subsequent rate of change in blood pressure and on the development of elevated levels of blood pressure. A sample of men recruited from the Boston area has been under surveillence since 1963. The subjects who were initially normotensive