This report, the second in a series on pulse pressure and pure systolic hypertension, examined in prospective analyses the associations between both the initial values of five variables and the changes in these variables and pulse pressure, utilizing data from three Chicago epidemiologic studies, in order to determine whether variables known to be related to blood pressure and hypertension are related to pure systolic hypertension or 'classical' hypertension. In these analyses follow-up pulse pressure, which was measured from 2-5 years after the initial measurement of the other variables, was redefined so that the association between the initial value or change and pulse pressure indicated whether the initial value or change was more strongly related to follow-up systolic or diastolic blood pressure. In these three studies, only the initial value for cigarette use had a consistent positive association with follow-up pulse pressure. Change in heart rate was generally positively related to follow-up pulse pressure, while the initial value was not. For relative weight and serum cholesterol, both the change and the initial tended to be negatively related. For glucose, the association was not consistent for either the initial value or the change. The results from these prospective analyses thus suggest that cigarette use is related to pure systolic hypertension, rather than 'classical' hypertension.