The reliability of casual blood pressures for reflecting blood pressure status and predicting cardiovascular sequelae of hypertension was examined in the Framingham cohort of 5209 men and women followed for 18 years. Blood pressures were more variable in persons with higher levels. After controlling for pressure level the degree of variability in an individual at one point in time did not correlate with the degree of varability at another time. Although a single casual measurement does not afford a precise characterization for an individual it was found to be highly predictive of future cardiovascular disease. A series of blood pressure measurements (averaged) improved the predictability somewhat but this seemed to be fully explicable by the greater stability of an average of several measurements as against a single measurement.