Correlates of present blood pressure status are analyzed in 576 subjects (271 males, average age 32 years) in Tecumseh, Michigan. In addition to the current values, anthropometric and blood pressure data are available when the subjects were, on average, 6.9 and 22.2 years of age. Data on 351 fathers and 368 mothers when they were, on average, 32 years old are also available. Moderate, but significant correlations were found between present and past blood pressure and between past weight or skinfold thickness and the present blood pressure. These correlations were much weaker for childhood values than for values at age 22. When multiple regression techniques were used with blood pressure values at age 22 and parental values as independent variables, 40% of the present systolic blood pressure variance could be explained. Prediction of present hypertension status (blood pressure greater than 140 and/or 90 mm Hg) was evaluated by discriminant analysis. Three variables (weight at age 22, systolic blood pressure at age 22, and father's diastolic blood pressure) entered the model and accurately predicted the present blood pressure classification in 89% of the sample. When current blood pressure status was assessed with respect to previous blood pressure classification (upper 20%), family background, and overweight, a gradient of risk for hypertension was found. On the low end of risk was high childhood pressure (risk 19.1% versus 12.1% in the overall population). The highest risk occurred for those with high pressure and overweight at 22 years who also had a family background of high blood pressure (44% versus 12.1%). The prediction of hypertension from young adulthood to the early fourth decade of life is feasible and permits delineation of populations targeted for primary prevention.