This prospective study determines the degree of tracking and investigates factors predicting a rise in blood pressure (BP) quartile in a cohort of 1369 subjects who were followed for 11 years from childhood into young adulthood. In 900 of these subjects BP, height, weight, physical fitness and BP responses to a maximal exercise testing were measured both at baseline and at follow-up. BP, weight, height and body mass index (BMI) were divided into sex-specific quartiles at both examinations. Tracking was evaluated by examining the tendency of remaining in the same quartile from baseline to follow-up and by measuring product-moment correlation coefficients. Tracking in the upper and lower quartile for BP, weight, height and BMI were significant. Odds ratios for staying in the upper or lower quartile through the follow-up period ranged from 1.6 to 2.4 for diastolic BP and from 2.1 to 3.1 for systolic BP. The range of correlation coefficients for the anthropometric measurements were 0.57-0.75, for diastolic BP 0.12-0.22 and for systolic BP 0.34-0.36 respectively. Changes in weight or relative weight as well as BP response to an exercise test were the factors which predicted a rise in quartile through the 11 years of follow-up. The existence of the inevitable regression to the mean problem in large longitudinal studies of BP was demonstrated by the finding of baseline BP being a significant factor in the prediction of rising in systolic, diastolic or both systolic and diastolic BP quartiles.