A cohort of children and adolescents (n = 801) aged 5 to 11 years living in Bogalusa, Louisiana was examined in three consecutive cross-sectional surveys over a 6 year period. The relationship between fasting insulin and glucose at baseline and longitudinal changes in blood pressure were examined. Significantly positive correlations were observed between fasting insulin and glucose at baseline and systolic and diastolic blood pressure at follow-up in white boys and girls (r = 0.19 to 0.38, P < .01), but not in blacks. After adjustment for several covariates, the association still existed in whites. In multiple regression analysis, fasting plasma insulin was observed to be a major contributor of subsequent systolic blood pressure levels independent of age, sex, height, obesity, and glucose levels in white children. These data suggest that insulin can be a determinant of blood pressure levels in children as suggested in adults. The relationship of insulin to blood pressure differs between black and white children and is likely modulated by the multiple mechanisms active in maintenance of blood pressure.