Since a positive association between blood pressure and plasma glucose level in child populations has been observed previously, we studied this relationship in a stratified random sample of children ages 7-15 years drawn from a total geographic population. We stratified this population (N = 3524) by diastolic blood pressure and weighted the extreme strata. The sample of children (N = 272) was free of secondary hypertensives and was tested by a restricted glucose tolerance test. As a parameter of cellular glucose kinetics, a "peripheral insulin resistance" was defined as one-hour glucose (mg/dl) X one-hour insulin (microU/ml). For white boys, this product showed an increase from 2745 in the low blood pressure strata to 6615 in the high blood pressure strata. To assess the relationship between blood pressure measured during the reexamination and fasting plasma glucose level, we classified the children according to their peripheral insulin resistance in a low, medium, and high group. After controlling for body weight, only white children in the high blood pressure strata with high insulin resistance had a positive association between fasting plasma glucose and systolic blood pressure. The relationship may be initiated by inappropriate intravascular water retention influencing the blood pressure in children with high peripheral insulin resistance and susceptible to hypertension.