Cumulative data indicate that essential hypertension begins to develop during the first two decades of life. However, it is currently impossible to identify which child will become a hypertensive adult. We investigated familial aggregation of blood pressure in 1,183 nuclear families (including mother, father, and their first two children) in a rural community in China. Information on blood pressure (BP) (via direct measurement) and related risk factors (via an interview) was obtained from all subjects. A generalized additive model was used to explore independent relations of the BP among the four family members, with adjustment for sex, age, height, weight, education, smoking, and alcohol use. A logistic regression model was used to assess the risk of high BP (>90th percentile of percent predicted value) in the second sibling by parental and the first sibling's BP. There was a significant parent-child and sib-sib correlation of BP. The parent-child correlation of BP was similar for the first and second child. There was no significant difference between mother-child and father-child correlation. The rate of high systolic BP in the second sibling was lowest (3.1%) when both parents and the first sibling were in the low BP tertile (low-low group) and highest (24.4%) when both parents and the first sibling were in the high BP tertile (high-high group). The rates were intermediate if only the parents (10.7%, high-low group) or only the first sibling (7.6%, low-high group) was in the high BP tertile. With the odds ratio (OR) for the low-low group as the reference value(1.0), the OR for the high-high group was 10.3 (95%CI 3.9-35.3); for high-low group 3.8 (95%CI 1.3-13.7); and for low-high group 2.6 (95%CI 0.9-9.87); indicating an interaction of the parents' BP and the first sibling's BP on an additive scale. The above analysis was performed for four age groups: 6-10, 11-14, 15-19, ≥20 years and no significant age-specific difference in association was found. A similar pattern was found for diastolic BP. The data indicate a strong familial aggregation of BP in this population and show that such a familial influence on BP can be detected from early childhood onward.