Patterns of systolic and diastolic blood pressure were assessed in adolescents who were 12-13 at the time of first examination. Three-year follow-up consisted of 5 check-ups. Upon the first examination, all adolescents were divided into three groups: those with low (below the 10th percentile), normal (between the 10th and 90th percentiles) and high (above the 90th percentile) arterial BP values. Repeated checkups covered all adolescents with low and high arterial BP as well as a 20% sample of normotensive subjects. Over the three years of follow-up; the greatest increment in systolic and diastolic arterial BP was noted in the hypotensive group, and the smallest, in the hypertensive group. Various hypotheses are discussed that might explain the demonstrated regularity; the "regression-to-mean" effect is proposed as a possible explanation.