Arterial blood pressure was studied retrospectively in 17,067 children and adolescents (aged 4-18 years - 8,647 males and 8,420 females) from cross-sectional data collected at the Preventive medicine centre in Nancy between 1977 to 1979. The distribution of blood pressure was expressed in percentiles by age and by height. Diastolic blood pressures are more widely scattered than systolic pressures as has been found in other studies and may be due to different methods and definitions used to determine the diastolic pressure. Arterial blood pressure seems to be more closely correlated to height than with age and it is suggested that when assessing blood pressure in children and adolescents curves of arterial blood pressure compared to height should be used. The definition of arterial hypertension remains difficult. On the basis of the centiles, 3 groups of children may be defined, one with borderline hypertension, one with proven hypertension and one with dangerous hypertension.