1. Blood pressure has been measured in a random sample of 5852 subjects (both sexes, 20-64 years old). The prevalence of hypertension (blood pressure greater than or equal to 160 mmHg and/or 95 mmHg; 150 mmHg and/or 90 mmHg under the age of 30 years) was 29-5%. 2. In a cohort of 972 detected hypertensive subjects, longitudinal changes (2 and 5 months after the initial examination) were evaluated. The phenomenon of regression toward the mean value was evident. 3. The implications of the variability of blood pressure in a community control programme of hypertension are discussed.