Blood pressure measurements in 7735 middle-aged men from general practices in 24 towns in England, Wales and Scotland provide information on the prevalence of hypertension and its management in Great Britain. Despite a substantial correlation (r = 0.70) between systolic and diastolic blood pressures, individuals can show considerable discrepancies between these two measurements; they are not interchangeable. This observation has important implications for the choice of criteria used to define hypertension. However defined, the prevalence of hypertension increases markedly with age, increasing body mass index and with heavy alcohol consumption. It is not related to smoking and only to a small extent to social class. Diastolic hypertension (greater than or equal to 90 mmHg) was present in 26% and systolic hypertension (greater than or equal to 160 mmHg) in 22% of these men. In both systolic and diastolic hypertension, only one quarter of affected men could recall having been diagnosed as hypertensive by a doctor, and only one third of these were on regular antihypertensive treatment. There is a threefold variation in the prevalence of measured hypertension in the 24 towns with a trend towards higher rates in Northern England and Scotland. No relationship was seen between the prevalence rates of measured hypertension in the towns and the rates of doctor diagnosis of hypertension. Cardiovascular mortality rates in the towns were correlated with the measured prevalence rates for systolic and diastolic hypertension (r = 0.70 and r = 0.57, respectively). The geographic variations in blood pressure and hypertension in Great Britain provide a major opportunity for research into the causes of 'essential' hypertension.