The prevalence of mixed hypertension (MHT), isolated systolic hypertension (ISH) and isolated diastolic hypertension (IDH) was estimated in the elderly population in the register of a large general practice in Wrexam, North Wales. Of the 3289 elderly patients, born in 1927 or before, entered in the register of surgery, 1901 attended for the first screening. The mean SBP rose with age until the age of 80-84 years in males and 75-79 years in females and then gradually declined. The mean DBP showed an earlier decline in mates than in females. The prevalence of hypertension at first screening was: mixed hypertension 9.8%, ISH 19.1% (DBP < 95 mmHg*, 23.1%*) and IDH 5.7% with a total prevalence of hypertension of 52.2%. The prevalence fell at each subsequent screening so that at the third screening MHT was 3.9%, ISH 4.2% (5.4%)* and IDH 1.0%, with a total prevalence of hypertension of 10.3%. The prevalence of ISH rose with age until 70-74 years of age and with the maximum prevalence in this age group and then gradually declined. There was a drastic drop in the prevalence of both mixed hypertension and IDH after the age of 70-74 years. This study provides data for this community and also supports earlier observations that hypertension is a common problem in the elderly and that ISH is the commonest form of hypertension in the elderly. It confirms the fall in mean DBP with age but reports a decline also in mean SBP after the age of 80-84 years in males and 75-79 years in females, There is also a declining prevalence of all forms of hypertension after the age of 70-74 years, which may be a reflection of the increased cohort mortality, elderly hypertensives after this age group being more susceptible to cardiovascular and cerebrovascular complications of hypertension and so making them inaccessible for screening. It also confirms the observation that the prevalence of isolated systolic hypertension varies according to the definition used and the number of clinic visits, supporting the suggestion that the diagnosis of hypertension by casual clinic measurements should be based on multiple readings registered on at least three separate visits.