The mortality of 3783 non-malignant hypertensive patients attending the Glasgow Blood Pressure Clinic between 1968 and 1983 and followed for an average of 6.5 years was compared with that in three control groups: the general population of Strathclyde a group of 15 422 subjects aged 45-64 years and screened in Renfrew and Paisley between 1972 and 1976, and a group of hypertensives seen in a blood pressure clinic based on general practice in Renfrew.Average blood pressure for men at entry to the Glasgow Clinic was 181/111 mmHg falling to 158/96 mmHg during treatment. Corresponding values for women were 185/109 mmHg and 161/96 mmHg. Seven hundred and fifty clinic patients (451 males) died during follow-up, the commonest causes of death in both sexes being myocardial infarction and stroke. All-cause age-adjusted mortality (deaths per 1000 patient-years) was 41.4 for men and 22.1 for women. At all ages in both sexes and for all levels of initial blood pressure mortality was less in patients whose blood pressure was reduced most. Without a randomized control group it is not certain that lower mortality in those with well controlled blood pressure was due to treatment, although this is the most likely explanation. Cigarette smoking, a history of myocardial infarction, angina or stroke, retinal arterio-venous nipping, raised blood urea, an abnormal electrocardiogram (ECG) and secondary hypertension were associated with increased risk, but heavy alcohol intake, obesity, haematocrit >45%, hypokalaemia and social class were not.Life table analysis showed that, despite some reduction of mortality by treatment, the relative risk to men and women in the clinic remained two- to five-times that of the general population. The benefits of treatment were not such as to restore normal expectation of life even when blood pressure was well controlled. Excess mortality in the clinic could not be explained by difference of smoking habit or social class. This suggests that there is in the hypertensive patients of the Glasgow Clinic an element of irreducible risk, that treatment may be beneficial in some respects but harmful in others, or that patients at particularly high risk are selectively referred to the clinic.