The antihypertensive effect and metabolic side effects of bendroflumethiazide have been compared with those of propranolol in two randomly selected groups, of 53 previously untreated middle-aged men during 6 years' treatment for mild to moderately severe essential hypertension. The blood pressure-reduction was the same in the two groups. During the follow-up 1 man in the bendroflumethiazide group and 3 in the propranolol group died while 2, 1 on each treatment, became diabetic. None had gout but serum urate increased in both groups. Glucose tolerance improved significantly in both groups during the first year and this improvement was sustained for the follow-up period. Serum potassium did not differ in the two groups during the first 5 years but during the sixth year it decreased in the diuretic group. Total potassium was, however, unchanged in both groups. These results indicate that the frequency of metabolic side effects during diuretic treatment of mild to moderately severe essential hypertension is low and has been grossly exaggerated. Since the antihypertensive effect and side effects were equal with both drugs, and since the diuretics are cheaper, they should be the drug of first choice in this type of hypertension.