In all subjects of the Australian therapeutic trial in mild hypertension, mean pressures for the two screening visits were within the range 95-109 mm Hg for diastolic blood-pressure phase V(DBP) and less than 200 mm Hg for systolic blood-pressure (SBP). In the 1943 control (placebo) subjects mean blood-pressures fell from 158/102 mm Hg at the first screening visit to 144/91 mm Hg 3 years later. At that time pressures remained within the mild hypertension range in 32%, ahd risen above it in 12%, and had fallen below in 48%. Trial end-points (ischaemic heart disease or cerebrovascular accident) occurred in 8%. The outcome was related to the level of initial pressure but not to other characteristics measured at entry. The mean initial pressures of 22 subjects who experienced a cerebrovascular event were higher than those of a matched group with no hypertensive complications, but the 88 subjects who experienced ischaemic-heart-disease events had initial pressures similar to those in a matched control group. The trial end-point rate was related to the average DBP of subjects throughout the trial in those with average DBP greater than or equal to 95 mm Hg, and at those levels subjects on active treatment had a higher incidence than subjects of the placebo group with the same DBP level. For those with average DBP below 95 mm Hg the incidence of trial end-points was not related to blood-pressure level or treatment. 16% of placebo subjects in this mild hypertensive population had a mean DBP of less than 95 mm Hg at the first three visits. If this were taken as an indication to withhold drug treatment, 3 years later one-quarter of them (4% of all subjects) would be found to be hypertensive or to have experienced a trial end-point, and thus inappropriately untreated, while the other 12% would have pressures below 95 mm Hg and have had no trial end-point.