We measured compliance with beta-blocking drugs in a group of 105 patients with uncomplicated hypertension, and without coronary heart disease or other major concomitant disease. We measured compliance with an electronic monitor (MEMS®) that records time and date of each opening of the container in which the patient's regularly prescribed beta blocker was dispensed. This measurement was compared with their physicians' estimates of their compliance with the prescribed beta blocker regimen. Patients were followed during a period of 2–6 months, for a total of 15,274 days of observation. In 820 (5.4 per cent) of observation days dosing was incorrect; in 619 days, one or more tablets were omitted, and in 201 days one or more extra doses were taken. In 64 of the 105 patients, we observed drug holidays: days without evidence of beta blocker dosing. Concordance was low between the physicians' estimates and MEMS-measured compliance, with a tendency toward physician overestimation of compliance. The study enrolled only patients long treated for uncomplicated disease, with no other serious concomitant illness; this special selection probably accounts for the relatively good compliance found.