Thirty mild hypertensives were treated for more than two months with either cardioselective (atenolol or metoprolol) or non-selective (propranolol or pindolol) beta-blockers; the patients were assigned to the drugs in a double-blind manner. A procedure was designed to distinguish between the effects of the drugs themselves while treatment continued, and the development of adaptive changes which would persist when the drugs had been eliminated from the body. Though individual responses to treatment varied in both groups, the mean effect of the cardioselective and non-selective drugs in the control of hypertension was similar. There was no evidence of the development of supersensitivity or "rebound". On the contrary, an adaptive bradycardia (that is a fall of not less than 10% in heart rate persisting 52 hours after stopping treatment) was observed at rest in 17/30 patients, and peak heart rates and blood pressures during exercise were lower in both groups than before treatment. Cardioselective drugs induced a significantly greater bradycardia at rest than non-selective, but on exercise increases in heart rate were reduced more by the non-selective drugs, so that the same peak heart rates were reached on exercise in both groups. Adaptation also affected QT. The results suggest that two factors govern the shortening of QT by increases in heart rate, a "metabolic" effect, determined by sympathetic drive, and a "biophysical" effect determined by heart rate. The adrenergic effect is attenuated by acute beta-blockade, or by adaptation to prolonged blockade, leaving a shallow, rate-determined, slope to the QT/RR regression.