If the demonstrated effects of relaxation training on blood pressure (BP) occur only during relaxation practice, then little effect on the morbidity and mortality of exssential hypertensives would be expected. This question was addressed by inpatient monitoring of five hypertensive patients' BP for 24 hours during six experimental days, including a no-treatment baseline, three days of relaxation training, and one day of recovery. Lowering of both systolic and diastolic pressures persisted beyond the end of the training sessions. Moreover, systolic BP was significantly lower during relaxation training days than during either baseline or recovery days, a difference particularly noticeable at night when patients were sleeping. The BPs of the three patients showing the largest initial effects of training averaged 12.5/7.3 mm Hg less during nights following relaxation sessions than during nights following no treatment.