Physical training offers a potential nonpharmacological strategy for control of mild and borderline hypertension, but its effect on blood pressure is controversial. We investigated the effects of endurance training on waking and sleeping blood pressure and on baroreflex sensitivity in 16 borderline hypertensive patients. First, 8 patients were assessed before and after a 6-month endurance training programme. Then, when it was clear that blood pressures were lower after training, a further 8 patients were studied not only at the end of the training programme but also after 4 months' abstention from exercise (detraining). Measurements were taken of baroreflex sensitivity (response to iv phenylephrine), blood pressure, R-R interval, and blood pressure and R-R variability. Ambulatory blood pressures were measured in 13 patients (7 trained, 6 detrained) and sleep blood pressures in 6 patients (3 trained, 3 detrained). Increased fitness was associated with a decline in resting arterial blood pressure of 9.7 (SE 2.0) mm Hg systolic and 6.8 (1.2) mm Hg diastolic, and with a decline in ambulatory blood pressure of 4.8 (1.4) mm Hg and 7.5 (2.1) mm Hg, respectively; both p less than 0.05. Baroreflex sensitivity was 14.0 (1.8) ms/mm Hg in the unfit and 17.5 (2.0) ms/mm Hg in the fit; p less than 0.05. Sleep blood pressures were not lower in the fit despite longer sleep R-R intervals. These findings indicate that, in some subjects with borderline or mild hypertension, a physical training programme is sufficient to bring the blood pressure within normal limits.