The effects of different beta-adrenoceptor agonists and antagonists on plasma noradrenaline and potassium concentrations were studied in patients with borderline hypertension. Heart rate, arterial pressure and the heart rate corrected duration of total electromechanical systole (QS2I) were also measured. Infusion of the beta-adrenoceptor agonists isoprenaline (non-selective) and salbutamol (beta2-selective), but not prenalterol (beta1-selective) caused dose-dependent increments in plasma noradrenaline. Isoprenaline and salbutamol decreased plasma potassium dose-dependently. For a given effect on heart rate and QS2I the fall in potassium was less pronounced after prenalterol. The effects of isoprenaline were also studied after the beta-adrenoceptor antagonists propranolol, 320 mg day−1 for 1 week (non-selective) and atenolol, 100 mg day−1 for 1 week (beta−1-selective). The effects of isoprenaline, when infused in equipotent chronotropic doses, on noradrenaline and potassium were not affected by atenolol, whereas they were completely abolished by propranolol. The rise in noradrenaline during beta-adrenoceptor stimulation could be explained by presynaptic facilitation of noradrenaline release. The fall in potassium probably reflects stimulation of Na-K-ATPase dependent transport of potassium into the cell. Both effects were seen after beta2- but not after beta1-adrenoceptor stimulation. Hypokalaemia and raised levels of noradrenaline are also known to occur under such stressful conditions as acute myocardial infarction, when circulating levels of the endogenous beta2-adrenoceptor agonist adrenaline are high. Blockade of these effects by beta-adrenoceptor antagonists may contribute to the cardioprotective effect of these drugs. This warrants further consideration of the clinical significance of beta-blocker selectivity.