1. (-)-Atenolol was used as a tool to assess the function of beta 1- and beta 2-adrenoceptors in human heart. Right atrial and left ventricular preparations from patients undergoing open heart surgery were set up to contract isometrically. Membrane particles were prepared for beta-adrenoceptor labelling with [3H]-(-)-bupranolol and adenylate cyclase assays. 2. The positive inotropic effects of (-)-noradrenaline were antagonized to a similar extent by (-)-atenolol in atrial and ventricular preparations. (-)-Atenolol consistently antagonized the effects of (-)-adrenaline to a lesser extent than those of (-)-noradrenaline in atrial preparations. In ventricular preparations (-)-atenolol antagonized the effects of low concentrations of (-)-adrenaline to a lesser extent than those of high concentrations. 3. pKB values (M) of (-)-atenolol, estimated with non-linear analysis from the blockade of the positive inotropic effects of the catecholamines, were 7.4 for beta 1-adrenoceptors and 6.0 for beta 2-adrenoceptors. 4. (-)-Atenolol inhibited the binding of [3H]-(-)-bupranolol to ventricular beta 1-adrenoceptors with a pKD (M) of 5.9 and to ventricular beta 2-adrenoceptors with a pKD of 4.6. 5. (-)-Atenolol inhibited the catecholamine-induced adenylate cyclase stimulation in the atrium and ventricle with pKB values of 5.8-6.4 for beta 1- and pKB values of 4.7-5.7 for beta 2-adrenoceptors. The binding and cyclase assays suggest a partial affinity loss for (-)-atenolol inherent to membrane preparations. 6. beta 1-Adrenoceptors mediate the maximum positive inotropic effects of (-)-noradrenaline in both the atrium and ventricle of man. beta 2-Adrenoceptors appear to be capable of mediating maximal positive inotropic effects of (-)-adrenaline in atrium. In contrast, ventricular beta 2-adrenoceptors mediated only submaximal effects of (-)-adrenaline.