In the normal heart the ratio of β1/β2-receptors in both atria and ventricles is about 75:25; in the failing heart the ratio is about 60:40. Stimulation of either β1- or β2-receptors results in a positive chronotropic and inotropic response. In the periphery, with the exception of lipolysis, renin release, control of intraocular pressure and intestinal relaxation, β2-related activity predominates. The nature of the β2-receptor is being unravelled and it has now been cloned. The gb-receptor antagonist is ‘anchored’ via disulfide bonding. Subsequent events involve the regulatory protein guanine nucleotide which couples the receptor to adenylate cyclase. β-receptor density may by up- or down-regulated. β-stimulation down-regulates (uncouples and internalizes or sequestrates) and β-antagonism up-regulates β-receptor numbers, but the functional implications of such changes are not always clear. A partial agonist occupies a receptor site and competitively inhibits the full agonist (e.g. noradrenaline). A partial agonist differs from a full agonist in that maximal response of a tissue is less. When background sympathetic activity is absent or very low a partial agonist will act as an agonist, e.g. increase heart rate, but when background tone is high the partial agonist will behave functionally as an antagonist, e.g. decrease heart rate. In animals partial agonist activity (PAA) can be assessed in many ways. In the catecholamine-depleted (reserpine or syrosingopine), vagotomized or pithed, intact animal β-activity can be assessed via changes in heart rate, cardiac contractility and atrioventricular conduction. Isolated organs can also be used such as atria, papillary muscle, tracheal, mesenteric artery and uterine preparations. The choice of animals is important as marked species differences in response can occur. In man assessing PAA is difficult due to the presence of an intact sympathetic system: the problem can be overcome by autonomic blockade of constrictor and vagal reflexes with prazosin, clonidine and atropine but leaving the β-receptor mediated responses unimpaired. β1 and β2-selective PAA can also be gauged via an increased sleeping heart rate (basal sympathetic tone) in the presence and absence of a β1 and β2-selective antagonist. β1-selective PAA can also cause an increase in resting systolic blood pressure. β2-selective PAA may be further assessed by a fall in DBP, increased blood flow, fall in peripheral resistance or increased finger tremor. Isolated human tissue such as lymphocytes (β2) or myocardial biopsy material, (β1 and β2) can be used to assess down-regulation of β-receptor density as a measure of PAA. Expression of PAA depends upon receptor affinity and tissue response (β-receptors are coupled to the effector enzyme via a guanine nucleotide regulatory protein) and the efficiency of coupling determines whether or not the drug acts as an antagonist (zero coupling), partial agonist or full agonist. For any one agent this may vary from tissue to tissue and from species to species. In man, β1-selective PAA, depending on its degree, will diminish or abolish antihypertensive efficacyat rest; in contrast, β2-selective PAA is associated with a fall in resting blood pressure due to peripheral vasodilation. The cardiovascular pharmacodynamic and clinical implications of PAA are discussed, as exemplified by a nonselective agent with nonselective PAA (pindolol); a nonselective agent with primarily β2-selective PAA (dilevalol); a β1-selective agent with low/moderate PAA (epanolol); and a β1-selective agent with moderate/high PAA (xamoterol). Also discussed are the effects of the above four types of PAA on various adverse reactions, i.e. fatigue, cold peripheries, heart failure, hypotension, bronchoconstriction, CNS side effects, tremor, rebound phenomena and plasma lipid changes.