1. The effects of 5-hydroxytryptamine (5-HT) were investigated on right atrial appendages obtained from patients treated with beta-adrenoceptor blocking agents who were undergoing open heart surgery. Atrial strips were paced under isometric conditions. 2. 5-HT increased contractile force to approximately one half of the force produced by a saturating concentration of (-)-isoprenaline. Both 5-HT and (-)-isoprenaline accelerated the onset of relaxation, as indicated by an abbreviation of time to peak force. 3. The effects of 5-HT were resistant to blockade by 0.4 microM (+/-)-propranolol, 1 microM (-)-pindolol, 0.4 microM methiothepin, 4 microM yohimbine, 0.4 microM ketanserin, 10 microM phenoxybenzamine, 1 microM methysergide, 2 microM MDL 72222 and 20 microM granisetron. 4. Cocaine 6 microM potentiated the effects of 5-HT, increasing the pEC50 from 6.6 to 7.4. The inotropic potency of 5-HT is five times greater than that of (-)-noradrenaline. 5. ICS 205930 antagonized competitively the effects of 5-HT with a pKB of 6.7. 6. In the presence of 0.4 microM (+/-)-propranolol, 10 microM 5-HT increased both adenosine 3':5' cyclic-monophosphate (cyclic AMP) levels and cyclic AMP-dependent protein kinase activity by approximately one half and two thirds respectively, of the corresponding effects of 200 microM (-)-isoprenaline. 7. Both the increase in cyclic AMP levels and the stimulation of protein kinase activity are consistent with the inotropic effects of 5-HT being mediated by cyclic AMP-dependent phosphorylation of Ca2+ channels and of proteins involved in contraction and relaxation. 8. The human atrial 5-HT receptor resembles the neuronal 'so called' 5-HT4 receptor of rodents both in increasing cyclic AMP levels and in its affinity for ICS 205930.