The antinociceptive actions of several beta-adrenoceptor agonist drugs have been studied in mice by use of a modified abdominal constriction test. All the drugs studied had high antinociceptive activity, with ID50 values in the nmol kg-1 range. (-)-Isoprenaline and (+/-)-isoxsuprine were the most potent, being about ten times more active than salbutamol, the least potent drug studied. All these drugs produced their action very rapidly and appear to act within the peritoneum. (-)-Isoprenaline had about six times the potency of the (+)-isomer. (+/-)-Propranolol caused rightward shifts, usually parallel, of the dose-response curves for (-)-isoprenaline. (+)-Propranolol was more than ten times less potent than the racemic drug. Practolol also caused parallel, rightward shifts of the dose-response curves for (-)-isoprenaline, and was about twice as potent as (+/-)-propranolol, whether given by subcutaneous or intraperitoneal injection. Atenolol and ICI 118551 had intermediate potencies. Propranolol, practolol and ICI 118551 were all considerably less potent in antagonizing the antinociceptive actions of fenoterol and RO363, than (-)-isoprenaline. None of these antagonist drugs showed more than a slight ability to discriminate between the beta 1- and beta 2-selective agonist drugs. No evidence was found for the involvement of opioid, dopamine, or alpha-adrenoceptors in the antinociceptive action of the beta-adrenoceptor agonist drugs. Evidence for and against the involvement of beta-adrenoceptors is discussed, and it is concluded that if these receptors do mediate the antinociceptive action they appear to be atypical.