1 A range of opioid receptor agonists were tested for activity in five antinociceptive models: the acetylcholine-induced abdominal constriction, tail-flick and hot plate tests in the mouse and the paw pressure test in the rat and guinea-pig. 2 Agonists acting preferentially at the kappa-opioid receptor were significantly more potent in the guinea-pig than in the rat paw pressure test, whereas mu-receptor preferring agonists were equipotent in the two tests. The mouse abdominal constriction test was of equal sensitivity to the guinea-pig pressure test for both types of agonist. 3 The mouse tail-flick and hot plate tests were progressively less sensitive than the other three tests, particularly to kappa-receptor preferring agonists. 4 The efficacy of an agonist can also markedly affect its activity in antinociceptive tests. Thus, partial kappa-agonists were weak or inactive in the rat paw pressure test, and partial agonists at both mu- and kappa-opioid receptors were relatively weak in the tests in which heat was the noxious stimulus, particularly the mouse hot plate test. 5 The mouse abdominal constriction test is suggested as the most appropriate antinociceptive model for testing a broad range of opioid agonists, whilst the relative potency of a drug in the rat and guinea-pig paw pressure tests may indicate the degree to which it is selective for kappa-opioid receptors in vivo.