Adult dogs were anesthetized and segments of small intestine were isolated and perfused via a mesenteric artery with Krebs-bicarbonate solution. Muscle responses along the circular and longitudinal axes were monitored with extraluminal strain gauge transducers. Agonists were administered as intraarterial bolus injections in volumes not exceeding 0.1 ml. Antagonists were dissolved in the Krebs perfusion solution. The C-terminal octapeptide of cholecystokinin (CCK-OP) produced dose-related tonic and phasic contractions of smooth muscle along the transverse axis and tonic and phasic relaxations along the longitudinal axis of the muscle segments. Smooth muscle responses to CCK-OP were reduced by atropine sulfate (0.1 mug/ml) and minimally affected by hexamethonium (10 mug/ml) perfusion. Tetrodotoxin (TTX, 5 ng/ml) selectively reduced muscle responses to dimethylphenylpiperazinium (0.5 mug) and CCK-OP (0.2 mug) but failed to alter responses to bethanechol (BeCh, 5 mug). Higher concentrations of TTX (10 ng/ml) abolished responses to CCK-OP. Depolarizing doses of nicotine (500 mug) selectively antagonized responses to 5-hydroxytryptamine (5HT, 5 mug) and CCK-OP (0.2 mug) but did not alter responses to BeCh (5 mug). Perfusion with tetraethylammonium (1 mg/ml) reduced muscle responses to depolarizing doses of nicotine and abolished the nicotine-induced antagonism of 5HT and CCK-OP. The intestinal smooth muscle response to CCK-OP in the dog is mediated through a neurogenic mechanism. CCK-OP interacts with a nonnicotinic receptor on postganglionic cholinergic neural elements in this preparation.