In view of the recent M1 and M2 subclassification of muscarinic receptors and the suggestion of separate populations of muscarinic receptors on oxyntic and histamine cells in the gastric mucosa, we have analysed the effects of McN-A 343, classified as an M1-selective agonist, on gastric acid secretion by the mouse, isolated, lumen-perfused stomach assay. Acid secretion stimulated by McN-A 343 was not inhibited by tetrodotoxin pretreatment, although it was competitively antagonized by atropine (pKB 7.90), suggesting a muscarinic site of action between postganglionic neurones and the final secretory event. Acid secretion stimulated by McN-A 343 was more sensitive than 5-methylfurmethide-stimulated secretion to H2-receptor blockade: the profile of inhibition was consistent with expectations for a model of indirect agonism, suggesting that McN-A 343 preferentially stimulated the release of endogenous histamine from mucosal histamine cells. In view of this selective action the McN-A 343-pirenzepine interaction was studied, the latter being classified as an M1-selective antagonist. Results were consistent with expectations for a competitive interaction but the pKB (6.69) was not significantly different from the value obtained at the oxyntic cell, using 5-methylfurmethide as agonist in the presence of H2-receptor blockade, in a previous study. We suggest that there is no need to postulate differences in oxyntic and histamine cell muscarinic receptors to account for the selective stimulant activity of McN-A 343 observed in this study and the relatively selective inhibition of gastric acid secretion by pirenzepine in vivo. McN-A 343 selectivity may be accounted for by a higher muscarinic receptor density on the histamine cell and pirenzepine selectivity by a smaller degree of loss into the gastric secretion compared to atropine.