A series of N-(4-amino-2-butynyl)-5-methyl-2-pyrrolidones modified only in the amino group was synthesized. The compounds were agonists, partial agonists, and antagonists on the isolated guinea pig ileum. They had greater affinity and lower intrinsic efficacy at ileal muscarinic receptors than the identically modified N-(4-amino-2-butynyl)-2-pyrrolidones and N-(4-amino-2-butynyl)succinimides. Dissociation constants in the three series were correlated, suggesting that the compounds had similar mode of binding to muscarinic receptors. The 5-methyl-2-pyrrolidones were 10- to 20-fold less potent as muscarinic agonists on the guinea pig urinary bladder than on the ileum and also elicited lower relative maximal responses on the bladder. For example, the trimethylammonium (9) and azetidino (10) analogues were equipotent (EC50 = 0.2 microM) with the selective muscarinic stimulant N-(1-methyl-4-pyrrolidino-2-butynyl)-N-methylacetamide, BM 5 (2), as agonists on the ileum, but on the bladder 9 and 10 were relatively weak partial agonists, whereas 2 was an antagonist. Compound 10, like 2 and the dimethylamino analogue 8, also differentiated between centrally mediated muscarinic effects in vivo as it was potent in producing analgesia and hypothermia but did not elicit tremor. Instead, 10 antagonized oxotremorine-induced tremor. Thus, 10 resembled 2 in its actions except that the greater intrinsic efficacy of 10 shifted the balance between agonist and antagonist properties slightly toward agonism. Manipulation of intrinsic efficacy by minor changes in chemical structure is emphasized as a means of attaining selectivity.