Muscarinic receptor densities were measured in membranes prepared from rat cerebral cortex using [3H]pirenzepine and [3H]quinuclidinyl benzilate. Isotherms of equilibrium binding data modeled to a single apparent binding site for both ligands. However, as has been reported previously, [3H]pirenzepine labeled only a small fraction of the binding sites that were labeled by [3H]quinuclidinyl benzilate. This observation has been used to support the hypothesis that subtypes of muscarinic receptors exist. Several investigators have previously suggested that antagonist binding to muscarinic receptors involves an isomerization of the receptor-antagonist complex, and it is only the isomerized form of the receptor that is identified by radioligand binding studies. To examine the possibility that the difference in the density of binding sites identified by [3H]pirenzepine and [3H]quinuclidinyl benzilate is due to differences in the degree of isomerization of the receptor associated with the binding of each ligand, the kinetics of the binding of [3H]pirenzepine and [3H]quinuclidinyl benzilate to membranes prepared from rat cerebral cortex were examined. The pseudo-first-order rate constant of association for both ligands showed a nonlinear (hyperbolic) dependence on ligand concentration. These results suggested that a rapidly equilibrating initial binding step was followed by a more slowly equilibrating isomerization of the initially formed ligand-receptor complex. The kinetic data were computer-modeled to obtain estimates of the equilibrium constants for both reaction steps. The equilibrium constants for the isomerization step were 0.1 and 0.004 for [3H]pirenzepine and [3H]quinuclidinyl benzilate, respectively. Our measurements, in agreement with others, suggested that only the fraction of receptors which isomerized were measurable using filtration binding assays. Although essentially all (99.6%) of the [3H]quinuclidinyl benzilate binding sites appeared to isomerize, only 90% of the [3H]pirenzepine binding sites isomerized, and thus only 90% were measured in our assay. It therefore appears that differences in receptor isomerization can partially, but not wholly, account for the differences between [3H]pirenzepine and [3H]quinuclidinyl benzilate binding in rat cerebral cortex.