An expanded version of the mobile receptor model has been assessed in studies on the binding of N-[3H]methylscopolamine and [35S]GTPgammaS to cardiac muscarinic receptors and their attendant G proteins in ventricular membranes from hamster. The model comprises two pools of receptor, one of which lacks G proteins, and a heterogeneous population of G proteins that compete for the receptor within the G protein-containing pool. To guide the formulation of the model itself and to define the various parameters, data were combined from assays performed under various conditions with native membranes and following irreversible blockade of about 80% of the receptors with propylbenzilylcholine mustard. Multiple G proteins are indicated primarily by multiple states of affinity evident in the dose-dependent effect of guanyl nucleotides on the binding of carbachol; G protein-free receptors are indicated by sites of low affinity for carbachol that survive treatment with the mustard. The expanded model generally succeeds where more frugal schemes have been inadequate, but it nevertheless fails to yield a mechanistically consistent description of the data. Guanyl nucleotides and partial alkylation do not affect the inhibitory potency of carbachol in a manner consistent with their supposed effect on the equilibrium between uncoupled and G protein-coupled receptors. As inferred from the model, G proteins are lost upon alkylation of the receptor, and their numbers are regulated by guanyl nucleotides. Parameters estimated via N-[3H]methylscopolamine are wholly inconsistent with the same parameters estimated via [35S]GTPgammaS. The failure of the model suggests that multiple states of affinity may not arise from a ligand-regulated equilibrium between free receptors and G proteins on the one hand and one or more RG complexes on the other.