G protein-coupled receptors display a dispersion of affinities in the binding of agonists, the breadth of which is a measure of efficacy. The implied heterogeneity commonly is thought to be induced by the G protein in an otherwise homogeneous population of receptors. In some early studies, however, purified M2 muscarinic receptor also revealed heterogeneity under some conditions. We have shown that M2 receptors extracted from Sf9 cells can be purified as monomers devoid of G protein and reconstituted almost exclusively as tetramers in phospholipid vesicles (POPC/POPS/cholesterol), as indicated by cross-linking. Monomers in solution appeared homogeneous, as indicated by Hill coefficients near 1 for the inhibitory effect of agonists on binding of the antagonist N-[3H]methylscopolamine; in contrast, reconstituted tetramers appeared heterogeneous (nH < 1) in a manner that correlated with efficacy as reported for GTPase activity (p = 0.013). It follows that efficacy is intrinsic to the receptor in its tetrameric state. To examine whether heterogeneity arises from pre-existing asymmetry within the tetramer or is induced through cooperativity between ligands, the inhibitory effect of two agonists (oxotremorine-M, arecoline) and two antagonists (N-methylscopolamine, atropine) was measured at two concentrations of N-[3H]methylscopolamine. The data were analyzed simultaneously in terms of four dissimilar and non-interacting sites (i.e., asymmetry) and four interacting sites (i.e., cooperativity). The radioligand was assumed to label only two of the four sites, because the capacity for N-[3H]methylscopolamine doubled upon dissolution of reconstituted receptors in digitonin; one-half of the reconstituted sites therefore were inaccessible to the radioligand. Whereas the data could not be described in terms intrinsic asymmetry alone, the fit improved markedly upon the introduction of cooperative interactions (p < 0.00001). Efficacy therefore appears to arise at least in part from cooperativity within the tetramer.