At least three pharmacologically different muscarinic receptor subtypes (M1, M2 and M3) have been identified in rat brain. While M1 and M2 subtypes can be directly labelled by selective ligands (3H-pirenzepine and 3H-AF-DX 116, respectively), there are no truly selective ligands for the M3 subtype. In the present study, we have investigated a possible method of studying the pharmacological M3 subtype in rat forebrain using the non-selective labelled antagonist 3H-N-methyl-scopolamine (3H-NMS) in the presence of unlabelled pirenzepine to protect the M1 subtype. The results obtained in kinetic experiments using 3H-NMS in presence of 30.10(-9) M unlabelled pirenzepine (Kon 1.2.10(-8) M-1 m-1, Koff 4.7.10(-2) m-1 and Kd 0.4.10(-9) M) are compatible with the studies carried out in rat pancreatic islets and submaxillary gland which contain predominantly the M3 subtype. We have also performed inhibition experiments with the selective antagonist AF-DX 116. Due to the small proportion of M2 receptors present in rat forebrain, this drug is able to discriminate between M1 and non M1 non M2 receptor subtypes in competition experiments with 3H-NMS versus AF-DX 116 (Ki values 0.28.10(-6) M and 4.3.10(-6) M, respectively). When the competition experiments were performed using 3H-NMS in presence of 30.10(-9) M unlabelled pirenzepine, the Ki value obtained was 3.8.10(-6) M, very close to the value obtained for the non M1 non M2 receptor in competition experiments with 3H-NMS versus AF-DX 116 and in excellent agreement with the affinity of this drug for the glandular M3 subtype. All these data suggest that the approach using the non-selective antagonist 3H-N-methyl-scopolamine in presence of unlabelled pirenzepine allows the study of the pharmacological M3 subtype in rat forebrain.