Our model of the human m1 muscarinic receptor has been refined on the basis of the recently published projection map of bovine rhodopsin. The refined model has a slightly different helix arrangement, which reveals the presence of an extra hydrophobic pocket located between helices 3, 4 and 5. The interaction of series of agonists and antagonists with the m1 muscarinic receptor has been studied experimentally by site-directed mutagenesis. In order to account for the observed results, three-dimensional models of m1 ligands docked in the target receptor are proposed. Qualitatively, the obtained models are in good agreement with the experimental observations. Agonists and partial agonists have a relatively small size. They can bind to the same region of the receptor using, however, different anchoring receptor residues. Antagonists are usually larger molecules, filling almost completely the same pocket as agonists. They can usually produce much stronger interactions with aromatic residues. Experimental data combined with molecular modelling studies highlight how subtle and diverse receptor-ligand interactions could be.