1. The objective was to determine the role of muscarinic receptor subtypes in mediating contraction of the human detrusor smooth muscle in vitro. 2. Contractile responses of human detrusor muscle strips to carbachol were obtained in the absence and presence of a range of muscarinic antagonists (pirenzepine, methoctramine, 4-diphenylacetoxy-N-methyl piperidine methiodide (4-DAMP)4, tropicamide, oxybutynin and tolterodine). Affinity estimates (pKB values) were calculated for the antagonists and correlated with values at the cloned muscarinic receptor subtypes quoted in the literature. 3. Pirenzepine, methoctramine and tropicamide drugs that have high affinities at M1, M2 and M4-receptors, respectively, all had low affinities on the human detrusor (pKB values of 6.8, 6.9 and 6.5, respectively), whilst the M3-selective antagonist 4-DAMP had a high affinity (9.5). Schild plots for all four antagonists had slopes of unity indicating an action at a single receptor. Oxybutynin and tolterodine also acted as competitive antagonists with affinity estimates of 7.6 and 8.1, respectively. 4. When the antagonist affinities obtained on the bladder were plotted against the values published for these antagonists at the cloned muscarinic receptor subtypes, the best correlations were obtained for the m3- and m5-muscarinic receptor subtypes. 5. These data suggest that direct contractile responses of the human detrusor muscle to muscarinic receptor stimulation in vitro are mediated solely via the M3-muscarinic receptor subtype with no contribution from the major M2-receptor population.