1. Single channel currents activated by 250 nM acetylcholine were recorded from cell-attached patches of BC3H1 mouse tumour cells grown in culture. Channels were recorded in the absence and presence of alphaxalone, diazepam, etomidate, fentanyl, ketamine, meperidine, or propofol. 2. All of the anaesthetics tested shortened channel open time but did not alter single channel current amplitude. Drug concentrations calculated to reduce the time constant of open-time distributions by 50% were 99 microM alphaxalone, 66 microM diazepam, 57 microM etomidate, 26 microM fentanyl, 15 microM ketamine, 16 microM meperidine, or 81 microM propofol. 3. Ketamine, meperidine, and propofol reduced channel open time at concentrations comparable to plasma levels attained during therapeutic use of these agents, while alphaxalone, diazepam, etomidate, and fentanyl reduced channel open time only at levels higher than those encountered clinically. 4. The potency of these drugs in decreasing channel open time appears to be directly correlated with their octanol/buffer partition coefficients. In contrast to expectations, however, agents with higher partition coefficients were less potent in altering channel open time. 5. Ketamine and meperidine produced a prominent third component in closed-time distributions, which were otherwise well described by the sum of two exponential components. Alphaxalone, diazepam, and etomidate also produced a small third component, while no additional component was seen with propofol or fentanyl. These additional components probably arise from creation of an additional closed state of the channel. 6. We conclude that these agents are not altering channel properties merely by exerting non-specific effects via the lipid bilayer and that they are probably not all acting by similar mechanisms.