Aromatic solvents, such as toluene, can cause depression of the central nervous system functions in both solvent-exposed workers and abusers. The mechanism by which toluene produces its effects is generally thought to be similar to that produced by general anaesthetics, including inert gases and alcohols. However, whether lipophilic compounds indirectly influence activity by perturbing membrane lipids or bind directly to proteins remains a major question. In a recent study, the sigmoidal admission rate-dependence of inert gas anaesthetic potency has been suggested to possibly reflect a direct narcotic-protein interaction. Therefore, experiments have been carried out using seven input toluene flows of 0.5, 1, 2, 3, 4, 5 and 6 l/min. Our results indicate that as the rate of toluene delivery increased, the concentration of toluene required to produce anaesthetic effects increased. Although this was fitted relatively well with linear regression, this fitted better when using a sigmoidal model (r = 0.998 vs. r = 0.971, P < 0.01). In addition, comparison with previous data on nitrous oxide shows a striking similarity between plots (r = 0.991) which appears consistent with a similar site of action for both agents. We suggest that all classes of lipophilic agents could produce their inhibitory effects at similar 'non-specific' sites of action of finite size and limited occupancy.