1. NIE-115 mouse neuroblastoma cells were studied under voltage clamp in the whole-cell patch-clamp configuration. Peak currents induced by bath application of 5-hydroxytryptamine (5-HT) were inwardly rectifying, reversed at 0.4 +/- 0.2 mV (mean +/- s.e.mean), and were approximately half-inhibited (at 1 microM 5-HT) by 2 nM of the 5-HT3 selective antagonist MDL-72222 (3-tropanyl-3,5-dichlorobenzoate). 2. Peak inward currents activated by a low concentration of 5-HT at a holding potential of -50 mV were potentiated by volatile general anaesthetics. At their human minimum alveolar concentrations (MACs), the degree of potentiation increased in the order isoflurane < halothane < enflurane < methoxyflurane. Potentiation by methoxyflurane was independent of membrane potential in the range -70 mV to +40 mV. The reversal potential was the same in the presence and absence of methoxyflurane. 3. Methoxyflurane shifted the 5-HT dose-response curve to lower 5-HT concentrations, without significantly changing the Hill coefficient or maximum response. The EC50 concentration for 5-HT decreased from 1.86 +/- 0.02 microM to 1.07 +/- 0.11 microM (means +/- s.e.mean) due to the presence of 1 MAC (270 microM) methoxyflurane. 4. In contrast to the volatile anaesthetics, the barbiturate anaesthetic, thiopentone, inhibited the 5-HT3 receptor. Hill analysis of thiopentone dose-response data gave an average IC50 = 117 +/- 8 microM thiopentone and Hill coefficient = 1.6 +/- 0.2 (means +/- s.e.mean). These parameters were not significantly different for data obtained at 5-HT concentrations above and below the control EC50 concentration for 5-HT, consistent with non-competitive inhibition. 5. The n-alcohols occupied an intermediate position between the volatile and barbiturate anaesthetics. The lower alcohols (butanol and hexanol) potentiated 5-HT responses at low alcohol concentrations but inhibited them at high concentrations. In contrast, the higher alcohols (octanol, decanol, dodecanol, tridecanol, tetradecanol and pentadecanol) produced no potentiation, but only inhibition, at all alcohol concentrations. 6. Inhibition of the 5-HT3 receptor by the n-alcohols exhibited a cutoff in potency similar to those previously found for tadpoles, luciferase enzymes and a neuronal nicotinic acetylcholine receptor channel.