We have examined the short-term effects of three volatile anaesthetics, halothane, isoflurane and desflurane, on mucociliary activity in the rabbit maxillary sinus in vivo. Mucociliary activity was recorded photoelectrically and the signal processed by fast Fourier transformation. Administration of 1.0 MAC of halothane, isoflurane or desflurane caused a temporary increase in mucociliary activity, with mean peak responses of 47.8 (SEM 13.0)%, 44.0 (9.6)% and 45.1 (23.7)% (n = 6), respectively. The response to all three compounds was biphasic; an initial peak was observed within 2 min and a second peak at 3-8 min. The second response was not significant for halothane. In contrast, desflurane produced a significant second peak while the first was small and failed to reach significance. Halothane displayed an initial peak within 2 min which was blocked by atropine but not by the neurokinin 1 (NK1) receptor antagonist CP-99. The second peak at 3-5 min was less pronounced for halothane than for isoflurane or desflurane. The second peak was not affected by atropine pretreatment, but was blocked by pretreatment with CP-99. A combination of atropine and CP-99 pretreatment abolished the mucociliary response to halothane. Atropine pretreatment did not affect, whereas CP-99 significantly reduced, the response to desflurane. We conclude that the NK1-mediated response was most pronounced for desflurane which is considered the most airway irritating compound of the three. It is likely that the size of the NK1-mediated response reflects the airway-irritating properties of the volatile anaesthetic used.