Three groups of horses scheduled for arthroscopic surgery were premedicated with 0.025 mg/kg acepromazine, 88 microg/kg romifidine and 25 mg/kg guaifenesin, and anaesthesia was induced with 2.2 mg/kg ketamine. In group 1 (14 horses), anaesthesia was maintained by halothane vaporised in oxygen, with supplementary doses of 20 microg/kg romifidine and 0.5 mg/kg ketamine every 20 minutes. The 17 horses in group 2 were maintained as for group 1; however, an additional dose of 0.05 mg/kg butorphanol was administered before the first incision. The 14 horses in group 3 were maintained as for group 2, except that the dose of romifidine was halved to 10 microg/kg. The mean (sd) dose of vaporised halothane (6.5 [1.1], 4.8 [1.3] and 4.8 [1.3] microl/kg/minute, for groups 1, 2 and 3, respectively), vaporiser dial setting (2.3 [0.2],1.6 [0.4] and 1.3 [0.1] per cent) and dose of dobutamine administered to correct episodes of hypotension (0.16 [0.08], 0.07 [0.09] and 0.0 [0.0] microg/kg/minute) were all significantly lower in groups 2 and 3 than in group 1. Mean arterial blood pressure was significantly higher in groups 2 and 3 than in group 1 (71 [3], 74 [5] and 80 [5] mmHg, for groups 1, 2 and 3, respectively).