This study was carried out to assess the efficacy of oral ondansetron, a new 5HT3 receptor antagonist, in patients undergoing thyroid surgery. It included 60 patients, randomly assigned to two groups, and receiving orally, 1 h before induction of anaesthesia, either 8 mg of ondansetron (n = 29) or a placebo (n = 30). One patient was excluded. The same anaesthetic protocol, consisting of 3 to 5 micrograms.kg-1 of fentanyl, 4 to 6 mg.kg-1 of thiopentone, and 0.5 mg.kg-1 of atracurium, was used in all. Anaesthesia was maintained with 50% nitrous oxide in oxygen with 0.8 to 1% endtidal concentration of isoflurane and additional boluses of 0.1 mg of fentanyl as required. The incidence and intensity of nausea, graded mild, moderate or severe, and the incidence of vomiting were recorded postoperatively. During the first twelve hours after surgery, 40% of patients in the placebo group had nausea (16.7% mild, 20% moderate and 6.7% severe), and 50% vomited. In the ondansetron group, nausea and vomiting occurred in 13.8% and 20.4% of patients respectively. The 4 patients in the latter group complained of major nausea. The differences between the groups were statistically significant: p = 0.025 for nausea and p = 0.042 for vomiting. It is concluded that oral ondansetron, 8 mg taken orally 1 h before surgery, significantly reduces the incidence of nausea and vomiting during the first twelve postoperative hours. As it is easy to use and has no side-effects, it might be of interest in day-case surgery patients, despite its high cost.