Ninety patients scheduled for general or orthopaedic surgical procedures were randomly assigned to receive one of three i.m. premedications: dixyrazine 0.5 mg kg-1; morphine 0.15 mg kg-1 and scopolamine 0.0065 mg kg-1; or placebo. The premedication was administered and evaluated in a double-blind fashion. The patients were anaesthetized with thiopentone, fentanyl, pancuronium, and ventilated with nitrous oxide in oxygen. The three premedications had no noticeable anxiolytic effect. Although there was no difference in the frequency of observed postoperative nausea and vomiting between the three groups, premedication with dixyrazine nonetheless reduced the patients' experience of postoperative nausea as well as their need for postoperative antiemetics. Although patients in the two treatment groups were significantly more sedated immediately before induction of anaesthesia than patients receiving placebo, the degree of postoperative sedation was similar in all three groups. Morphine-scopolamine caused more postoperative dizziness than dixyrazine and placebo. Lack of recall was produced by both morphine-scopolamine and dixyrazine. It is concluded that premedication with dixyrazine is a useful alternative, especially in patients who have previously experienced postoperative nausea and vomiting.