Ninety unpremedicated patients undergoing mask anaesthesia were assigned to one of three groups according to the volatile anaesthetic and the acute intravenous premedication administered. Group I received saline placebo as premedication and halothane by inhalation. Group II received saline placebo and isoflurane by inhalation. Group III received nalbuphine 0.1 mg.kg-1 IV as premedication and isoflurance by inhalation. Mean time to loss of consciousness (71 sec) did not differ among groups. The dosage of thiopentone required to induce loss of consciousness was decreased by 15 per cent (from 3.9 to 3.3 mg.kg-1) by nalbuphine premedication (P less than 0.05), and time to induction of surgical anaesthesia using isoflurane was decreased by 15 per cent (P less than 0.05). The incidence of reflex actions (coughing, laryngospasm, breath holding, hiccoughs and movement) during induction was no different in the saline-premedicated halothane or isoflurane groups. Acute intravenous nalbuphine premedication decreased significantly the incidence of reflex actions during induction of isoflurane anaesthesia from 77 per cent to 37 per cent (P less than 0.02). Desaturation episodes (SaO2 less than 90 per cent) were more frequent with isoflurane inductions compared with halothane (55 per cent vs 17 per cent, P less than 0.01). Apnoeic episodes accounted for the majority of desaturations associated with nalbuphine premedication, while excitatory reflexes (coughing and laryngospasm) accounted for more desaturations with isoflurane alone.