The number of apnoeic episodes and arterial oxygen desaturations were measured preoperatively and for sixty hours postoperatively in twenty ASA status 1-2 patients scheduled for elective gastrectomy. Monitoring included continuous pulse oximetry, respiratory inductive plethysmography and repeated arterial blood gas analysis. The number and magnitude of apnoeas and desaturation episodes were compared between two postoperative analgesic regimens of epidural morphine; intermittent bolus injection (Group B, n = 10), and patient controlled administration with continuous infusion (Group P, n = 10). Morphine dose, P a C02 and mean Sp02 values were similar between the two groups. Although the number of central apnoeas with Sp02 < 90% was greater in Group B, other episodes of apnoea or desaturation were similarly seen preoperatively. In the postoperative period, central apnoeas with spa; < 90% were significantly increased in Group B, while no change was seen in Group P. Apnoeas with 51'02 < 80% were only seen in Group B. We conclude from these results that postoperative apnoeas and episodic desaturations are greatly influenced by the different modes of opioid administration.