The incidence of postoperative respiratory apnoea was compared between five patients receiving a continuous i.v. infusion of morphine (mean 73.6 mg) and five patients receiving a continuous extradural infusion of 0.25% bupivacaine (mean 192 mg) in the 24-h period following upper abdominal surgery. Monitoring consisted of airflow detection by a carbon dioxide analyser, chest wall movement detected by pneumatic capsules, and continuous electrocardiograph recorded with a Holter ambulatory monitor. Both obstructive (P less than 0.05) and central apnoea (P less than 0.05) occurred more frequently in patients who had a morphine infusion. There was also a higher incidence of tachyarrhythmias (P less than 0.05) and ventricular ectopic beats (P less than 0.05) in the morphine infusion group.