The results are reviewed which were obtained with epidural administration of fentanyl for the relief of pain in 134 patients who had undergone abdominal surgery or had sustained multiple fractures of the ribs. Single doses of 0.1 mg of fentanyl were given epidurally via an indwelling catheter inserted between L1/L2 or L2/L3 or, in the case of 7 patients with multiple rib fractures, between Th8/Th9. The results were satisfactory to excellent. An increase in the volume of the fentanyl solution from 10 ml to 20 ml increased the number of patients who were free from pain from 61 per cent to 81 per cent. The analgesic effects of individual doses lasted 2-8 hours (average 5.5 h). Persons with multiple rib fractures needed further injections on average every 3.5 hours. In these cases epidural fentanyl administration also proved valuable in combination with artificial ventilation. There were no serious side-effects. A comparison of the clinical efficacy of fentanyl proved that the same doses of the drug were more effective if given epidurally than if injected intramuscularly. The serum concentrations of fentanyl showed considerable individual variations after both intramuscular and epidural administration with peak values varying correspondingly. Epidural injections caused a steeper rise in concentration than did intramuscular administration. Although none of the patients developed respiratory depression the possibility of respiratory failure should be taken into account when employing epidural fentanyl analgesia. The method is very effective in intensive therapy for relieving postoperative or post-traumatic pain, but is unsuitable for the relief of chronic pain because the analgesic effect is comparatively short-lived.