Whilst developing a rat model for studies of gastric protection, we noticed that the anaesthetic agent 'Hypnorm', containing the opiate fentanyl 0.315 mg/ml and the butyrophenone fluanisone 10 mg/ml, was itself protective against indomethacin-induced damage: unrestrained animals given indomethacin (20 mg/kg) subcutaneously had an ulcer score of 9 +/- 1 mm2, compared with 1 +/- 1 mm2 in animals pre-treated with Hypnorm (0.8 ml/kg) and then given indomethacin (p less than 0.01). Further investigation showed this effect to be due to fentanyl-inhibiting gastric acid secretion: doses of fentanyl (90 and 180 micrograms/kg) which decreased indomethacin-induced damage also caused a rise in intragastric pH from 2.7 +/- 0.6 in controls to 5.1 +/- 0.8 and 5.0 +/- 0.8, respectively. However, the response of fentanyl varied depending on the dose given: fentanyl, 3.6 micrograms/kg did not affect indomethacin-induced damage, 8 +/- 2 vs. 9 +/- 1 mm2; fentanyl, 18 micrograms/kg potentiated damage, 15 +/- 4 mm2 (p less than 0.05), whereas fentanyl, 90 micrograms/kg and 180 micrograms/kg decreased damage, 2 +/- 1 mm2 and 0.1 +/- 0.1 mm2, respectively (p less than 0.01). Neither the butyrophenone haloperidol (8.3 mg/kg) nor the alpha-adrenergic receptor antagonist phenoxybenzamine (3 mg/kg) protected against indomethacin-induced damage. We conclude that fentanyl affects intragastric pH and can both potentiate and protect against indomethacin-induced damage. Furthermore, the potentiation of gastric damage by fentanyl occurred at doses similar to those used for human anesthesia, so clinical studies are suggested.