Upper gastrointestinal endoscopy was performed on 342 out-patients. The patients were allocated to four groups according to premedication given. The premedication used was atropine 0.1 mg/10 kg (A), atropine 0.1 mg/10 kg and diazepam 5 mg (AD), atropine 0.1 mg/10 kg and fentanyl 0.2 mg (AF), atropine 0.1 mg/10 kg diazepam 5 mg and fentanyl 0.2 mg (ADF). Premedication was given about 30 minutes before the procedure intramuscularly; the mouth and pharynx were sprayed with 10% lidoc,ine. The patients as well as the endoscopist considered the premedication in groups A and AD to be satisfactory. From the patient's point of view there were hardly any differences between the var;ous groups, except in drowsiness, which occurred more often in groups where fentanyl had been used. From the endoscopist's point of view groups A and AD were preferred because they offered better working conditions for the procedure. The longest period of observation was required in group ADF. On the basis of these results premedication with only atropine or if desired a combination of atropine and diazepam in addition to local sprayed anaesthesia of the mouth and pharynx is sufficient, and is recommended for endoscopy of the upper gastrointestinal tract.