In a randomized trial the efficiency of synthetic secretin (0,25 clinical units/kg bw/h; 15 patients) was compared with that of synthetic somatostatin (250 micrograms/h, 15 patients) in treatment of severe acute non arterial bleeding from peptic ulcers or stress lesions in the upper digestive tract. In all patients the presence of hemorrhage at beginning of the therapy was established by endoscopic examination. Both test preparations were given as continuous infusions over 48 h and had identical effects on the hemorrhage: in 12 patients (80%) out of each group bleeding stopped during the infusion of secretin or somatostatin, 6 patients (20%) had to be operated on for persistant bleeding or severe recurrence.