To study the peripheral effect of somatostatin in dogs a tiny Silastic catheter was implanted in an almost terminal branch of the mesenteric artery, providing blood supply to a jejunal segment of 5-10 cm. In 4 dogs one catheter was implanted, perfusing a small bowel segment about 60 cm below the angle of Treitz; in 3 dogs two catheters were implanted, perfusing two small bowel segments about 30 and 60 cm below the angle of Treitz. The catheter was kept open by continuous perfusion with a diluted heparin solution by means of a portable insulin infusion pump fixed to the animal's protection jacket. Small bowel motility was recorded electromyographically. Experiments were started 2 wk after surgery and were performed in conscious fasted animals. Control experiments using intraarterial saline perfusion showed normal migrating motor complexes in the perfused segment. Intraarterial perfusion of somatostatin (50-200 ng/kg X h) induced ectopic activity fronts, which always started just distal to the perfused segment and progressed down the small bowel at a normal propagation velocity. During the occurrence of ectopic fronts no activity fronts were observed in the small bowel proximal to the perfusion site. Apart from ectopic fronts, normal activity fronts, starting in the duodenum and passing the perfused segment, were also observed. The most proximal part of the jejunum was more sensitive to intraarterial somatostatin than the more distal jejunum. Intravenous infusion of the same doses of somatostatin had no effect. Somatostatin, therefore, seems to induce ectopic fronts via a local mechanism. We propose that somatostatin is able to relieve the intestine locally from an inhibitory mechanism that prevents the development of activity fronts.