Patients with duodenal ulcer were classified into responders to H2-receptor antagonist and non-responders in whom duodenal ulcer did not heal within three months with the antagonist. In these patients and healthy controls, modified sham-feeding (MSF) test was performed to elucidate the pathophysiological differences between these groups, especially vagal activity of acid output. After tetragastrin 4 micrograms/kg/hr infusion for one hour, MSF was superimposed to the infusion (G + MSF). The acid output of non-responders (8 cases) and controls (8 cases) significantly increased in G + MSF. On the other hand, the acid output of responders (6 cases) significantly decreased. The acid output of non-responders in G + MSF was also significantly higher than that of controls. The results suggest that vagal activity of non-responders is higher than responders, and that there is disturbance of inhibitory mechanism in vagal system.