Three experiments were done to better assess the gastrointestinal (GI) site(s) of action of GLP-1 on food intake in rats. First, near-spontaneous nocturnal chow meal size (MS), intermeal intervals (IMI) length and satiety ratios (SR = MS/IMI) were measured after infusion of saline, 0.025 or 0.5 nmol/kg GLP-1 into the celiac artery (CA, supplying the stomach and upper duodenum), cranial mesenteric artery (CMA, supplying small and all of the large intestine except the rectum), femoral artery (FA, control) or portal vein (PV, control). Second, infusion of 0.5 nmol/kg GLP-1 was tested after pretreatment with the GLP-1 receptor (GLP-1R) antagonist exendin-4(3–39) via the same routes. Third, the regional distribution of GLP-1R in the rat GI tract was determined using rtPCR. CA, CMA and FA GLP-1 reduced first MS relative to saline, with the CMA route more effective than the others. Only CMA GLP-1 prolonged the IMI. None of the infusions affected second MS or later eating. CA and CMA GLP-1 increased the SR, with the CMA route more effective than the CA route. CMA exendin-4 (3–39) infusion reduced the effect of CMA GLP-1. Finally GLP-1R expression was found throughout the GI tract. The results suggest that exogenous GLP-1 acts in multiple GI sites to reduce feeding under our conditions and that GLP-1R in the area supplied by the CMA, i.e., the small and part of the large intestine, plays the leading role.