Exogenous cholecystokinin (CCK) suppresses food intake by acting on vagal sensory neurons. However, CCK doses used in behavioral experiments are generally much larger than those necessary to produce electrophysiological changes in vagal afferents. We made automated measurements of liquid food intake before, during, and after infusion of low doses of CCK octapeptide (CCK-8) through a chronic aortic catheter with its tip seated just above the celiac juncture. In parallel experiments, we made similar infusions while collecting blood from the hepatic portal and jugular veins for CCK assay. Injection of 10, 30, 50, and 70 pmol of CCK-8 suppressed feeding in a dose-dependent manner beginning 1 min postinfusion. The lowest dose to produce statistically significant suppression of preinfusion intake was 30 pmol. Infusion of the same CCK-8 doses into the jugular vein did not suppress feeding. Near-celiac injection of 30 pmol of CCK-8 produced systemic plasma CCK concentrations averaging 6.5 +/- 1 pM compared with less than 1 pM after saline injection. These findings show that exogenous CCK, by acting on tissues perfused by the celiac artery, can suppress feeding at doses that 1) are similar to those producing effects on the firing of vagal neurons and 2) do not increase plasma CCK concentrations above postprandial levels.