Brain corticotropin-releasing factor/urocortin (CRF/Ucn) systems are hypothesized to control feeding, with central administration of 'type 2' urocortins producing delayed anorexia. The present study sought to identify the receptor subtype, brain site, and behavioral mode of action through which Ucn 3 reduces nocturnal food intake in rats. Non-food-deprived male Wistar rats (n=176) were administered Ucn 3 into the lateral (LV) or fourth ventricle, or into the ventromedial or paraventricular nuclei of the hypothalamus (VMN, PVN) or the medial amygdala (MeA), regions in which Ucn 3 is expressed in proximity to CRF(2) receptors. LV Ucn 3 suppressed ingestion during the third-fourth post-injection hours. LV Ucn 3 anorexia was reversed by cotreatment with astressin(2)-B, a selective CRF(2) antagonist and not observed following equimole subcutaneous or fourth ventricle administration. Bilateral intra-VMN and intra-PVN infusion, more potently than LV infusion, reduced the quantity (57-73%) and duration of ingestion (32-68%) during the third-fourth post-infusion hours. LV, intra-PVN and intra-VMN infusion of Ucn 3 slowed the eating rate and reduced intake by prolonging the post-meal interval. Intra-VMN Ucn 3 reduced feeding bout size, and intra-PVN Ucn 3 reduced the regularity of eating from pellet to pellet. Ucn 3 effects were behaviorally specific, because minimal effective anorectic Ucn 3 doses did not alter drinking rate or promote a conditioned taste aversion, and site-specific, because intra-MeA Ucn 3 produced a nibbling pattern of more, but smaller meals without altering total intake. The results implicate the VMN and PVN of the hypothalamus as sites for Ucn 3-CRF(2) control of food intake.