Feeding and blood glucose responses to local injection of nanoliter volumes of 5-thio-D-glucose (5TG), a potent antimetabolic glucose analogue, were studied at 142 hindbrain and 61 hypothalamic cannula sites. A site was considered positive if 5TG elicited at least 1.5 g more food intake or a hyperglycemic response at least 25 mg/dl greater than the respective responses elicited by vehicle injection in the same rat. Of 61 hypothalamic cannula sites tested, none were positive for blood glucose and only one was positive for feeding. Increasing the 5TG dose to 48 ug did not produce additional positive results at hypothalamic sites. In contrast, 66 hindbrain sites were positive for feeding and 49 were positive for blood glucose, with 33 of these being positive for both responses. The distribution of positive sites for feeding and hyperglycemia overlapped almost completely. Positive sites were concentrated in two distinct zones: one in the ventrolateral and one in the dorsomedial medulla. In both locations, the glucoreceptive areas extended approximately from the level of the area postrema (AP) to the pontomedullary junction. Glucoreceptive zones were co-distributed with epinephrine cell groups C1-C3, suggesting that epinephrine neurons may be important components of the neural circuitry for glucoregulation. Localization of glucoreceptive sites will facilitate positive identification of glucoreceptor cells and the direct analysis of the neural mechanisms through which they influence food intake and metabolic responses.