These experiments were designed to identify brain sites at which opioids might act to influence ingestive behavior and to determine which opioid receptor types are involved. After food deprivation, rats were given microinjections of naloxone into several brain regions and food intake was measured. Injections into or near the paraventricular (PVN) or ventromedial (VMH) hypothalamic nuclei or the globus pallidus (GP) reduced food intake; injections into the striatum or lateral hypothalamus (LH) were ineffective. A second study examined the ingestive effects of roughly equimolar doses (1.43-1.75 nmol) of dynorphin A (DYN), beta-endorphin (beta-END), and D-Ala2,D-Leu5-enkephalin (DADLE) when injected into 4 different brain regions. Only DYN significantly increased food intake, and this effect was seen only with injections into the PVN and VMH. Beta-END stimulated water intake when injected into the PVN, VMH and GP but not the LH. Further studies indicated that with PVN injections, DYN was effective at a dose as low as 0.47 nmol, and that a higher dose of DADLE (4.39 nmol) did stimulate food intake. These studies support an important role for dynorphin and the kappa opioid receptor in the regulation of feeding and suggest that the opioid regulation of food and water intake can be differentiated both by sites of action and by effective agonists.