In tests of frequency threshold for brain stimulation-induced feeding, naloxone (s.c.) did not affect the first in a brief series of threshold estimates but elevated subsequent estimates progressively. It was demonstrated that neither the time-course of drug action nor any cumulative disruptive effect of brain stimulation itself, accounts for the progressive elevation of threshold. Self-stimulation in 'feeding' electrodes was therefore studied, in combination with hunger manipulations, to inferentially evaluate naloxone's effect on feeding mechanisms. Results suggest naloxone's anoretic effect does not reflect heightened responsiveness of a satiety mechanism. Reversal by naloxone of the potentiating effect of hunger on self-stimulation, however, suggests the anoretic effect is due to blockade of an opioid process associated with hunger that otherwise enhances the reward value of food.