BALB/c mice were unilaterally implanted with a guide-cannula, the tip of which was positioned 1 mm above either the lateral hypothalamus (LH) or the mesencephalic central gray area (CG). On each experimental day, a stainless-steel injection cannula was inserted into the LH or the CG and self-administration of two doses of morphine (50 and 5 ng) was compared in the two brain structures using a spatial discrimination task in a Y-maze. At the dose of 50 ng, mice injected into the LH rapidly discriminated the reinforced arm from the neutral arm of the maze in order to self-administer morphine. In contrast, at this same dose, mice of the CG group do not show any regular self-administration behavior. At the dose of 5 ng, both LH and CG injected mice show a regular self-administration response. The rate of discrimination was similar in the two groups. When naloxone (5 ng) was mixed with morphine (5 ng), the number of self-administrations progressively decreased in both brain areas. This decrease was both larger and more rapid in CG than in LH. Marked signs of physical dependence (escapes from the maze) were observed in the two groups during this phase. Finally, when morphine alone (5 ng) was again made available, a regular self-administration response reappeared in the two brain structures. These data suggest (1) that morphine has reinforcing effects in both LH and CG and (2) that in these two brain structures self-injection of this drug is dependent on an opiate receptor mediated mechanism.