Naloxone-injected (1.0 mg/kg) and saline-injected control rats were subjected to a two-trial test of object localization memory. In trial I rats were allowed to explore for 5 min an enclosed T-maze with an object (plastic bottle) placed in one maze arm. Then, the object was removed and after a 20-min retention interval rats were faced with two empty arms of the same maze (trial II). Control rats showed good retention of the place occupied by the object, displaying a significant preference (74%, P < 0.032) for the arm which previously contained the object. Naloxone treated rats responded at chance levels (53%). An accurate performance in this task is normally based on information provided by spatial cues outside the maze (cognitive map), so that a random performance of Naloxone treated rats could supposedly be related to some disorders in the internal representation of the environment.