To determine whether tolerance to morphine-induced anorexia involves associative mechanisms, rats were given chronic injections of morphine (Group 1, 10 mg/kg; Group 2, 20 mg/kg) in the presence of one compound cue on alternate days and injections of saline in the presence of another compound cue on the intervening days. After tolerance developed to the initial suppression of intake, three tests of Pavlovian conditioning were conducted. On the compensatory response test, in which saline injections were given in the presence of the morphine cue, only Group 2 showed a significant increase in milk intake. On the explicit unpairing test and the environmental specificity test, in which morphine injections were given in the presence of the saline cue or in an entirely different room, respectively, neither group showed a significant loss of tolerance. The failure to demonstrate cue-dependent tolerance in this paradigm may have been due in part to inadvertent temporal conditioning and in part to the rapid development of nonassociative tolerance.