The results of several experiments indicated that the hyperthermic effect of morphine in rats becomes attenuated over the course of successive administrations by a conditional, compensatory, hypothermic response elicited by cues present at the time of morphine administration, thus accounting for hyperthermic tolerance: (a) Rats with a history of morphine administration display a tolerant response to the hyperthermic effect of the drug and a compensatory hypothermia following a placebo if these substances are administered following cues that previously signaled morphine--neither the tolerant reaction to morphine nor the hypothermic response to the placebo results when animals are injected following cues that previously signaled injection of physiological saline (Experiments 1A and 1B); (b) presenting environmental cues previously associated with morphine, but without the drug, abolished established tolerance, that is, pyretic tolerance can be extinguished (Experiment 2); (c) placebo sessions interspersed between morphine sessions impeded the acquisition of tolerance, that is pyretic tolerance is retarded by partial reinforcement (Experiment 3). These findings, implicating a Pavlovian conditioning process in hyperthermic tolerance, are not readily interpretable by tolerance models that do not attribute any role to drug-associated environmental cues in the acquisition of tolerance.