Recent research indicates that the sensitization that results from repeated drug or non-drug stress exposure may develop into a pattern of alternating increases and decreases (i.e., oscillation) in response to each subsequent stressor exposure. Oscillation, with or without prior sensitization, has been observed for a number of drug and non-drug stressors, and for various neurochemical and endocrine endpoints. The present studies investigated whether oscillation also occurs in the behavioral and endocrine effects of repeated morphine treatment and if a drug that normalizes the mood swings of bipolar disorder in humans will also attenuate drug oscillation in this animal model. In the first experiment, rats were given 1-5 pretreatments with morphine (15 mg/kg, i.p.), separated by 1-week intervals with the last injection occurring 1 hour prior to being tested for stressor-induced (i.e., 5 seconds, 2-mA electric footshock) hypoalgesia, as measured by latency to paw-lift or jump from a hot-plate. Plasma beta-endorphin also was measured. The second experiment replicated the behavioral findings of the first study and, in addition, assessed the effect of continuous lithium chloride, in the drinking water, on morphine-induced oscillation. Caffeine was used as a partial control for the lithium. The results were that one injection of morphine enhanced stress-induced hypoalgesia and subsequent morphine administrations resulted in oscillation. Beta-endorphin exhibited sensitization but not oscillation, suggesting that it did not mediate oscillation of the behavioral response. In addition, lithium, but not caffeine, eliminated oscillations of the behavioral response without affecting its initial enhancement.