Animals made ill by intraperitoneal injection with toxins, such as lithium chloride (LiCl) or lipopolysaccharides (LPS), or presented with cues associated with LiCl become hyperalgesic [Pain 56 (1994) 227]. The descending pronociceptive neurocircuitry and spinal pharmacology that underlie these effects bear the same features as those that mediate analgesic tolerance to morphine [Neurosci. Biobehav. Rev. 23 (1999) 1059]. Thus, we examined whether LiCl, LPS or cues paired with LiCl could reduce morphine analgesia. Morphine analgesia in the tail flick test was reduced 24 h but not 7 days following injection with LiCl, and 24 h following injection with LPS. In addition, morphine analgesia was reduced in the hot plate test 40 min and 24 h following LiCl. Furthermore, these effects occurred in the absence of detectable hyperalgesia indicating that illness-induced tolerance was not the result of an increase in pain sensitivity offsetting analgesia. Finally, rats tested in a context associated with LiCl demonstrated less morphine analgesia than rats tested in a context not associated with LiCl or rats naive to LiCl suggesting that illness activates descending mechanisms that antagonize analgesia rather than simply desensitizing opioid receptors. Thus, in addition to provoking hyperalgesia, illness-inducing agents also activate endogenous antianalgesic mechanisms.