Nerve ligation injury in rats results in reduced nociceptive and non-nociceptive thresholds, similar to some aspects of clinical conditions of neuropathic pain. Since underlying mechanisms of hyperalgesia and allodynia may differ, the present study investigated the pharmacology of morphine and MK-801 in rats subjected to a tight ligation of the L5 and L6 nerve roots or to a sham-operation procedure. Response to acute nociception was measured by (a) withdrawal of a hindpaw from a radiant heat source, (b) withdrawal of the tail from a radiant heat source or (c) the latency to a rapid flick of the tail following immersion in water at different noxious temperatures. Mechanical thresholds were determined by measuring response threshold to probing the hindpaw with von Frey filaments. Nerve ligation produced a significant, stable and long-lasting decrease in threshold to mechanical stimulation (i.e., tactile allodynia) when compared to sham-operated controls. Standardization of the diameter of the filaments (to that of the largest filament) did not alter the response threshold in nerve-injured animals. Nerve ligation produced decreased response latency of the ipsilateral paw (i.e., hyperalgesia) when compared to that of sham-operated rats. Tail-flick latencies to thermal stimuli induced by water at constant temperatures (48 degrees, 52 degrees or 55 degrees C) or by radiant heat were not significantly different between nerve-injured and sham-operated groups. At doses which were not behaviorally toxic, MK-801 had no effect on tactile allodynia. At these doses, MK-801 blocked decreased paw withdrawal latency to radiant heat in nerve-injured rats, but did not significantly elevate the response threshold of sham-operated rats. Systemic (i.p.) or intracerebroventricular (i.c.v.) doses of morphine previously shown to be antiallodynic in nerve-ligated rats did not affect the response to probing with von Frey filaments in sham-operated controls. Intrathecal (i.t.) morphine did not change paw withdrawal thresholds elicited by von Frey filaments of either nerve-ligated rats (as previously reported) or of sham-operated rats at doses maximally effective against thermal stimuli applied to the tail or foot. Spinal morphine produced dose-dependent antinociception in both nerve-injured and sham-operated groups in the foot-flick test but was less potent in the nerve-injured group. Presuppression of hyperalgesia of the foot with i.t. MK-801 in nerve-injured animals did not alter the potency of i.t. morphine. I.t. morphine was also active in the tail-flick tests with decreased potency in nerve-injured animals and, at some stimulus intensities, with a decreased efficacy as well. These data emphasize the distinction between the inactivity of morphine to suppress mechanical withdrawal thresholds (as elicited by von Frey filaments) and the activity of this compound to block the response to an acute thermal nociceptive stimulus in sham-operated or nerve-injured rats. It appears that nerve ligation injury produces a thermal allodynia/hyperalgesia which is likely dependent upon opioid-sensitive small-diameter primary afferent fibers and a mechanical allodynia which may be largely independent of small-fiber input.