In a rat model of peripheral mononeuropathy produced by moderate constriction of the sciatic nerve, we have shown that various i.v. doses of morphine and selective opioid agonists produce potent and long-lasting antinociceptive effects on the vocalization threshold to paw pressure. For all the opioids, the antinociceptive effects were more marked for the paw on the nerve-injured side (nerve-injured paw) than for the sham-operated paw. One contributory mechanism could be a peripheral action of the opioid agonists in the nerve-injured paw. This hypothesis was tested in the present study, using systemic morphine and low doses of local naloxone or its quaternary salt naloxone methiodide, exhibiting peripherally acting antagonist properties. The effects of escalating doses of naloxone (0.5-2 mu g injected i.v. or intraplantar into the nerve-injured paw) or naloxone methiodide (5-30 mu g into the nerve-injured paw) on the antinociceptive effect of morphine (1 mg/kg i.v.) were evaluated using the vocalization threshold to paw pressure in neuropathic rats at two weeks after placing ligatures, a time when the behavioural pain-related disorders have reached a maximum. We show that (i) the intraplantar injection of naloxone (0.5, 1 and 2 mu g in a volume of 0.1 ml) or naloxone methiodide (10, 20 and 30 mu g in a volume of 0.1 ml) into the nerve-injured paw significantly and dose-dependently reduced the effect of morphine on this paw; (ii) the local injection of equal doses of naloxone or naloxone methiodide into the sham-operated paw did not significantly reduce the effect of morphine on either paw; (iii) the intraplantar injection of 0.1 ml saline or the top dose of naloxone or naloxone methiodide alone into the nerve-injured paw had no significant effect; (iv) the same doses of naloxone injected i.v. failed to modify the antinociceptive effects of morphine on either the nerve-injured or the sham-operated paws. These results suggest that the augmented analgesic potency of morphine on the vocalization threshold to paw pressure in rats with a peripheral mononeuropathy may partly result from a peripheral action of this substance, as is the case in models of unilateral inflammatory pain.