The treatment of pain arising from nerve injury can be difficult and the opioid sensitivity of neuropathic pain remains debatable. Clinical and animal studies report a wide range in the effectiveness of morphine, ranging from inadequate to potent analgesia. In this electrophysiological study we compare the effectiveness of spinal versus systemic administration of morphine on the natural and electrically evoked responses of spinal neurones of rats with a selective spinal nerve (L5/6) ligation. Recordings were made 1 week and/or 2 weeks after ligation. We have also compared the effects of morphine, by the two routes, on normal and sham operated animals. In spinal nerve ligated rats, morphine (0.1-5 microg) administered via the intrathecal route produced greater dose-dependent inhibitions of the neuronal responses compared with those produced by the systemic route (1-6 mg/kg). The dose response curves for intrathecal morphine on the C-fibre evoked and noxious natural stimuli evoked neuronal responses (mechanical and thermal) of spinal nerve ligated rats were to the left of those of sham operated and normal rats, suggesting an enhanced potency of intrathecal morphine after nerve injury. This was clearest for the lower doses of the opioid. The effects of spinal morphine on the responses to low intensity stimuli were similar in all groups of rats. In contrast to the spinal route, systemic morphine was less effective in inhibiting the evoked neuronal responses of spinal nerve ligated rats. This was especially clear for the C-fibre evoked and noxious natural stimuli evoked responses (mechanical and thermal) of spine nerve ligated rats. Our results suggest that the effectiveness of morphine may be partly related to the timing of the treatment relative to the duration of the neuropathy, the route of administration and also the neuropathic symptom. Spinal opioids may be a useful approach to pain control in neuropathic pain states where systemic routes produce inadequate analgesia.