1. Single-cell recordings were made from neurones in various spinal laminae in anaesthetized or decerebrated, spinalized or intact rats and cats. Cells were activated by controlled peripheral sensory stimuli which mimicked natural conditions and with some cells also by micro-electrophoretically administered excitatory amino acid analogues. Such responses were tested with amino acid antagonists administered both micro-electrophoretically and intravenously. 2. With cells in the dorsal horn, the dissociative anaesthetic ketamine, administered either micro-electrophoretically or intravenously at doses which selectively reduce responses to N-methylaspartate, had no consistent effect on any of the sensory responses examined. 3. The non-selective amino acid antagonist cis-2,3-piperidine dicarboxylate was somewhat more effective at reducing sensory responses. 4. With motoneurones, intravenous N-methylaspartate-blocking doses of ketamine consistently reduced nociceptive responses. Non-nociceptive responses were less affected. 5. With ventral horn interneurones, intravenous but not micro-electrophoretic ketamine reduced nociceptive responses on about half the cells tested. 6. These results are interpreted in terms of the physiological role of the N-methylaspartate class of excitatory amino acid receptor in mediating responses in the ventral but not dorsal horn of the spinal cord to peripheral somatic stimuli.