Twenty patients with chronic pain who previously had obtained analgesia from epidural clonidine and lignocaine agreed to participate in a double-blind crossover study of lumbar epidural clonidine (150 micrograms), lignocaine (40 mg) and the combination of clonidine (150 microgram) and lignocaine (40 mg), all drugs were given in a volume of 3 ml. There were 11 women and 9 men with a mean age 53 years (range: 23-78 years); 9 patients had low back and leg pain, 9 had neuropathic pain, 1 had pelvic pain and 1 Wegner's granulomatosis. Pain intensity and pain relief, as well as sensory and motor blockade, were assessed for 3 h following each injection. The combination was reported as the best pain relief by 12 of the 17 patients who completed all three arms of the study; 4 patients reported that clonidine was the best, 1 patient reported that none of the injections provided any analgesia and no patient reported that lignocaine was the best. SPID analysis revealed a significant difference between the combination and lignocaine (P < 0.05) but no other significant difference. TOTPAR analysis revealed no significant difference between any of the injections. All 3 injections produced evidence of neurological blockade; clonidine produced sensory blockade in 3 patients and motor blockade in 3 patients. Lignocaine produced sensory blockade in 6 patients and motor in 8 patients, while the combination produced evidence of neurological blockade in all 17 patients, sensory in 6 and motor in 11 patients. Overall there was no relationship between neurological blockade and analgesia. The reported side effects appeared to be related to clonidine. These data indicate that in these patients with chronic pain epidural clonidine had a supra-additive effect and behaved more like a co-analgesic than a pure analgesic.