It has previously been shown that, in normal subjects, heterotopic painful stimuli induce parallel decreases in the sensation of pain and the nociceptive spinal flexion reflex RIII simultaneously evoked by electrical stimulation of the sural nerve. By contrast, heterotopic non-noxious stimuli were ineffective in this respect. In order to assess the possible involvement of supraspinal structures in these changes, we have now compared the effects of nociceptive electrical stimuli applied to the fourth and fifth fingers of the hand on the contralateral RIII reflex in 5 normal subjects and 5 tetraplegic patients suffering from a clinically complete spinal cord transection of traumatic origin. The tetraplegic patient had a C5, C6 or C7 spinal cord transection with lower segments, notably C8 and T1, clinically uninjured. In these patients, sural nerve stimulation induced a biceps femoris reflex which did not differ from the RIII reflex recorded in normal subjects, either in terms of latency or of threshold. In both groups, the RIII reflex threshold was determined and a juxta-threshold RIII reflex was studied before, during and after a 2 min period of conditioning stimulation. The conditioning procedure consisted of nociceptive electrical stimulation applied to the digital branches of the ulnar nerve which arises from C8 and T1, i.e., below the level of the patients' spinal transection. In normal subjects such conditioning elicited painful sensations of 'pin prick', 'burning' and 'squeezing'. It simultaneously induced a strong inhibition of the RIII reflex which outlasted the conditioning period by several minutes. By contrast, non-nociceptive electrical conditioning stimuli (0.1 ms duration pulses of 6 to 8 mA intensity, at 3/s), similarly applied to the digital branches of the ulnar nerve were totally ineffective. In tetraplegic patients not only was the RIII reflex never decreased by the heterotopic nociceptive stimulation but it was, in fact, slightly enhanced. These results show that the effects of heterotopic nociceptive stimulation observed in normal man are possibly mediated by a complex loop involving supraspinal structures. This further underlines the similarity between this phenomenon and the Diffuse Noxious Inhibitory Controls (DNIC) described in the rat, since the latter are observed in intact but not in spinal animals. These findings are discussed with reference to counter-irritation phenomena and procedures aimed at producing analgesia by somatic electrical stimulation.