Alpha-2-adrenoceptor agonists can activate varying antinociceptive mechanisms depending on the dose and the route of administration, although the main site for their antinociceptive effect in physiological pain conditions seems to be the spinal dorsal horn. In this paper the investigations on the underlying mechanisms are reviewed, with particular emphasis on novel studies using a highly selective and potent alpha-2-adrenoceptor agonist medetomidine. In behavioral studies alpha-2-adrenoceptor agonists, including medetomidine, produce antinociception following systemic administration or local application to the spinal cord. Following systemic administrations the antinociceptive effect at subanesthetic doses in physiological pain conditions is weak and the affective-motivational component of pain is more sensitive to suppression than sensory-discriminative aspect of pain. Electrophysiological studies at the spinal cord level indicate that alpha-2-adrenoceptor agonists suppress the responses of the pain-relay neurons both due to pre- and postsynaptic spinal cord level mechanisms following systemic or intrathecal drug administration. The antinociceptive effect at the spinal cord level is selective, since non-nociceptive signals are not attenuated. According to electrophysiological single neuron recordings, nociceptive signals at supraspinal levels are more sensitive to antinociceptive effects of systemically administered alpha-2-adrenoceptor agonists than nociceptive signals at the spinal cord level. However, according to immunocytochemical imaging of immediate early gene response to noxious stimulation, medetomidine strongly suppresses responses at the spinal cord level already following a systemic dose (100 micrograms/kg i.p.) which is not enough to produce a significant suppression of spinally-initiated nocifensive tail reflexes nor the electrophysiologically recorded responses to nociceptive spinal neurons. Thus, the higher sensitivity of supraspinal neuronal responses and their behavioral correlates to the antinoceptive effect of medetomidine obviously reflects the cumulative effect of medetomidine at several areas along the polysynaptic pathway to the rostral parts of the brain. Paradoxically, the response of immediate early genes in the medial thalamic neurons is only slightly, influenced by antinociceptive doses of medetomidine. Alpha-2-adrenoceptors have significant interactions with other receptors (e.g. opioid, serotonin and muscarine) in producing antinociception at the spinal cord level.