We compared the effects of systemic morphine on normal (heat and cold) pain and paradoxical burning pain evoked by the simultaneous application of innocuous warm and cold stimuli to the skin. Twelve healthy volunteers participated in a randomised, double-blind, cross-over study to compare the effects of intravenous administration of morphine (0.025 or 0.1mg/kg) or placebo (saline). Stimuli were applied to the palm of the right hand with a thermode ("thermal grill") composed of six bars, whose temperatures were controlled by Peltier elements. For each session, we measured the heat and cold pain thresholds and then successively measured the intensity of: (i) paradoxical pain evoked by a combination of non-noxious warm and cold stimuli; (ii) "normal" pain evoked by suprathreshold heat or cold stimuli; (iii) non-painful sensations evoked by warm or cold stimuli at temperatures used to produce paradoxical pain. Measurements were performed before 20min after the administration of morphine or placebo and 5min after the administration of the morphine antagonist, naloxone. The administration of 0.1mg/kg of morphine, but not 0.025mg/kg, induced a significant and naloxone-reversible reduction of paradoxical pain intensity, which was directly correlated with the reduction of normal cold pain. No differences were observed for non-painful thermal sensations. The paradoxical burning pain evoked by a thermal grill can be modified pharmacologically by analgesics and share some mechanisms with normal pain. This unique experimental "illusion of pain" may represent a new model to test analgesics in healthy volunteers.