A controversy exists as to whether morphine attenuates spinal cord nociceptive transmission through a supraspinal site of action. The approach of examining the effects of morphine on spinal cord nociceptive transmission in the presence and absence of spinal cord conduction has led to conflicting conclusions. We have compared the effects of morphine on the rat tail-flick reflex (TFR) in lightly anaesthetized animals in the presence and absence of a spinal cord cold-block. Morphine, administered systemically, was found to be more potent in increasing the latency of the reflex when the spinal cord conduction was present. However, when low doses of morphine were injected intrathecally, morphine was more potent when spinal cord conduction was blocked. These data indicate that systematically administered morphine, at low doses, has a supraspinal site of action in prolonging the onset of the TFR. Conflicting results on this issue do not appear to be due to plasticity changes following spinal cord section or lesions, psychological stress in conscious animals or the presence of tonic bulbospinal inhibition.