The modulatory effects of electrical and chemical (glutamate) stimulation in the rostral ventromedial medulla (RVM) on spinal nociceptive transmission and a spinal nociceptive reflex were studied in rats. Electrical stimulation at a total 86 sites in the RVM in the medial raphe nuclei (n = 54) and adjacent gigantocellular areas (n = 32) produced biphasic (facilitatory and inhibitory, n = 43) or only inhibitory (n = 43) modulation of the tail-flick (TF) reflex. At these 43 biphasic sites in the RVM, facilitation of the TF reflex was produced at low intensities of stimulation (5-25 microA) and inhibition was produced at greater intensities of stimulation (50-200 microA). At 43 sites in the RVM, electrical stimulation only produced intensity-dependent inhibition of the TF reflex. Activation of cell bodies in the RVM by glutamate microinjection reproduced the biphasic modulatory effects of electrical stimulation. At biphasic sites previously characterized by electrical stimulation, glutamate at a low concentration (5 nmol) produced facilitation of the TF reflex; a greater concentration (50 nmol) only inhibited the TF reflex. In electrophysiological experiments, electrical stimulation at 62 sites in the RVM produced biphasic (n = 26), only inhibitory (n = 26), or only facilitatory (n = 10) modulation of responses of lumbar spinal dorsal horn neurons to noxious cutaneous thermal (50 degrees C) or mechanical (75.9 g) stimulation. Facilitatory effects were produced at lesser intensities of stimulation and inhibitory effects were produced at greater intensities of stimulation. The apparent latencies to stimulation-produced facilitation and inhibition, determined with the use of a cumulative sum method and bin-by-bin analysis of spinal neuron responses to noxious thermal stimulation of the skin, were 231 and 90 ms, respectively. The spinal pathways conveying descending facilitatory and inhibitory influences were found to be different. Descending facilitatory influences on the TF reflex were conveyed in ventral/ventrolateral funiculi, whereas inhibitory influences were conveyed in dorsolateral funiculi. The results indicate that descending inhibitory and facilitatory influences can be simultaneously engaged throughout the RVM, including nucleus raphe magnus, and that such influences are conveyed in different spinal funiculi.