A controversy exists concerning the role of the neuropeptide somatostatin for the transmission or inhibition of nociceptive information in the spinal cord. To better correlate electrophysiological effects of somatostatin at single cell level with results obtained with intrathecal injections of somatostatin in behaving animals and human pain patients we applied somatostatin to the spinal cord by controlled superfusion of the recording segment in vivo. The hypothesis of an opioid link and possible neurotoxic effects of somatostatin were also addressed. In cats deeply anaesthetized with pentobarbitone, halothane and nitrous oxide, extracellular recordings were made from 27 neurons located in laminae I-VI. All neurons responded to both innocuous mechanical and noxious radiant heat stimuli applied to the glabrous skin of the ipsilateral hindpaw. The dorsal surface of the spinal cord was superfused at the recording segment by means of a Perspex chamber (7 x 7 mm). Somatostatin superfusions at 1.2 microM had no effect on responses to noxious heat. Responses were, however, depressed by somatostatin at 61 microM to 59.7 +/- 5.1% of control and by somatostatin at 1.53 mM to 39.9 +/- 9.5% of control. This inhibition was not antagonized by the mu-opiate antagonist naloxone applied to the spinal cord at concentrations of 2.7 mM, either together with somatostatin, or after the inhibition by somatostatin had fully developed. Neuronal responses were linear functions of the skin temperatures for stimulation intensities between 42 degrees C and 52 degrees C. The slopes of these stimulus response functions were reduced during somatostatin superfusion at 61 microM to 46.8 +/- 9.3% of control, without changing the temperature thresholds for responding (42.5 +/- 0.6 degrees C). Somatostatin superfusion at 61 microM had no effect on the number of action potentials evoked by innocuous skin brushing, or by electrical stimulation of primary afferent A-fibres in cutaneous nerves. The amplitude of intraspinally recorded field potentials evoked by these electrical nerve stimuli was also unaffected by somatostatin. The inhibition of nociceptive spinal dorsal horn neurons by spinally administered morphine was assessed in eight experiments. Morphine reduced noxious heat-evoked responses to 42.1 +/- 9.6% of control at 0.3 mM and to 51.8 +/- 6.9% of control at 3.0 mM. The slopes of the stimulus-response functions were reduced by morphine at 0.3 mM to 53.1 +/- 11.3% of control, without changing the temperature thresholds (42.7 degrees C). Naloxone superfusion (2.7 mM) reliably antagonized the inhibition by morphine. Brush-evoked responses were not, or much less, affected by spinal morphine.(ABSTRACT TRUNCATED AT 400 WORDS)