Although numerous data support the existence of a presynaptic inhibitory control by opioids of substance P-containing primary afferent fibres entering the dorsal horn of the spinal cord, the exact nature of the opioid receptor involved in this control is still a matter of debate. In the present study, the potential role of delta opioid receptors was investigated by looking for the possible effects of selective delta-ligands on the in vivo release of substance P-like material from the whole spinal cord in halothane-anaesthetized rats. Perfusion of the intrathecal space allowed the collection of substance P-like material that was released at a constant rate of approximately 0.65 pg substance P equivalents/min for at least 135 min. The addition of Tyr-D-Thr-Gly-Phe-Leu-Thr (10-mu-M) or dermenkephalin (10-mu-M), two selective delta agonists, to the perfusing fluid produced a marked reduction (-50-65%) in substance P-like material outflow which could be prevented by the selective delta antagonist naltrindole (10-mu-M) but not by naloxone (10-mu-M), which acts preferentially on mu opioid receptors. Furthermore, naltrindole alone (or the association of this antagonist plus dermenkephalin) enhanced the outflow of substance P-like material (+170%) as expected from the blockade of a tonic inhibitory control due to the stimulation of delta receptors by endogenous opioids. In contrast, naloxone alone did not modify the spontaneous release of substance P-like material from the spinal cord of halothane-anaesthetized rats, indicating that mu opioid receptors are probably not playing an important role in this tonic opioid control of substance P-containing fibres. These data suggest that the antinociceptive effect of delta agonists injected via the intrathecal route may involve, at least partly, a presynaptic inhibitory control of small diameter substance P-containing primary afferent fibres.