1. In rats, the interaction between the mu-opioid agonist dermorphin and the delta-opioid agonist [D-Ala2, Glu4]deltorphin was studied in binding experiments to delta-opioid receptors and in the antinociceptive test to radiant heat. 2. When injected i.c.v., doses of [D-Ala2, Glu4]deltorphin higher than 20 nmol produced antinociception in the rat tail-flick test to radiant heat. Lower doses were inactive. None of the doses tested elicited the maximum achievable response. This partial antinociception was accomplished with an in vivo occupancy of more than 97% of brain delta-opioid receptors and of 17% of mu-opioid receptors. Naloxone (0.1 mg kg-1, s.c.), and naloxonazine (10 mg kg-1, i.v., 24 h before), but not the selective delta-opioid antagonist naltrindole, antagonized the antinociception. 3. In vitro competitive inhibition studies in rat brain membranes showed that [D-Ala2, Glu4]deltorphin displaced [3H]-naltrindole from two delta-binding sites of high and low affinity. The addition of 100 microM Gpp[NH]p produced a three fold increase in the [D-Ala2, Glu4]deltorphin Ki value for both binding sites. The addition of 10 nM dermorphin increased the Ki value of the delta-agonist for the high affinity site five times. When Gpp[NH]p was added to the incubation medium together with 10 nM dermorphin, the high affinity Ki of the delta-agonist increased 15 times. 4. Co-administration into the rat brain ventricles of subanalgesic doses of dermorphin and [D-Ala2, Glu4]deltorphin resulted in synergistic antinociceptive responses. 5. Pretreatment with naloxone or with the non-equilibrium mu-antagonists naloxonazine and beta-funaltrexamine completely abolished the antinociceptive response of the mu-delta agonist combinations. 6. Pretreatment with the delta-opioid antagonists naltrindole and DALCE reduced the antinociceptive response of the dermorphin-[D-Ala2, Glu4]deltorphin combinations to a value near that observed after the mu-agonist alone. At the dosage used, naltrindole occupied more than 98% of brain delta-opioid receptors without affecting mu-opioid-receptors. 7. These data suggest that in the rat tail-flick test to radiant heat, mu- and delta-opioid agonists co-operate positively in evoking an antinociceptive response. Although interactions between different opioid pathways cannot be excluded, in vitro binding results indicate that this co-operative antinociception is probably mediated by co-activation of the delta-opioid receptors at the cellular level by the mu- and delta-agonist.