Intrathecal administration of dynorphin A in rats produced dose-dependent antinociceptive effects in the tail-flick test to radiant heat and on a limb-flexion test to pressure. The potency of dynorphin A as an analgesic agent on the tail-flick test was dependent upon the duration of cannula implantation. When a short-term procedure was used (drug injected 1 day after catheter implantation) dynorphin A was approximately equipotent to morphine, whereas in animals with long-term implants (drug injected 7 or more days after catheter implantation) dynorphin A was an order of magnitude less potent than morphine. [D-Ala2,D-Leu5]enkephalin was the most potent opioid tested, and in the tail-flick test (long-term procedure) it was about 2 orders of magnitude more potent than dynorphin A and 7 times more potent than morphine. [Leu]enkephalin had no detectable antinociceptive effects. Low doses of naloxone (1 and 2 mg/kg s.c.) completely blocked the antinociceptive effects of morphine and [D-Ala2,D-Leu5]enkephalin, but neither low nor high (40 mg/kg s.c.) doses clearly blocked the antinociceptive effects of dynorphin A. Thus, dynorphin A has an antinociceptive action at the level of the spinal cord, and mu opioid receptors do not mediate these effects. In addition, high doses of dynorphin A (20 nmol or greater) produced long-lasting hindlimb paralysis, which suggests that dynorphin peptides may play a role in motor function in the spinal cord. This paralytic action of dynorphin A was not antagonized by naloxone in doses up to 32 mg/kg s.c.