The effects of intrathecally administered opiates (morphine sulfate and meperidine), alpha-adrenergic agonists (clonidine and ST-91) and baclofen were examined on the shock titration threshold of macaque monkeys chronically prepared with intrathecal (I) or epidural (E) catheters. Spinal opiates produced a long-lasting analgesia which was antagonized by naloxone. The order of potency was I morphine greater than I meperidine greater than E meperidine greater than E morphine. Clonidine and ST-91, also produced a dose-dependent, long-lasting elevation in the shock titration threshold, antagonized by phentolamine, but not naloxone. L-baclofen, but not D-baclofen, resulted in a dose-dependent elevation of shock titration threshold, which was not antagonized by naloxone. Repeated administration at 24-h intervals over a 7-day period of morphine, clonidine or baclofen, resulted in a significant reduction in the analgetic effects of each drug. Cross tolerance between the three classes of agents was not observed. Intrathecal co-administration of inactive doses of ST-91 and morphine resulted in a near maximal increase in the shock titration threshold, which failed to show any significant tolerance over 21 days. Intrathecal ST-91 and morphine produced no change in either muscle strength, tendon reflexes, respiratory rate, urine formation, or the ability to locomote. Baclofen, in contrast, produced a dose-dependent decrease in muscle strength. That the intrathecal drugs did not produce anesthesia was demonstrated by their failure to block the avoidance response to ensuing ear shock cued by a light tactile stimulus applied to the hind paw. These results clearly indicate that a powerful analgesia can be produced by selectively activating adrenergic, opiate, and baclofenergic receptor systems in the spinal cord.