The prostaglandin hyperalgesia and tail immersion tests were used to evaluate the analgesic action of morphine, codeine, d-propoxyphene and pentazocine following intraperitoneal, intraplantar and intracerebroventricular administration to rats. In the prostaglandin hyperalgesia test, all drugs produced a dose-dependent analgesia by the various routes. The rank order of potency after intraperitoneal administration was morphine (100) greater than d-propoxyphene (4) greater than pentazocine (2) greater than codeine (1). Although morphine (ID50 = 4 micrograms) was a very potent analgesic when given intracerebroventricularly, very shallow dose-response curves were obtained with the other substances which promoted less than 30% of inhibition at doses up to 250 micrograms. In the paw, morphine (ID50 = 5 micrograms) was only 5-8 times more potent than pentazocine, propoxyphene and codeine. Thus, in contrast with morphine, intraplantar administration of codeine, pentazocine and d-propoxyphene is much more effective than intracerebroventricular administration. In the tail immersion test the smallest intraperitoneal doses which affected the reaction time were 9 mg/kg morphine, 16.2 mg/kg codeine and pentazocine and 48.6 mg/kg d-propoxyphene. When injected intracerebroventricularly morphine (10 micrograms) was the only opiate that caused a detectable analgesic effect. In the prostaglandin hyperalgesia test, a small dose of naloxone (1 micrograms) given into the rat paw significantly antagonized the analgesic effect of d-propoxyphene, codeine and pentazocine administered either intraperitoneally or intraplantarly. These results clearly indicate that a method involving or mimicking inflammatory hyperalgesia is much more sensitive in detecting opiate analgesia than a method which uses heat as a nociceptive stimulus. Furthermore, our results support the proposition that part of the overall analgesia which follows the systemic administration of opiates is due to a peripheral antinociceptive action.