The kappa opioid agonists U50,488, bremazocine, ethylketazocine and tifluadom and the opioid antagonist naltrexone were examined alone and in combination with morphine in a squirrel monkey shock titration procedure, before and during chronic morphine administration. Before chronic morphine administration (prechronic phase), all opioids except naltrexone produced dose-dependent increases in median shock level when administered alone. When combined with a dose of morphine that increased median shock level to 90% of maximum (ED90), naltrexone, U50,488 and bremazocine completely antagonized the effects of morphine in most monkeys, whereas ethylketazocine and tifluadom partially antagonized the effects of this dose of morphine. After 10 weeks of daily morphine administration (chronic phase), the average ED90 for morphine was increased 1 log unit. In contrast, average ED50 values for U50,488, bremazocine and tifluadom were decreaed 1/4 to 1/2 log unit, whereas the average ED50 for ethylketazocine did not change from the prechronic to chronic phases. When combined with morphine during the chronic phase, naltrexone completely antagonized the effects of the morphine ED90 at approximately the same doses as during the prechronic phase. In contrast, antagonist activity decreased for U50,488 and bremazocine, increased for ethylketazocine and did not change consistently for tifluadom, compared with the prechronic phase. The present study demonstrates that chronic morphine administration alters both the agonist and antagonist activity of kappa opioids. Changes in antagonist activity of kappa opioids after chronic morphine administration may be explained by concurrent changes in their agonist potency and the extent to which their agonist effects are mu-mediated.