Morphine releases endogenous opioids into the circulation of dogs. To test the stereospecificity of this effect, as well as to determine whether morphine also releases endogenous opioids centrally, which might be involved in its antinociceptive action, the effects of (-)-morphine sulfate (10 mg/kg, sc) or (+)-morphine hydrobromide on antinociception in a dog tail-flick test, on semi-quantified morphine-induced signs of salivation, emesis, defecation and ataxia, and on the plasma and cerebrospinal fluid (CSF) levels of endogenous opioid peptides were studied. Plasma and CSF levels of immunoreactive beta-endorphin (i-BE), met-enkephalin (i-ME), leu-enkephalin (i-LE), and dynorphin (i-DY) were quantified by radioimmunoassay in octadecylsilyl-silica cartridge extracts. Immunoreactive morphine (i-M) levels were measured in unextracted samples. (-)-Morphine treatment significantly increased antinociception, morphine-induced signs, i-M levels in plasma and CSF, and i-BE, i-ME, and i-LE levels in plasma, but not CSF. Levels of i-DY remained constant in plasma and CSF. (+)-Morphine treatment did not alter any of these parameters, indicating that the effects of morphine on nociception, behavioral signs, and plasma endogenous opioids in dogs were stereoselective. It is concluded that morphine does not cause an increase in immunoreactive endogenous opioid peptides in the CSF at the time of its peak antinociceptive effect.