The rejection of [beta-125I]endorphin and D-[3H]Ala2-Met-enkephalinamide into the lateral ventricles of mice revealed that radioactivity could be transported rapidly via the cerebrospinal fluid (CSF) down to the lowest part of the spinal cord. The maximal levels of both compounds in thoracic, lumbar and sacral sections of the spinal cord occurred 10 min after administration, which coincided with the peak antinociceptive action of morphine after i.c.v. administration. When [3H]Leu-enkephalin was administered i.c.v., the level of radioactivity in the spinal cord did not increase with time and these levels in spinal cord were considerably lower than those of either [beta125I]endorphin or D-[3H]Ala2-Met-enkephalinamide. This difference in distribution of these compounds, coupled with the observation that [beta-125I] endorphin and D-[3H]Ala2-Met-enkephalinamide were more stable than [3H]Leu-enkephalin in CSF, suggested that radioactivity in the spinal cord of mice treated with [beta-125I]endorphin or D-[3H]Ala2-Met-enkephalinamide was due to unchanged material. Very low levels of radioactivity were found in brain or spinal cord after i.v. administration of either [beta-125I]endorphin or D-[3H]Ala2-Met-enkephalinamide. This latter observation supports the view that endorphins administered i.c.v. enter the spinal cord directly through CSF rather than reentry from the general circulation. The results presented herein support the hypothesis that the antinociceptive action of morphine in mice as measured by the tail-flick procedure can be mediated via the release of endogenous compound(s) from brain into CSF which are transported to lower sections of the spinal cord where they inhibit the tail-flick response.