Microinjections of morphine in the rat periaqueductal gray matter (PAG) inhibited intestinal transit in linear relation to the log of the dose administered (in the range from 5 to 20 micrograms/rat). This linear regression was parallel with that obtained on intracerebroventricular (i.c.v.) or intraperitoneal (i.p.) administration of morphine and the intracerebral (i.c.) route was calculated to be 4 times more potent than the i.c.v. route and 189 times more potent than the i.p. route. Monolateral electrolytic lesions into the PAG abolished the intestinal effect of i.c.v. morphine to a large extent. The relevance of other brain areas and the type of opiate receptors involved in this central effect of morphine are discussed.