Experiments were carried out in rats to (1) elaborate upon the sepcificity of drug action in the periaqueductal gray matter (PAG), and (2) to evaluate the posible congruence of PAG sites of morphine-induced and stimulation-produced analgesis (SPA) applied at virtually identical PAG loci. It was demonstrated that the effect of morphine intracerebrally (i,c.) administered into the PAG was not duplicated by other centrally acting agents (chlorpromazine, chlordiazepoxide, pentobarbital or naloxone) administered i.c. at the same PAG site. This selective action of morphine in the PAG was further demonstrated not to be test-bound since morphine significantly altered responding in all four of the analgesiometric tests employed. Thus, multiple i.c. injections of drugs at the same PAG locus were useful in demonstrating site specificity of drug action where behavioral and electroencephalographic methods alone had previously provided ambiguous information. Morphine-induced analgesia and SPA, evaluated at virtually coincident PAG sites, revealed only a general congruence of efficacious loci. The most effective PAG loci for morphine-induced analgesia were not the same as those for SPA; analgesia effected by one analgesia-producing manipulation did not reliably predict that analgesia would also be produced by the other analgesia-producing manipulation at the PAG sites examined. In general, the more efficacious analgesia-producing PAG loci were localized in the ventral-ventrolateral PAG.