The pharmacology of memory has been recently studied by the infusion of drugs into the hippocampus (HIP), amygdala (AMY), medial septum (MS), and entorhinal cortex (EC) at various times after training or at the time of retention testing. It was found to be remarkably similar to that of long-term potentiation (LTP). Memory and LTP are blocked early on by antagonists of glutamate N-methyl-D-aspartate (NMDA) or metabotropic receptors (mGLUs), by the antagonist of the presynaptic membrane receptor to PAF, BN 52021, by the inhibitor of heme oxygenase, ZnPP, by the inhibitor of NO synthase, N-nitro-arginine, by GABA type A receptor agonists, or by muscarinic blockers. Both memory and LTP are enhanced, at this early stage, by glutamate, mGLU agonists, GABA-A antagonists, muscarinic agonists, and norepinephrine. In the next 1-3 h, memory and LTP are accompanied by enhanced activity of protein kinases and are blocked by specific inhibitors of calcium/calmodulin dependent protein kinase II and protein kinase C. At the time of expression, memory and LTP are blocked by antagonists of glutamate AMPA receptors and are accompanied by an enhanced sensitivity of these receptors. Memories that depend on HIP are affected by drugs given into the HIP but not the MS or AMY, memories that depend on the AMY are affected by drugs given into the AMY, and memories that depend on the HIP, AMY, and MS are affected by drugs given into the three structures. Drug effects on the EC are less specific as to memory type, and are seen only 1-3 h after those in the HIP, MS, or AMY, suggesting a slow flux of information from the latter to the former and an integrative role of the EC. The data suggest that memories rely on interrelated LTP processes in the HIP, AMY, MS, and EC. (C) 1995 Academic Press, Inc.