The role of nitric oxide on passive avoidance learning was studied by administering L-arginine or D-arginine to male rats in a passive avoidance paradigm. L-Arginine administered into the lateral brain ventricle at a dose of 1.25 microg showed a tendency to increase the passive avoidance latency, and 2.5 microg exerted almost maximal action, but the action gradually increased still further up to 20 microg tested. D-Arginine had no action. Peripheral administration (intraperitoneal) of L-arginine facilitated the consolidation of passive avoidance learning in a dose-dependent manner. A significant increase in passive avoidance response was obtained following an injection of 100 mg/kg L-arginine. When L-arginine was given i.c.v. with a selected dose of 5 microg, 30 min prior to a learning trial, the latency of the passive avoidance response was likewise lengthened. However, when L-arginine was given 30 min before the 24-hr testing (retrieval), it was ineffective. It was also ineffective when given 6 hr after the training trial. However, when L-arginine was administered immediately following the training trial, the action in improving the consolidation could be detected 6 hr after the training trial. Nitro-L-arginine, which blocks nitric oxide synthase, can also block the facilitation of consolidation caused by the nitric oxide donor L-arginine. The nitric oxide synthase inhibitor per se in different doses had no action on the learning of a passive avoidance task. The results indicate that nitric oxide is able to facilitate the learning and consolidation of memory in a passive avoidance paradigm, but it is ineffective in retrieval processes. The results also suggest that, under the experimental circumstances used, nitric oxide is involved only in the facilitated learning and memory processes caused by pharmacological effect of L-arginine, and not involved in normal learning processes.