This series of experiments investigated whether the NMDA receptor antagonist D-2-amino-5-phosphonopentanoate (D-AP5) could induce impairments of spatial learning across a dose range comparable to its impairment of hippocampal long-term potentiation (LTP) in vivo. Estimations of the extracellular concentration of D-AP5 in hippocampus using microdialysis were also made to compare whether these impairments occur at concentrations similar to those required to impair LTP in the in vitro hippocampal slice. Rats were chronically infused with D-AP5 into the lateral ventricle at a range of concentrations (0–50 mM) via osmotic minipumps. They were first trained to find and escape onto a hidden platform in an open-field water maze task. After the behavioral learning, they were anesthetized with urethane and an attempt was made to evoke and monitor hippocampal LTP. Extracellular samples of D-AP5 in hippocampus were then taken using microdialysis, and finally, the animals were killed and tissue samples dissected. The microdialysis and tissue samples were analyzed for D-AP5 content using HPLC with fluorescence detection. The results established, first, that D-AP5 impairs spatial learning in a linear dose-dependent manner, highly correlated with its corresponding impairment of hippocampal LTP in vivo. No concentration of D-AP5 was observed to block LTP without affecting learning. Second, the microdialysis estimates indicated that, subject to certain assumptions, D-AP5 causes these impairments at extracellular concentrations comparable to those that impair LTP in vitro. Third, comparison of the whole tissue and microdialysis samples revealed a concentration ratio of approximately 30:1, indicating that 97% of the intracerebral D-AP5 is inaccessible to the dialysis probes. Infusion of 20 mM EGTA was found to cause a sevenfold increase in D-AP5 in the dialysis perfusates, suggesting that at least part of the inaccessible D-AP5 is trapped by a calcium-dependent mechanism. Two further behavioral control studies indicated that the D-AP5-induced impairment of spatial learning is unlikely to be secondary to a drug- induced motor disturbance, and that the performance of the D-AP5 group whose concentration was just sufficient to block hippocampal LTP completely was statistically indistinguishable from that of a group of rats with bilateral hippocampal lesions induced by ibotenic acid. Taken together, these findings offer support for the hypothesis that activation of NMDA receptors is necessary for certain kinds of learning.