Initial suggestions on the involvement of glutamate in memory came from electrophysiological studies on LTP that is blocked by NMDA-antagonists. Then Morris and colleagues (1986) provided the first evidence that icv infusion of the competitive NMDA antagonist 2-amino-5-phosphonovaleric acid (APV) to rats, inhibits both LTP in vivo and spatial learning in a Morris water maze. This was followed by a great amount of evidence confirming the initial finding in various learning tasks. The present paper is devoted to critical review of the literature focusing on the following problems: which glutamate receptors are involved ?, in which tests NMDA antagonists inhibit learning ?; which types of memory are affected ?; which brain structures are involved ? do NMDA receptor antagonists invariably impair learning ?; is the effect of NMDA receptors antagonists on learning specific ?; does the stimulation of NMDA receptors result in cognitive enhancement ?.