NMDA receptor antagonists (MK-801, memantine, amantadine, CGP37849), and the AMPA receptor antagonist NBQX, were tested in a continuous multiple-trial two-choice dark avoidance paradigm. During training MK-801 (0.2mg/kg) and CGP-37849 (12mg/kg) increased the number of trials to criterion. Amantadine (92mg/kg) increased all parameters except the number of trials to criterion. CGP37849 (6 and 12mg/kg) increased the number of shocks per trial. Injection of MK-801 (0.2mg/kg), memantine (20mg/kg), amantadine (92mg/kg) or CGP37849 (6 or 12mg/kg) before training decreased the latency to enter the dark compartment in the retention test 24h later. NBQX affected neither training nor retention parameters. NMDA antagonists were also tested in a single-trial two-choice dark avoidance procedure. Pre-training but not pre-retention test injection of MK-801 (0.2mg/kg), memantine (20mg/kg), amantadine (92mg/kg) or CGP37849 (12mg/kg) decreased the latency to enter the dark compartment during retention; scopolamine had similar effects. Hence, NMDA receptor antagonists, but not the AMPA receptor antagonist NBQX, impair memory storage which is also seen in animals trained to criterion, indicating that interference with the acquisition processes is not solely responsible for the memory deficit. This deficit occurs at doses that do not appear to interfere with the recall and the performance of the previously learned response.