Cognitive deficits of schizophrenia constitute a disabling part of the disease predicting treatment success as well as functional outcome. Phencyclidine (PCP), a non-competitive NMDA receptor antagonist was used to model schizophrenic cognitive dysfunctions of learning and memory using the Morris water maze paradigm for reference memory. In experiment 1 male Sprauge-Dawley rats were acutely administered PCP (0.5, 1.0 and 2.0 mg/kg s.c.) before the first swim session on each of the four acquisition days. Probe test for reference memory was performed 2 days after the last acquisition day; the first probe without drug treatment to assess reference memory and a second probe with prior drug treatment to control for state dependency effects of PCP. In experiment 2 the effects of pre-treatment (10 min before PCP) with the nitric oxide synthase inhibitor, L-NAME (10 mg/kg s.c.), on the PCP (2 mg/kg)-induced spatial memory deficit was evaluated in the Morris water maze paradigm for reference memory. The results showed that PCP in a dose of 2 mg/kg disrupts spatial learning as estimated by prolonged search time to find platform during acquisition as well as the reference memory test as measured by less time spent in target quadrant during probe trial. No state dependency effects of PCP were found. Pre-treatment with L-NAME completely reversed the PCP-induced disruption of acquisition learning. The reference memory disruption was, however, not completely restored as measured by probe trial.