In order to assess the effects of glucose on drug-induced spatial learning deficits, three experiments were conducted using the Morris water maze. Scopolamine and glucose were injected ip at various stages of training. Rats of Wistar strain served as subjects. In Experiment 1, scopolamine (0.4 mg/kg) and 10, 100, or 500 mg/kg of glucose were administered every day from the start of training, and the effect on acquisition was evaluated. In Experiment 2, scopolamine and 100 or 500 mg/kg of glucose were administered after 6 days of training, and the effect on performance was assessed. In Experiment 3, scopolamine and 500 mg/kg of glucose were injected after 2 days of training, and the effect on the following trial was tested. In all experiments, scopolamine impaired acquisition/performance of the task. Glucose at 500 mg/kg showed a significant enhancing effect on acquisition regardless of scopolamine injection only when injected daily from the start of training (Experiment 1). Glucose injected after the performance has reached asymptote (Experiment 2) did not affect performance, and glucose in the middle of training showed a slight but insignificant enhancing effect (Experiment 3). These results may suggest that the effect of glucose changes as a function of the degree of learning of the spatial learning task. The possibility of task specificity of the glucose effect was also discussed in relation to the cholinergic systems and local cerebral glucose utilization.