Aged Long-Evans rats were screened for spatial memory deficits using the Morris water maze task. Rats found to have impaired performance on the task (aged-impaired, AI) were then treated with a selective muscarinic M2 receptor antagonist, 5,11-dihydro-8-chloro-11-[[4-[3-[(2,2-dimethyl-1-oxopentyl)ethylamino]propyl]-1-piperidinyl]acetyl]-6H-pyrido[2,3-b][1,4]benzodiazepin-6-one (BIBN-99; 0.5 mg/kg, s.c.), for 3 successive days while receiving additional water maze training. BIBN-99 significantly improved performance in the task during the 3 days of drug treatment. Treatment was then ceased for the remainder of the study and rats were tested again in the water maze on days 10, 17, and 24. Compared to vehicle-treated rats, enhanced performance was observed in the AI rats that had previously been treated with BIBN-99. These results indicate that BIBN-99 enhances spatial learning in AI animals and that enhanced (or long-term) memory persists in the absence of the drug. In a second experiment, a 2-month delay was imposed in between the original water maze screening and the drug treatment regime. Again, BIBN-99 significantly improved performance in AI rats. This latter study suggests that reference memory does not decay, even in an AI animal that had displayed poor learning following original water maze screening. Together, these studies help provide further insight into possible mechanism(s) of reference memory and its potential clinical usefulness.