Male Swiss mice were allowed to explore a novel environment, provided by an open-field activity chamber for a 10-min period. The procedure was repeated twice within a 24-h interval. The difference in the exploratory activity between the first (training) and the second exposure (testing) to the chamber was taken as an index of retention of this habituation task. Posttraining intraperitoneal administration of insulin (8, 20, or 80 IU/kg) impaired retention in a dose-related manner, although only the dose of 20 IU/kg of insulin produced significant effects. Thus, the dose-response curve adopted a U-shaped form. Insulin (20 IU/kg) given to untrained mice did not modify their exploratory performance when recorded 24 h later. The effects of insulin on retention were time dependent, suggesting an action on memory storage. An ineffective dose (8 IU/kg) of insulin given together with an ineffective dose of a central acting muscarinic cholinergic antagonist atropine (0.5 mg/kg) or with a central acting nicotinic cholinergic antagonist mecamylamine (5 mg/kg) interacted to impair retention. In contrast, neither methylatropine (0.5 mg/kg), a peripherally acting muscarinic receptor blocker, nor hexamethonium (5 mg/kg), a peripherally acting nicotinic receptor blocker, interacted with the subeffective dose of insulin on retention. The impairing effects of insulin (20 IU/kg) on retention were reversed by the simultaneous administration of physostigmine (70 microg/kg) but not neostigmine (70 microg/kg). We suggest that insulin impairs memory storage of one form of learning elicited by stimuli repeatedly presented without reinforcement, probably through a decrement of brain acetylcholine synthesis.