The authors studied effects of subanesthetic doses of the d- and l-ketamine stereoisomers on maze performance in mice to determine whether the stereoisomers differed in their ability to disrupt a stable cognitive behavior. Twenty-four Swiss-Webster (CFW) mice were trained to stability in a four-compartment modular maze, using water as a reward. Each compartment contained a central partition with a barrier at the distal end of one of the two passageways. A fixed barrier-sequence was employed. Elapsed time to traverse all four compartments and total number of errors (the number of times a wrong compartment was entered) were measured. A cohort design was employed with the following four groups: saline control, d-ketamine, l-ketamine, racemate. Two subanesthetic doses, 7.5 and 15 mg/kg of each form of the drug were given subcutaneously at five-day intervals. Both the d-isomer and the racemate significantly prolonged elapsed time at 15 mg/kg, the d-isomer having the greatest effect. The l-isomer did not alter elapsed time at either dose but appeared to increase spontaneous locomotor activity after injection. Relative to errors, at the 7.5 mg/kg dose there were no changes from control with any form of ketamine. However, at the 15 mg/kg dose, total errors significantly increased both with the racemate and the d-isomer. The performance decrements observed with the racemate appear to be attributable largely to the d-component.