This report describes the performance of rats in a delayed non matching-to-position task using an automated visual memory apparatus. A new testing procedure has been used to restrict the problem of motor mediation. In this procedure the rat must first activate a sample stimulus (either the right or left visual display panel). It has then to respond to one of three retractable levers that were extended throughout the retention delay before it can receive a reinforcer (reward pellet vs. nothing) for a response made to either one of the two visual stimulus display panels. A response on the levers was independent from a response on the stimulus display panels. These retractable levers were set in a compartment that protruded 15 cm from the back wall below the display panels, forming a platform which the rat needed to climb in order to reach the stimulus display panels. By requiring the rat to respond to one of these levers the rat could not linger by the correct display panel. Furthermore, as the rat could not anticipate the location of the response lever which would permitted the trial to continue nor the duration of the retention delays before a lever press allow access to the choice stimuli, the rat could not position itself by the correct panel during the retention delay. Rats were able to learn this task readily and perform above chance over delays of up to 16 sec, even though the task demands were likely to restrict any motor mediation strategies.