Nine naive rats of the PVG strain were trained to non-match to sample in a visual automated apparatus using repeated stimuli. Their memory was then tested at different retention delays (0, 2, 4, 8, 12 sec) presented in mixed order within a session. The duration of the intertrial intervals (ITI= 30 sec) was adjusted to the duration of the next retention interval (RT) in order to equilibrate proactive interference across the different delays (ITI= 30 sec + next RT). The rats learned quickly to non-match to sample and performed the memory task well above the chance level at different retention intervals (in every session each rat achieved a score of 69% or more for at least one delay). However, their memory performance was independent of the duration of the retention interval. In the present paper, we describe the conditions under which these rats were trained and their memory tested.