With a relatively complex maze. reliable forgetting is clearly seen when the training-test interval is 25 days. This forgetting is evidenced by the longer time taken to run the maze and in an increase in the number of errors from the last training trial to the first test trial. In this case. forgetting is a lapse. not a loss. since performance attains the last training trial level at a subsequent test. Furthermore, a reminder which does not in itself contain sufficient information to facilitate performance of a naive animal, significantly improves maze performance of animals which have "forgotten," even on the first retention test. With the use of additional control groups, it is shown that there must be a memory lapse before contextual cues can be demonstrated to be effective in facilitating memory retrieval.The importance of context (extra list retrieval cues) in information retrieval has long been recognized in human subjects (see review by Tulving & Thompson, 1973). Recent studies suggest that contextual cues playa role in the retrieval of memory in nonhuman subjects as well (Hickis, Robles, & Thomas, 1977;Spear, 1971Spear, , 1973 Tornie, 1976). Animal memory studies, using such diverse techniques as experimental amnesia (Quartermain, McEwen, & Azrnitia, 1972;Sara, 1973), proactive interference (Spear, 1971), or spontaneous forgetting (Gatti, Pais, & Weeks, 1975), have shown that the attributes found in the experimental environment can be extremely important in retrieval of weak memories. A recent study of spontaneous forgetting in very young rats has corroborated the importance of contextual cues in reinstatement of memory in animals (Haroutunian & Riccio, 1977).The present series of experiments examines the effects of pretest presentation of background stimuli or cues on retrieval of a forgotten habit. The evaluation of the effects of environmental manipulations on retrieval processes requires a behavioral paradigm Requests for reprints should be sent to B. Deweer or S. J. Sara, whose present address is: Division of Behavioral Neurology, New York University School of Medicine, 341 E. 25th Street, New York, New York 10016. in which significant forgetting over time can be clearly demonstrated, forgetting being defined as a performance decrement after a long training-test interval.The "antimnemonic bias" in animal studies is evidenced by the relatively few studies concerning longterm retention (see Winograd, 1971, for a historical survey). What data are available are rather conflicting. The studies of retention, over years, of keypecking in pigeons (Skinner, 1950) have traditionally been cited as an example of the durability of welllearned habits. However, it has been pointed out that careful examination of Skinner's data reveals a good deal of forgetting (Spear, 1976). Nevertheless, in a recent study of retention of a negatively reinforced discrimination task in gerbils, the investigators concluded that "the original learning endured, unaffected by the passage of time" (Neylon & Brosgole,