Choice accuracy by rats in a delayed-alternation paradigm was shown to decrease over a 120-sec retention interval. The decrement in choice accuracy was reversed by presentation of the appetitive reinforcer outside ofthe apparatus during the retention interval. This suggests that the reinforcer served to reactivate the target spatial memory and that the short-term retention deficit in the absence of such memory reactivation was not due to a loss of information. The results are discussed with respect to recent criticisms of memory-reactivation treatments in short-term retention tasks.In principle, associative deficits can be caused by failure of information storage at the time of acquisition, irreversible decay of information during the retention interval, or failure of information retrieval at the time of retention testing. Although it is clear that processes that occur in temporal proximity to the learning event and during the retention interval are important for later retention (e.g., McGaugh, 1966), several memory theorists have emphasized the importance of retrieval processes (e. g., Lewis, 1979;Miller, Kasprow, & Schachtman, 1986;Spear, 1971Spear, , 1973Spear, , 1978.A critical observation in support of retrieval-based models of memory has been the reversal of a variety of associative performance deficits through memory reactivation, that is, "reminder" treatments. A reminder treatment re-presents, prior to testing, some subset of the stimuli present during the original learning episode. It has been hypothesized that such exposure serves to stimulate further processing of the learning event so that future retrieval is facilitated (Spear, 1978). Most studies that have demonstrated the reversal of associative performance deficits have included conditions to preclude nonassociative interpretations of reminder-induced behavioral change (see Miller et al., 1986, for a recent review).The vast majority of successful restorations of associative performance by reminder treatments have used longterm retention tasks. Few studies using animals as subjects have investigated the effect of reminder treatments on short-term retention, but those directed toward the problem have obtained positive results. Feldman and Gordon (1979) alleviated forgetting over a 9O-sec retention interval in aT-maze delayed-alternation preparation using rats. In delayed alternation, an animal is initially forced