Rats were given extensive training in a complicated appetitive maze. Following a retention interval of 7 days they were given a single electroconvulsive shock (ESC). Animals for which memory cues had been reinstated just prior to ECS were found to be more amnesic upon retest than animals whose memories had not been reinstated. Thus, amnesia can be produced for wellestablished memories if the memory cues are present at the time the amnesic agent occurs.The gradient of amnesia as a function of the learning-ECS interval has long been a primary bulwark supporting the consolidation theory (Dawson, 1971;McGaugh, 1966). According to this theory, the younger the engram the more susceptible it is to disruption, and, in time, the engram consolidates to the point that it can no longer be disrupted. Another interpretation for this gradient has been proposed by Misanin, Miller, and Lewis (1968). They argued that memory recency was the important variable rather than the age of the engram. When a memory is active, or at least in a transition phase from being passive to being active, it is most open to blocking, and therefore ECS will produce an amnestic gradient. To support their notion that ECS is a blocker of active memories rather than a disrupter of labile new memories, they demonstrated an amnesia for an older, well-consolidated memory that was active at the time ECS was presented. This is a form of "cue-dependent amnesia,"