Explanations of the misinformation effect were considered in an experiment using a reversed eyewitness suggestibility design (Lindsay & Johnson, 1989b). Forty-eight subjects read a narrative describing a photograph that they subsequently viewed. For half the subjects, the narrative contained misinformation. Recognition tests for objects appearing in the photograph were administered in either a verbal or a pictorial modality. A misinformation effect was found in the verbal condition, as found by Lindsay and Johnson. With pictorial probes there was no misinformation effect, indicating that reinstatement of the appropriate modality cues can eliminate the influence of misleading post-event information upon memory and permit the retrieval of the target memory. The experiment was conceived of and discussed within the headed records framework.