The effect of presentation modality on recall was studied in two experiments. In the first, lists of alternating auditory and visual words were presented at a 15-second rate, with a visual distractor task interpolated between each word. Recall was asked for immediately after each list, and again after all lists had been presented. A large modality effect (auditory advantage) was obtained in immediate recall but no difference between modalities was found in final recall. In the second experiment, this modality effect was attenuated with the use of phonologically similar words, and the effect was replicated with a fast presentation rate. An echoic memory interpretation of the modality effect is offered, in which it is suggested that echoic memory persists until the time of recall to be used directly and not via a long-term modality independent memory system.