In Experiment 1, frequency-discrimination thresholds were estimated in a 2-interval, forced-choice, backward masking procedure with a masker acoustically dissimilar to the targets. Young subjects were more efficient in escaping the effects of masking than were their elderly counterparts. In Experiment 2, young and elderly subjects performed the same task, with a masker acoustically similar to the targets and with a target-dissimilar masker. Under target-similar masking and at short target-masker intervals, the elderly demonstrated significant improvement, reaching the level of performance of the young, whereas under the target-dissimilar masker, the age-related differences were restored. Both age-related slowing of information processing and increase in stimulus persistence can account for the results of Experiment 1, but only increased stimulus persistence explains the results of Experiment 2.