In four experiments, young (18-26 years, M = 21) and elderly (over 65 years, M = 72) people were compared for recognition memory of (a) graphic stimuli (faces of presidents and vice presidents, engineering symbols, and free forms) and (b) everyday odors. On graphic stimuli, the elderly consistently matched the young, but on odors the performance of the elderly was worse. Their poorer olfactory performance was observed after only 26 s, but became truly marked after 1 hr or more. Somewhere between 1 hr and 2 weeks, their odor performance fell to chance, but their graphic performance remained well above chance. Although the young did forget both graphic and odor materials progressively, their performance always stayed above chance over a 6-month period. Experiments 1 and 2 revealed that the elderly are less sensitive to odors than the young (with thresholds about 10-fold higher), which may explain, in part, their poorer olfactory memory performance. Knowledge that the subjects brought to the tasks by way of familiarity with and ability to name odors and faces played a positive role in recognition memory. Because of this positive role, together with the negative role played by verbal distraction, we conclude that odor recognition memory depends, perhaps heavily, on semantic processing. Impaired semantic processing may result even when odors are simply rendered desaturated, or pastel because of the weakening of olfactory sensitivity with aging.