Odor-evoked memories are well described in literature and anecdote but have yet to be experimentally characterized. The primary purpose of this study was to provide an experimentally controlled account of the nature of odor-evoked memories experienced by men and women. Two groups of 20 subjects (10 males, 10 females) were instructed to approach an odor-rating task from a subjective (emotional) or objective (analytical) orienting set. Each subject rated 20 different odors on five rating scales (pleasantness, familiarity, intensity, arousal, degree of interest), attempted to name each odor, and then indicated whether or not the odor evoked a personal memory. When a memory was evoked the subject wrote out a brief description and rated the memory on five scales (emotionality, clarity, specificity, rarity, age). The results demonstrated that, consistent with folklore, odor-evoked memories tended to be highly emotional, vivid, specific, rare and relatively old. The most familiar odors were found to evoke the greatest number of personal memories and were most often correctly named. However, 32% of the memories were evoked without an odor name, implying that memories associated with odors do not necessarily require a verbal label in order to be elicited. Females described more emotionally intense and clearer memories than did males, and females, under the subjective set, described the most vivid memories.