Three experiments were designed to decide whether temporal information is coded more accurately for intervals defined by auditory events or for those defined by visual events. In the first experiment, the irregular-list technique was used, in which a short list of items was presented, the items all separated by different interstimulus intervals. Following presentation, the subject was given three items from the list, in their correct serial order, and was asked to judge the relative interstimulus intervals. Performance was indistinguishable whether the items were presented auditorily or visually. In the second experiment, two unfilled intervals were defined by three nonverbal signals in either the auditory or the visual modality. After delays of 0, 9, or 18 sec (the latter two filled with distractor activity), the subjects were directed to make a verbal estimate of the length of one ofthe two intervals, which ranged from 1 to 4 sec and from 10 to 13 sec. Again, performance was not dependent on the modality of the time markers. The results of Experiment 3, which was procedurally similar to Experiment 2 but with filled rather than empty intervals, showed significant modality differences in one measure only. Within the range of intervals employed in the present study, our results provide, at best, only modest support for theories that predict more accurate temporal coding in memory for auditory, rather than visual, stimulus presentation.The recency effect in memory captures one of the most conspicuous laws of memory and metamemory: we remember well what just happened and less well what happened in the more remote past. Currently, one prominent theoretical approach to recency is founded on the proposition that items from the end of a series have a privileged status because of their discriminability along the temporal dimension (Baddeley & Hitch, 1977;Hitch, 1985;Murdock, 1960). Glenberg (1987;Glenberg & Swanson, 1986) has set out a particularly well-formulated version of this idea. His temporal distinctiveness theory deals successfully with several aspects of recency including, notably, the long-term recency effect of Bjork and Whitten (1974) and its dependence on the relative spacing of list items to each other and the time lapse prior to recall.Briefly, the central proposition of Glenberg's (1987;Glenberg & Swanson, 1986) retrieval-based distinctiveness theory is that encoding of new information includes a description of the time of occurrence of the new information. Retrieval proceeds by means of temporally defined search sets when more efficacious retrieval cues (i.e., semantic cues) are not available. The size of a given search set is a function of how long ago the information was presented: temporally more distant events are asThis research was supported by NSF Grant BNS 92-19661 to Robert G. Crowder. We wish to thank Arthur M. Glenberg, Alice F. Healy, Alee Hellstrom, and an anonymous reviewer for their thoughtful comments on an earlier version of this manuscript. Correspondence regarding this article should ...