In recent years there has been a good deal of debate about the role of rote, repetitive rehearsal (called Type I or maintenance rehearsal) on the establishment of memory traces that outlast the rehearsal process itself. One advance in the technology used to study this problem is the operational definition of maintenance rehearsal proposed by Glenberg and Adams (1978, Journal of Verbal Learning and Verbal Behavior, 17, 455-463). These authors argued that maintenance rehearsal should be defined as the continuous maintenance of information in memory using minimal cognitive capacity. Here this definition was adopted and extended in a paradigm in which the mental resources devoted to maintenance rehearsal could be systematically varied. The experiment revealed that there is, indeed, an effect of maintenance rehearsal on long-term recognition performance and that this effect depends on the mental resources devoted to the rehearsal process. The functions of rehearsal are two in number: to maintain information in a temporarily active state during short-term tasks and to create memory traces with some permanence. The latter of these functions has often been attributed to the class of processes called elaborative rehearsal, processes such as chunking, forming images, or recoding material in various ways. However, various sources of evidence have been accumulating to suggest that maintenance rehearsal--mainly rote repetitionmay also play some role in creating longer term memories. If this finding is correct, it is significant for two reasons. From a practical point of view, rote repetition is the strategy that is probably used most frequently as a mnemonic, so it is important