Assessing the effect of memory improvement strategies upon long-term forgetting, two studies investigated the influence of popular mnemonic devices (the keyword method and the "method of loci") upon forgetting relative to rote rehearsal. The first study (79 subjects) compared the forgetting of French vocabulary words learned either by using the keyword technique or by rote rehearsal. In the second study (76 subjects) the "method of loci" was compared to rote rehearsal for serial recall of unrelated word lists. Results indicated that subjects using mnemonic devices forgot at a faster rate than subjects rote rehearsing the same information. Thus, contrary to widely held expectations, mnemonic devices do not appear to confer any long-term advantage to the retention of material so learned. Perhaps the deleterious effect of mnemonic devices upon memory is due to the interference that may accompany use of an unfamiliar learning strategy. This suggests that training in the use of mnemonic devices should be of an extended and distributed nature so that potential interference effects can be minimized. (Two figures are included.) (SR)