The purpose of the experiment was to determine whether item fluctuation (intertrial forgetting and spontaneous recovery) depends upon the time allowed for recall and/or the demand characteristics of the experimental situation (manipulated via instructions to S). One hundred. and forty-four university undergraduates studied a 36-item list and then had two successive recall trials which were 9, 36, or 144 sec in duration. On the second recall trial, Ss were instructed to recall previously recalled items only ("old"), previously nonrecalled items only ("new"), or both previously recalled items and previously nonrecalled items ("both"). The results suggest two conclusions: (1) Although recall trial duration affects the overall amount of recall, it does not affect the amount of item fluctuation; and (2) an instruction to recall only "old" items or to recall only "new" items results in suppression of the nonrequested items without a corresponding facilitation for the requested items.