We recently demonstrated in human subjects that muscle-tension-induced arousal can enhance later retention performance and that this effect is attenuated by beta-adrenergic receptor antagonists. In that study, each subject established a baseline for muscle tension by squeezing a hand dynamometer for 30 s with maximum force. This may have served to "prime" subsequent arousal produced by muscle tension. Two experiments were performed to address this issue. At the beginning of each experiment, young adult subjects were asked to squeeze the hand dynamometer at maximum effort either for 30 s (Prime) or for only 1 s (No prime). Then, during the task, arousal was induced by having each subject exert a moderate amount of tension (25 to 50% of baseline maximum). In the first experiment, subjects were shown four consecutive lists of 20 highly imageable nouns, given immediate recall tests of each, and then given a comprehensive recall and recognition test at the conclusion of the experiment. Moderate arousal was induced once for each list (at encoding, consolidation, or retrieval) or not at all for one list. The sequence of arousal induction was counterbalanced. Significant enhancement of delayed recall was seen in the 30-s group for those lists in which arousal was induced during the consolidation or retrieval period with no significant effects in the 1-s group. These results demonstrate that arousal can modulate memory consolidation when induced shortly after learning and that an initial priming event may affect the response to subsequent similar arousing events. In the second experiment, subjects read paragraphs, some of which contained highlighted words (working memory task); half of the subjects were given the 30-s procedure and half the 1-s procedure. Only those subjects in the 30-s group showed significant arousal-induced enhancement of delayed recognition of the highlighted words. Again, no significant effect on retention performance was seen in the group that squeezed the hand dynamometer for only 1 s during the priming period. Pulse data suggested that there may be somewhat greater heart-rate reactivity in the 30-s group. These findings suggest that memory modulation by arousal may be primed, or enhanced, by a relevant preliminary arousal event.