Subjects in an independent groups free learning experiment recalled list of low- or high-arousal words, matched for imagery and frequency and exposed randomly for 3 seconds and 9 seconds. Extrapolating neural consolidation theory to previous work on serial position effects led to the predictions that (1) arousal facilitates primacy; (2) arousal inhibits recency; (3) arousal facilitates primacy increasingly with a longer exposure time; (4) arousal inhibits recency less with a longer exposure time. It was also predicted that (5) increased exposure time facilitates primacy. Although arousal significantly facilitated recall over both exposure times, none of the predictions were supported. Several analyses suggested that the arousal effect was not a function of differential clustering within list types. The results were interpreted as consistent with the claim that word arousal facilitates recall primarily through an increase in autonomic, but not necessarily cortical, arousal. (Author)