Seven paired-associates were constructed using words (for which scales values on arousal were derived by paired-comparison technique) as stimulus terms and digits (two through eight) as response terms. Forty subjects were randomly assigned to one of four conditions-cued or free recall and short or long-term tests following a single learning trial. Recall of pairs was very poor in all conditions. That level of arousal of words aids in the learning of the pairs was not substantiated by an analysis of the free-recall data on recall of the pairs, the stimulus terms, and the response terms. (Author)