The purposes of this study were to determine if test stimulus was a member of the memory set and if items in an interactive image held in short term memory (STM) could be scanned simultaneously. In experiment one, 50 university subjects compared a test word with a set of one to three words held in STM. The rate of STM search was obtained by measuring reaction time (RT) under four instructional conditions: subvocal repetition, separation imagery, interaction imagery, and sentence generation. The results indicated that RT was a linear function of memory set size in the interaction imagery condition. All other conditions also yielded a linear increase in RT with set size. In experiment two, memory set and test stimuli were both pictorial. Three groups of subjects compared items in the memory set and test stimuli. There were three experimental conditions: control--two pictured objects side by side but not interacting; provided interaction--two objects shown interacting with one another or overlapping; and subject generated interaction--subjects were given the same stimuli as the control group but were to visualize some kind of interaction between the pictured objects. The results indicated that paired associate scores in both imagery groups were superior to those in the control group. (WR)