In this study of children's memory, two types of presentation modes (pictures or words) were crossed with two levels of memory strategy ("make a picture in your mind" or "think of the word") to comprise four treatments: see picture/think picture, see picture/think word, see word/think picture, see word/think word. The study was conducted with 144 second graders across three reading levels from a suburban public school. Stimuli were 40 words chosen from the concrete nouns in the Pavio, et al. (1968) list and simple line-drawing visuals representing each word. Stimuli were slide projected onto a screen for three seconds. After each slide the screen was blank for three seconds to allow subjects to think according to instructions of the object or word. A randomly ordered and counterbalanced criterion test composed of 40 multiple-choice items--20 word items and 20 picture items--was administered two minutes after presentation and again in alternate form after three days. Results for memory strategy, presentation mode, reading level, test mode, immediate and delayed test performance, and variable interactions are presented. The data indicate that memory is better for pictures than for words in both immediate and delayed tests, but the "think picture" strategy was not effective on the delayed test. Low reading group subjects. who had used the "think word" strategy performed slightly better than "think picture" counterparts on the immediate test and substantially better on the delayed test. (Author/RH)