This study investigated the position of illustrations and how they affect the sequential organization of text. The primary purpose was to describe learners' reading sequence and preference for page designs with variations of illustration placement. The 70 subjects were undergraduate and graduate students enrolled in education courses at Memphis State University. Each subject was given a set of stapled pages containing eight randomly-sequenced page designs, each separated by a response sheet. For the "preferred" or natural reading sequence of the subjects to be determined, meaningless text which had the same structure and appearance as realistic text was used. The desktop publishing-generated text was printed in a two-column format for two of the designs and in a three-column format for the other six. In each design, the illustration cut across at least two columns. In addition, a two-point horizontal bar was inserted above the illustration in four of the designs. For each design subjects were asked to trace, with a pencil, the sequence with which they would naturally read the text and to indicate their reactions to the design on a five-point scale. A two-column format combined with illustrations dividing the columns was preferred over a three-column format incorporating an illustration dividing all columns. Results imply that illustration placement is a critical attribute in the design of instructional text. (5 references) (GL)