Processing complex multivariate information effectively when relational properties of information sub-groups are ambiguous is difficult for man and man-machine systems. However, the information processing task is made easier through code study, cybernetic planning, and accurate display mechanisms. An exploratory laboratory study designed for the University of Missouri School of Medicine and the Department of Electrical Engineering, after pretesting by the Department of Information Science, sought feasible coding strategies for displaying multivariate biochemical data gathered on a set of patients. The coding strategies served as "facilitators" of the human perceptual process to permit an accurage placement of the patients into two groups--normal and diseased. Geometrical designs successfully functioned as "codes" for a variety of body chemistry states. Users were able to interpret the visually displayed designs because of a natural human facility for pattern discrimination and recognition. The results of this laboratory test encourage further work of this kind. (CH)