This paper describes a study in which a Picture Readability Index (PRI) was used to investigate initial and extended perceptions of photographs. Readability criteria for evaluating instructional text seems to have been in place for a long time, yet instructional visuals like photographs and illustrations have typically been subject to no such criteria. The PRI, developed from research in the areas of semiotics, linguistics, perception, visual literacy, and cognitive psychology, seeks to apply measurable readability criteria to these visuals. It considers first impressions gained from brief exposure, and also examines how a picture and caption are processed together during prolonged exposure. Data is coded and entered onto a nomograph for comparison between affective and cognitive domain classifications. The paper-and-pencil version of the PRI test, that was administered to a small group of students, however, is somewhat limited by its length and complexity, which may cause fatigue to have an influence on responses. Thus researchers set out to adapt the PRI test for the computer; besides eliminating the fatigue factor, the computer-assisted version would also organize data and simplify the process of projecting the image for viewing by the respondent. Adapting the PRI test into a computer-administered format involved selecting an authoring tool, flowcharting to modularize the instrument, storyboarding, scanning in photographs and drawings, creating test questions, and constructing and revising the multimedia prototype. Future study may include a computerized-adaptive version of the PRI, in which the progression of questions on the monitor will depend on prior responses. (Contains 27 references.) (BEW)