The first attempt at standardization in the evaluation of rheumatoid arthritis was reported by Steinbrocker in 1949. Subsequently, Kellgren in the late 1950s, Sharp in 1971, and Larsen in 1973 reported detailed techniques for semiquantitative evaluation of radiographic changes and indicated that joint films may be the most accurate and reproducible basis for assessing the severity and progression of rheumatoid arthritis. High-resolution magnification radiography, achieved by optical magnification of fine-grain film or direct radiographic magnification, has proved particularly valuable for the reliable and sensitive evaluation of articular abnormalities in arthritis. Studies have supported the clinical usefulness of these techniques in assessing early rheumatoid arthritis. A semiquantitative radiographic approach recently used in a multicenter clinical trial is illustrated herein. Results indicate a high interobserver correlation for grades of erosion and joint-space narrowing. The sensitivity with which observers graded the findings, however, varied among observers despite the use of reference standards. There was poor correlation among observers for assessing changes in grade with time so that evaluation of clinical response was limited. Observations from this analysis provide the basis for improved methods in the design of clinical therapeutic trials.