Multidimensional scaling (MDS) has been used in a wide variety of research fields; psychology, political science, anthropology, marketing research, urban and regional planning, and so on. In practical terms, MDS is a statistical method to make a picture of the information in the data. It enables us to examine the "hidden structure" of a set of data. When the set of data is large, MDS is extremely useful, since it is easier and more informative to look at a picture than the data themselves. In this paper, MDS is applied to the interregional migration data of Japanese regions for the years 1960-85. Findings show that the two-dimensional configuration of regions estimated by MDS generally corresponds with the geographical locations of regions, and the structure of interregional migration was very stable over the study years 1960-85. It is also suggested that MDS is a useful tool to identify the relationships between regions using the spatial interaction data.