This paper investigates the residential structure of Tokyo, the largest city in Japan, in the 1910s middle period of Taisho by applying factorial ecological method. The principal factor method is applied to the original data matrix of 1920, which contains 816 districts in columns and 19 variables in rows. Six common factors whose eigen-values exceed 1.0 are obtained. The first factor is interpreted as the family status and the second factor as the public officials and liberal workers. It becomes clear that the fundamental residential structure of Tokyo in this period is composed of five residential groups; the large household of commercial worker, the unmarried persons who work in large stores, the medium-size household of manufacturing workers in the prime of life consisting of married couple and their children, various manufacturing workers and public officials and liberal workers. Consequently, it is clear that the residential structure of Tokyo in the middle period of Taisho cannot be explained by the simple division which divides Tokyo into eastern and western areas as considered in previous studies, but is at the stage of the residential structure of the industrializing city.