Abstract In this paper, new quantitative methods were applied to elucidate the geographical distribution of Japanese dialects. (1) Regional differences (individual differences) of usage rate of standard language at each locality in GAJ (Grammar Atlas of Japanese Dialects) were dealt with. (2) In addition to the railway distance from Tokyo to each locality, the walking distance from the railway stations to the survey localities was measured and used as the “railway-walking distance”. (3) Using a scatter-plot in what is called a “limestone cave model”, the values on the axis of standard language usage were reversed, and the Japan Islands were divided into east and west according to their direction from Tokyo. It showed that the standard language usage rate and distance show a proportional relationship even at each locality (individual informant) value. (4) Focusing on the prefectures near Tokyo, the distance from Tokyo regulates the usage rate. From this, the rapid spread of standard language from modern Tokyo and influence of railway can be ascertained. It was found that the phenomena pointed out by the average values by prefectures of vocabulary items in LAJ are valid to the data for each locality (individual) of GAJ.