Recently there have been many studies about the structural change of the metropolitan area. The first approach to the theme was that from the view of counterurbanization in the United States. But in this study the districts which showed the highest rate of population growth were not remote rural areas but the nonmetropolitan areas adjacent to SMSAs. Then P. Gorden (1979) or P. Hall & D. Hay (1980) disputed that the counterurbanization could be explained in terms of the suburbanization. They found a problem with the structure of metropolitan areas in the study of national scale migration named ‘counterurbanization’.
In the four stages of urban development, which Klaassen et al. (1979, 1980) showed, the critical point between the second stage, suburbanization, and the third new stage, desurbanization, was whether the population of the whole metropolitan area increases or decreases. It was the same condition as counterurbanization. When the most rapidly growing district of population then is an adjacent district to the metropolitan area, we can no longer grasp the urban growth within the framework of the metropolitan area. This population decreasing process of the whole metropolitan area is first explained by K. O'Conner (1980) and Y. Taguti (1981) as follows: The new urbanizing zone adjacent to the metropolitan area, exurb, in which more commuters work in the suburbs, does not belong to the metropolitan area defined as a commuting area to the central city. Then, if the population of the adjacent districts increases, the metropolitan area does not expand there as before.
The structural change of the metropolitan area has also been analyzed in terms of the suburbanization of economic activities of the central city as by Muller, et al. In this point of view the population suburbanizes first, then manufacturing and retail activities of daily food necessities. In the third stage, large shopping centers are constructed in the suburbs and offices gravitate to them. However the question is the suburbanization of the decision making sector in offices. If that sector remains in the CBD as P. W. Daniels (1974) shows, the regional structure of economic activities in the metropolitan area never change. Economically it is the nodal region which has an apparent node CBD. But in terms of daily behavior of the people, not many people need to go to the CBD or central city frequently. It is an other region than an economic nodal region. Strictly speaking the metropolitan area now is the daily behavior area defined by commuting or shopping to the central city. It is the daily behavior in the area of the central city and around it that is changing now not only in the U. S., but in Japan and European countries.
In Japan we don't yet find a decrease of metropolitan area population. However the suburbanization of the central city activities are of course under way just as R. A. Erickson (1983) and A. Kellerman (1985) have shown for the metropolitan areas in the United States of the 1950's. In the 1960's city centers rapidly declined in America though we can't conclude whether they are in an advanced stage of urban development or in circumstances peculiar to the metropolitan areas of the United Sates. But the suburbanization of employment now in Japan makes the commuting rate to the central city decrease and it means that factors are increasing which cannot be explained within the framework of the metropolitan area. We need a new framework which can explain those factors.
We can cite two types given by E.J. Taaffe (1963) about the new structure of the region in order to understand the metropolitan area hitherto. One type is the ‘concentration of the peripheral laborsheds’ and the other is the ‘dispersion’ of these.
