By analyzing nodal flows of migration among cities in the 1980 population census, I found that Japanese urban systems are characterized by a Christaller-type hierarchical structure (Morikawa, 1985). I then recognized the existence of improved Pred-type structure urban systems by analyzing the functional relations between 145 head banks and their branches (Morikawa, 1994). Thus, I have considered both structures as being representative of different dimensions of Japanese urban systems to the present. It is now important to discuss whether improved Pred-type urban systems are actually developing under globalization; however, it is unavoidable to clearly explain the characteristics of both dimensions in Japanese urban systems in regard to the whole framework. Therefore, to investigate the structures of urban systems in detail (Figure 1), I analyzed not only nodal flows but also other major flows of migration. That is, the total in- and out -migration of over 3,000 people between 81 major cities each containing more than 300,000 inhabitants, including smaller prefectural capitals, as defined in volume 7 of the 1990 and 2000 population censes. As a result, I found that in 2000 the hierarchical structure crowned by Tokyo is as dominant now as it was previously and net -in-migration to the Tokyo metropolitan area is still significant, although it decreased somewhat compared with the situation in 1990. At the same time, I recognized that by analzing the functional relations between head banks and their branches, similar network structures between major cities tend to be developed far away from the Tokyo metropolitan area, as shown in Tables 3 to 9. Accordingly, I understand why urban system structures such as the improved Pred-type structure, exist within the dominant hierarchical structure. Although Sapporo, Sendai, Hiroshima and Fukuoka are usually regarded as regional capitals in which branch offices of large firms are concentrated, there are remarkable differences in the statuses of these four cities; the migration area of Hiroshima is very small compared with the other three, while Sapporo is regarded as the grand capital of Hokkaido. Although we cannot see the migration patterns of smaller cities and towns directly from census data, we can infer that they are so strongly subordinated to the neighboring prefectural and regional capitals that they, unlike major cities, cannot participate in network systems among larger cities. Such economic and cultural differences between major and smaller cities affect the economic and cultural stagnation of the latter. In comparing migration ratios of 1990, those of major cities have decreased, except for those within the Tokyo metropolitan area. Migration will depend upon the long-term depression sparked by the collapse of "bubble economy" rather than on the impact of globalization. In contrast to the Tokyo metropolitan area, the decline of the Osaka metropolitan area is remarkable. Hiroshima is also declining. I find a sign of the structuring transformation of Japanese urban systems from the decline of both cities.