The main purpose of this article is to analyze the process of demographic urbanization during the Meiji and Taisho periods in the Tokai area, central Japan, using "Kokoh-Chosa", the population statistics which were available before modern censuses began. The study period covers the years 1885-1920. In those times, Japan experienced a rapid population increase and an important urbanization process never seen before. The reason for these changes is attributed to modernization, and many researchers have been interested in such vicissitudes.
Although population changes in modern times have been considered to be as a result of a demographic transition, some historical demographers in Japan argue that this started around 1920. They insist that the beginnings of population increase trends were derived not from decreasing mortality but from upward fertility trends, so that the demographic transition occurred much later than the start of population growth trends. Moreover, in such processes there seemed to be some differences between cities and peripheral areas, and such regional differences are associated with urbanization processes.
Geographers have attempted to portray the processes of urbanization in those times in terms of social and economic indices. However, only a few studies have focused on the processes of Japanese demographic urbanization since "Kokoh-Chosa" possesses some significant problems as discussed below. In this paper, the problems of "Kokoh-Chosa" will first be made clear, and then there will be an attempt to adjust the data. "Kokoh-Chosa" gives us demographic information over 50 years before modern censuses, and it includes much valuable data, such as the numbers of births and deaths in all cities and counties. Thus, by adjusting the data and then using it, discussions on the processes of demographic urbanization become very fruitful.
The most serious problem of "Kokoh-Chosa" was that the population was overestimated, especially in city areas, and that this accumulated from year to year. This problem resulted from the population registration system of those days which was defective due to its complexity. While the population of "Kokoh-Chosa" was overestimated, this problem was recognized in those days and thus the excess population was sometimes removed. That kind of process was referred to as "Kiryu-Seiri". As a result, there were sometimes some sudden gaps in the population data series. The author solves the problem of Kiryu-seiri by calculating the excess population.
Using the adjusted data, the author firstly analyzes the changes in crude birth rates and death rates in cities and counties. This has much to do with the discussion about demographic transition processes. Secondly, urbanization processes are analyzed by means of several maps showing the population change every five years. In particular, the author separates total population growth into its natural and social components, and discusses the geographical differences in population dynamics.
The results of the discussions are summarized as follows.
Population increases, which may cause demographic urbanization, were due to high fertility, mainly in county areas. On the other hand, there are few changes in mortality trends in the study period with a few exceptions such as the mortality crisis in 1918. Birth rates were higher than death rates at almost all times in all regions, but natural increase rates in cities were lower than those in county areas. Thus, we can say that demographic urbanization was due to many migrants from counties to cities.
In the late 19th century, demographic urbanization had already appeared, but the net-migrants were mainly seen in the cities and their neighbors, while there were very few net-migrants in peripheral counties. Since the 20th century
