Abstract:

The lack of governmental and systematic urban population data limit the study progress in urban systems in China's historical period, which has great negative effect on quantitative analysis of urban system rank-size distribution. This paper digitized the "China Postal Atlas" in 1936, based on its 13240 postal points and more than 40000 km postal routes, and examined the spatial distribution of the kernel density of postal points and routes, using the Thiessen Polygon, Zonal Statistics and other GIS tools to give value to the postal capacity T of modern Chinese cities. T-value is used as a proxy for city size in modern China to analyze the rank-size rule. The conclusions are as follows. (1) The distribution of rank-size of cities in modern China based on the T-value was in accordance with Zipf's law. (2) In 1936, there were five postal aggregation areas in China: Beijing-Tianjin-Hebei region, the Yangtze River Delta, the Pearl River Delta, Wuhan and Chengdu-Baxian region, which is highly consistent with the current distribution of China's urban agglomerations. (3) In 1936, the spatial distribution of postal points and routes was very different on both sides of the "Hu Huangyong Line". The northwestern half to the line had a slightly smaller proportion of points than that of population, but a much higher proportion of high-grade points and routes than that of population. (4) In this paper, the sample size of urban system in modern China increased from about 200 (urban population) to 2360 (T-value). For the first time, the "three-stage" distribution caused by the unique "depression in the middle section" of the city size curve was found, which was quite different from that of the United States and China today. (5) The postal capacity T-value can be considered as a proxy for city size in urban geography studies in modern China.

Key words: China Postal Atlas, urban system, rank-size rule, urban population, Hu Huanyong Line, China