Among many of the climatic classifications of the world, the Köppen's is one of the most useful and widely adopted one for many practical purposes, although it has been about thirty years since the publications of his climatic map of the world (Klimakarte der Erde, 1928, Gotha). However, this does not mean that his scheme is completely satisfactory and reasonable, because it contains many differences with actual distribution of geographical phenomena in the light of more recent knowledge. Aside from methodological and theoretical considerations, we can find the following two important weak points in the Köppen map. (A) Actual wrong position of the boundary lines due to the inadequateness of reliable climatic data. Especially in Asia and Africa, the net of the meteorological observation was very thin, in addition, the length of observation period was short, so that the placing of climatic boundaries was not exact. (B) Unsatisfactory choice of numerical value limits between two climates. This results from the fact that Keippen built his system mainly on the information from the northern hemisphere and from Europe. For example, the D-E boundary corresponding to the tree-limit given by the 10°C isothermal line of the warmest month, is fairly consistent with the actual features in northeastern Europe and northern Siberia ; it does agree as well in northern Canada and there is much deviation in the high latitude region of South America. The same difficulty is experienced in the b-c boundary. Furthermore, the division of C climate (Cf, Cw and Cs) is fairly satisfactory in Europe and Western America, but it is not acceptable in the eastern side of Asia and N. America. In the present paper, the study is limited to the first item (A) and I have reconstructed the Köppen map of Japan, Korea, China, Indian Peninsula and Indonesia, using the new available data, but not changing the Köppen's numerical values. As for item (B), discussion will appear in the “Geographical Review of Japan” in the near future. In brief, the changes are as follows : Japan : Japan proper is altered very little even with new data. In Northern Japan, D climate appears at a relatively low altitude, the Mizusawa-Morioka Plain and a part of the Aomori Plain belong to D, rather than C climate. The greater part of the Japanese mainland is Cf a, but Dr. Sekiguchi found also that Cwa occupies small separate areas mainly on the Pacific side of Honsyu, Shikoku and Kyusyu Islands. Hokkaido consists of D climate excepting the southwest periphery of the island, but it is divided into Dfa, Dfb and Dfc (Fig. 1.), instead of Dfb only as on the original map. Sakhalin : no change Taiwan (Formosa) : A-C line is shifted somewhat to the north, and A is divided into Aw and Af instead of Aw only, as shown in Fig. 2. Korea : C-D line is moved southward in the inland region, and D is divided into Dfa, Dwa and Dwb instead of Dwa only, as shown in Fig. 3 as compared with the original map. China proper and Manchuria : Modification is greatest in this region. Cwa is limited to a narrow strip and the former wide region is replaced by Cfa. Also BS climate spreads far south as compared with Köppen, reaching to the northern side of Shantung Peninsula. Hence the Dw climate in Northern China is restricted to the very small area near Peking-Tientsin district. In Manchuria, however, Dwa is the typical climate, and only the western part is occupied by BS climate as shown in Fig. 4. Indian Peninsula : BS climate on the Deccan Plateau has been enlarged (Fig. 5.) and this may have some geographical significance for India. Other regions show only small changes from the original map. Indonesia : A considerable area of Am climate is interposed between Af and Aw in Java Island as shown in Fig. 6.
Gozoku-yashiki-mura is a village type originated from the premises of local powerful families in the southern part of Tohoku district. The origin of the older ones of them goes back to Heian period (8-12 centuries), and in some places the sites of residence of still earlier period are well preserved. Many of the rulers of early days who were the heads of the mansions became the farmers later, and settled in this part of the country. Their ways of living are to be surmised by the remaining earthen walls, moats surrounding the mansions, family shrines and the like. The question is why this kind of powerful families could have survived through long ages in this particular district like a cultural fossil. It is partly explained by the remoteness of Tohoku district which enabled them not to be too much influenced by the changes in the center. It is also because the feudal lords of later period adopted the policy to take advantage of their influence over local people, so long as it was not against their own interests. We can surmise the development of rural settlement here, through the analysis of present aspects of such premises. These are the grounds some 55 X 55 meters large, surrounded by moats and earthen walls. Some of them are even larger, and near them there lived their hereditary retainers called Nago or Fudai. Besides the historical reasons, here to be noticed is the existence of conditions for the preservation of such a way of living. The general nature of landform here is a very gentle slope, which developed some local habits different from most parts of this country. Such are the habit of using stream water for drinking, and that to let the rubbish flow into the fields as fertilizer, by means of irrigation canals. These habits all put together, helped to make the disseminated village an economically favorable type of settlement.