In order to study the local distribution of winds in the Miyako Island, one of the Ryukyu Islands, Southwest Japan, an attempt was made to observe the wind-shaped trees of Casuarina equisetifolia as an indcator of prevailing wind conditions. According to the climatological values observed at the Miyakojima meteorological observatory located at a point of 24°47'N, 125°I7'E, 39m a. s.l. , NE-NNE winds prevail from October to March. These winds are relatively stronger than the winds from May to August. Namely, monthly mean wind velocity is 6.5m/s in the cold season, and 5.3m/s in August. S-SSW winds are striking from May to July. The wind-shaped trees were observed at the 83 points in the island with an area of 159km^2 and an average of 3-4 trees was obtained at each point. Fig. 1 illustrates the scale of wind-shaped trees, which is determined in reference to the previous studies in the other regions. The result is shown in Fig. 2. Important facts revealed in this figure are: 1) The coastal region with a width of 500m-1,000m facing NE has the prevailing NE-ENE winds. On the other hand, there is also the ENE wind region at the south-western part of the island. Between these two regions, SSE-SE winds develop in a region with a width of 8km. They become S wind near the Hirara City. 2) Comparing the observed result of wind-shaped trees with those by the instruments at the meteorological observatory, the wind-shaped trees of Casuarina equisetifolia indicate the wind conditions during the warmer season, especially from August to September. The summer conditions of monthly mean wind velocity 5m/s with monthly mean temperatures of 27-28℃ overcome the winter conditions of monthly mean wind velocity 6.5m/s with monthly mean temperatures of 17-19℃ in the tree growing forms. 3) It is supposed that the winds are weaker in summer and stronger in winter in the ENE-NE wind regions than in the region near the meteorological observatory. 4) The coastal region facing NE has strong ENE-NE winds, because the north-easterly winter monsoon hits directly this coast. On the other hand, the other ENE-NE wind region at the south-western part of the island is considered to de appeared as strong falling lee winds crossing over the low hills with a height of 80-100m a. s. l. 5) An interesting phenomena of "shrub dunes" composed of Robinia pseudo-acacia \vas observed at the southeastern part of the islands, as shown in Fig. 3. Above the "shrub dune", the wind-shaped trees of Casuarina equisetifolia were seen. The upper part of the dune and the Casuarina trees do show smooth stream lines. The cross-section of the "shrub cune" resembles to that of barkhan-type sand dune.
The purpose of this paper is to clarify the aspects of the potential pattern of intercity associations in Korean cities in 1960 and 1975, with a view to identify the degree to which the cities do indeed operate as a system. The magnitude of direct associations and accessibilities among cities are measured by the gravity model in terms of railroad time distance and railroad distance among 35 cities. Basing on the results of direct associations among cities by the gravity model, the volume of indirect associations among cities are estimated by the largest flow. But in order to understand the underliyng potential associations, the writer also tried to use the second largest flow among cities. The main results are summarized as follows: 1. Through an analysis of the accessibilities of the 15 year period, it was proved that the promption of time distance has increased accessibilities of large center and their satellite cities to the detriment of medium and small sized centers. The result is that the promption of accessibilities has been concentrated in Seoul and the larger centers. But on the other hand, the function of the medium and small sized centers as nodal points is gradually on the decline. As expected, the accessibilities of cities increase in proportion to their population size (Figs. 1a-2). 2. The potential interaction pattern among cities on the basis of time distance and railroad distance has changed during the 15 years. In 1960, the largest flow among cities has revealed a pattern concentrating in Seoul, Busan, and Daegu. Meanwhile, in 1975, there seems a strong tendency in Seoul only. This means that with the decrease in time distance, the sphere of one day's journey is made possible in Korea, in consequence, Seoul has come to dominate the interactions with its primacy. The pattern of the second largest flow has also changed as It was found in the largest flow. In 1960, Seoul, Incheon, and Busan had divided the pattern of interactions in Korean urban system into two parts. Meanwhile, in 1975, these patterns have come to be concentrated in Incheon and Busan. As a consequence of the fact that Seoul has dominated the pattern of interactions among cities in 1975, accordingly, Busan and Incheon have come to dominate the underlying pattern of interaction among cities. When Seoul was excluded from the potential interaction among cities, Incheon develops a strong association with cities in the Joong Bu region. This means that the high primacy rate of Seoul gives rise to its prominance in interaction with all cities of Korea (Figs. 3-6). The potential pattern among cities on the basis of railroad distance is similar to the pattern as in the analysis of time distance. The difference between the results of the two datas have revealed a somewhat different interaction pattern such as the association of the cities in Jeonlanam Do, Jeonlabuk Do, and Gangweon Do with Incheon and Busan. These tendencies may be related to the change of transportation system during the 15 years (Figs. 7-10). 3. Ward's cluster analysis on the basis of time distance in 1960 and 1975 has indicated the change of groups during the 15 years. This is found that Seoul, Incheon, satellites of Seoul and industrial cities have increased their accessibilities and interactions among cities during the 15 years. In contrast, the regional administrative centers and the small sized centers have shown a pattern different from that mentioned above. An examination of the linkage tree suggests these changes well. The regional administrative centers and the satellite cities of Seoul formed one group in 1960, but the regional administrative centers and small sized centers are clustered at the earlier stages of grouping before being combined into one group in 1975 (Fig. 11). This is only a result of potential pattern among cities in Korean urban system. Through the analysis mentioned above, it is estimated that
The political foundation in ancient Aki (now western part of Hiroshima prefecture) was formed by the local dominant clans. The first clan, called Aki-no-kuni-no-mryatsuko (安芸国造) dominanated old family s called Amabe (海部) "Lebensraum". The second clan was the so-called Sae(i)ki-be (左伯部) who tried to maintain control over the Setouchi Inland Sea and restrain Kibi(吉備)'s power in the east, and was dispositioned by the central noble Ohtomo (大伴). Lastly, Ohshi-no-atai (凡直) clan executed regional government to follow the above territory at Ritsuryo^- (律令) era (A.D. 8th century). With reference to the above historical situation, this paper attempts to analyze the location, structure and planning of the provincial offices -Saiki-gunga (左伯郡衙), Aki-gunga (安芸郡衙) and Aki-kolufu (安芸国府)- by the cadastral research on the geomorphological survey. 1) Saiki-gunga was located in the lower parts of the alluvial fan formed by the Yahata-gawa river. This fan was deposited after the Yayoi era (before Ritsuryo^- era) and remained the Jo^-ri (条理) footpath. The ancient seashore probably had a 10 meter contour line of today, because the archaeological sites located in Yayoi era had a 20 meter line, the Jo^-ri remains do not exist the south part (near the Ko^-ri-bashi (郡橋) bridge) and the south area was developed after middle ages. On this part remain the place names of Ko^-ri-chimei-gun (郡地名郡) as Gunga, divided 4 cho^-s (丁) square following better 40 kens (間) partition (N32°E) than 60 kens (=106 meters) on the Jo^-ri system. 2) Aki-gunga is likely to have appeared at Saiki-gunga. The archaeological sites in this area situat on the alluvial fan formed by the Misoo-gawa river. This area is comprehended two blocks of settlements by the cadastral maps at Meiji era. The former one on the south part area was divided 3 cho^-s square following 60 kens (N20°W) - or maybe 4 cho^-s - on Jo^-ri system as Gunga for the place name of Ohtoshi (大年) as the warehouse. The latter one was divided 4 cho^-s square following 40 kens partition (N5°W), because a few wells on the spring belt and the place names in reference to Kokufu, for example Kokuchyo^- (国丁), Tadokoro-nawate (田所縄手) and So^-ia (総社). Finally, the Gunga's partition was based on the Jo^-ri system. After, Kokufu's one was enforced differ from basic direction. Both offices were probably completed at Ritsuryo^- era. Their location showed a Nodal function between the inland traffic system and the marine transportation system based on the above political background.