Regional difference and variability of precipitation during the Baiu season (June and July) in East Asia have been studied preliminary in terms of the frequency distribution of daily precipitation and the precipitation brought about in each precipitation intensity class determined empirically. In order to demonstrate the spatial distribution of dominant daily precipitation amount (class), cluster analysis was conducted using the contribution ratios of the precipitation brought about in respective classes to total precipitation at each station using the data from 1977 to 1981. The 234 stations were classified into eight groups. Regional division was conducted based on this classification of stations. South-west Japan, Korean Peninsula and Central China Plane where the Baiu phenomenon is predominant were identified as same group. In these regions, considerable amount of total precipitation is brought about by heavy rainfalls greater than 50 mm/day. The class of maximum contribution ratio shifts heavy rainfall side with increasing of regional averaged total precipitation. Nevertheless, the class of the miner peak appears in 15-30 mm/day independently with the total precipitation amount. This fact suggests the existence of the rainfall system that causes almost the same precipitation to the comparatively rainy regions as well as the peculiar rainfall system to the region that controls total precipitation. Moreover, relationship between frequency distribution of daily precipitation and total precipitation was analyzed preliminary in terms of inter-annual variation using the data from 1951 to 1990 at Shionomisaki in the South-west Japan and Akita in North-east Japan. The result indicates that the variation of Baiu precipitation is caused not orLly by the simple variation of number of rairLfail days but variation of proportion of heavy rainfall days to total rainfall days and whether extremely heavy rainfall appears or not, corresponding to the rank of total rainfall amount. This suggests that the wide range variation in precipitation consists of multiple backgrounds that control precipitation variations of respective ranges.
Up to this time, most of studies on children's environment perception have not taken account of some factors and phenomena in modern society preventing their development of perception of the environment. Moreover, these studies have taken only one district in one region as a study area and generalized the result, which has been made clear, as a typical case in this region. But, this case is not always applied to all regions, for one region, such as seen in a large-scale city, is divided into some districts which have different features respectively. Standing on these points, the author put the purpose of this paper on next three points. 1) To try to generalize the feature and the developmental process of children's perception of the environment in the neighborhood based on perceived amount, perceived space and perceived structure by using the analysis of their free-hand sketch-map. 2) To reveal the mechanism of development in environment perception by observing children's behavior in their living space basing on "the path of daily activity". 3) Taking account of the above two mentioned points, to try to grasp today's social and physical environment surrounding children. As a study area, the author chose Hiroshima City. This city consists of many districts which have differences in landscape and function because of city's enlarging accompanied with the affiliation of towns and villages around city. So, in studying, he divided this city into urban district, suburban district, moutain village district and island district. The main results are summarized as follows: 1) In urban and island districts, children's perceived amount increases with the advance of grade. On the other hand, in suburban and mountain village districts, their perceived amount decreases at the peak of forth grade. This is the reason why the simplification is recognized in their perception. 2) Children's perception of the environment is affected strongly by feature of the region where they live and their behavior in their living space. 3) In urban and suburban districts, children's perceived space enlarges, with the advance of grade, only in the narrow area of each school district, so their perceived space in these districts does not develop remarkably. In moutain village and island district, their perceived space spreads only around their home and reduces, in spite of the advance of grade, so their perceived space in these districts stops developping. 4) In urban and suburban districts, the neighborhood for children is limited in the narrow area of school district. In moutain village and island district, it is limited around their home. 5) In urban district, children's perceived structure develops with the advance of grade, however, in suburban district, it stops developping at the peak of forth grade. In moutain village and island district, the development of their perceived structure is unbalanced, so it dose not develop. 6) From the analysis of the path of daily activity, it is assumed that today's many social factors, such as school, cram school which is called "juku", mass media and club activity in school curricurum prevented children's daily activity, and that these preventions stopped their development of perception of the environment, in spite of the advance of grade. Finally, as a view in this study for the future, the author thinks that it is more useful to use the conception of time geography in order to analyse the prevention accompanied with behavior in detail, and to apply the results of this paper to improvement of geographical education. Furthermore, it is required that the way of free-hand sketch-map to analyse children's perception should be improved, because their sketch-map is filled with each children's peculiar character.