An analysis of the diffusion process of information or of pervading fashionable custom may make it possible to infer the regional system as a diffusion channel. From this viewpoint the writer attempted to explain how a regional system had operated in the Okagemairi diffusion in the year of 1771. Though the custom of visiting the Ise Shrines once in one's life had been widespread among the Japanese during the Fdo era (1603-1867), under the feudal system, it was impossible for wives and children, or employees to visit without participating in the Okagemairi, a phenomenon of group pilgrimage. A custom of Okagemairi usually began with rumors about the falling of charms of the Ise Shrines from heavens as a start, and people along the route offered food and money or free lodging to vistiors. As visitors consequently could travel without money, many people participated in the Okage-mairi within a short period.
From Mr. Kosen Mori's diary, who lived at Matsuzaka in Ise province, we come to know the date when the first participants from each province passed through Matsuzaka, as shown in the 4th column of Table I. The writer determined the date when participants started from province, as shown in the 5th column o Table 1, assuming that they walked 30km. per day along the shortest path shown in Fig. 2. He assumed this date as that of the diffusion, or as the date when the Okagemairi spread to each province.
Trend surface analysis of the date of diffusion revealed the following points (Fig. 5) (1) The fact that the explained variance of quadratic trend surface is 39.26% suggest: that the Okagemairi contagiously spread from Yamashiro, the starting point of diffusion. (2) The fact that the diffusion to Musashi and Settsu was much earlier than the general trend, and the fact that the diffusion to Tajima, Inaba, Mimasaka and Hoki was much later than the general trend, suggest that the population size effect was partially in operation (Fig. 6).
(3) As there is almost no correlation between residual and population size (γ-0.1959), the determinant to warp the ellipse-like diffusion pattern seems to be a short-circuit effect through sea routes rather than the population size effect. This point is suggested by the fact that the diffusion to Nagato and its surrounding provinces, where the westward Liner linking the Hokuriku region with Osaka via the Inland Sea of Seto, the Setouchi Liner and the Kyushu Liner stopped, was much earlier than the general trend, and also by the fact that the diffusion to Ecchu and Echigo in the Hokuriku region, which had been connected with Osaka and Kyoto by ship before the Edo era, was much earlier than the general trend.
(4) The fact that the Okagemairi did not spread to Noto, Hida, Kai, Kazusa and Awa (_??_), and the fact that the diffusion to Suruga was much later than the general trend, suggest that believers in the Jodoshinshu sect and the Nichirenshu sect in these provinces resisted the Ise faith (Fig. 4).
(5) The fact that most of the provinces mentioned in (4) are also located in mountainous or peninsular areas, and the fact that the diffusion to Shinano which is surrounded by the Chubu Mountains was much later than the general trend, suggest that physical and site factors delayed the Okagemairi diffusion.
Finally, in order to examine the above points from the aspect of spatial process, the writer simulated the information diffusion from Yamashiro by the gravity-type Monte Carlo simulation model whose flow-diagram and program are shown in Fig. 7 and in Appendix respectively. Fig. 8 shows the average output of 50 simulation runs of Model I whose coefficient of distance friction was determined to be 3.0 by the iterative method.

View full abstract

Is the microrelief on the flood plain such as natural levees, point bars and abandoned river channels growing during the recent floods? Some of them seem to have been formed under the control of eustatic base-level changes since the last transgression (so-called “Jomon transgression”), when relating the depth of archaeological sites and remnants to microreliefs on the fluvial lowland. The height of the last transgression in the Jomon period, the peak of which was during the late stage of the Proto Jomon to the early stage of the Early Jomon, about 8, 000 to 6, 000 years B. P., is estimated to be either 2-3 or 6 meters above the present sea-level. The level of succeeding minor regression in the Yayoi period, B. C. 300 to A. D. 300 years, is estimated to be 3 meters below the present sea-level in Japan.
Among the three characteristic sections composing the river plains, the lowermost deltaic plain section was once an offshore part in the former estuary during the Jomon transgression. The other two sections in the upstream side, i.e., the flood plain section characterized by natural levees and back-swamps and the fan plain section which are distributed in the upper side from the former river mouth, also may not have been independant from the sea-level fluctuation. Responding to those base-level changes, a sequence of filling up and cutting down above and below the present flood plain level and subsequent refilling up to the present flood plain level have occurred in the upstream side from the former river mouth (denoted as Type A). The spreading of fluvial deposits and the growth of deltaic plains have occurred responding to the regression in the former estuaries (TypeB). Type A -the Yoshinogawa plain. The Yoshinogawa plain is located in the tectonic graben along the Median Tectonic Line, where the late Pleistocene fans and terraces have been dislocated by the mosaic-like block faulting. The amount of dislocation during the Holocene is estimated to de on the order of several meters fine heights of the fault scarps cutting through the alluvial fans and elevation of wave-cut notches formed during the, Jomon transgression. The height of the Jomon transgression is 2.5 meters above the present sea-level in the Tokushima deltaic plain.
The geomorphic surfaces on the flood plain are classified into three; one space the up-per, middle and lower surfaces. The upper and middle surfaces distributed between Ikeda and Miyoshi, 75 to 65km from the sea, have scarps of more than 10 meters high above the present river floor(Figs. 2-a, b). The upper surface has usually remained as point bar during floods, but has been drowned twice during the last 50 years. Therefore, it may be considered as natural level now under construction. The middle surface is abandoned river channel or strath in shape and drowned frequently during floods.
At Ogaki, Miyoshi town, the ruins of ancient hamlets from the Yayoi to Mediaeval Age have been found on the upper surface, and the depth of the ancient cottage floor is within the horizon of paddy field soil, only a few tens of centimeters below the ground surface. A celamic fragment of the Late Jomon period was picked up on the upper surface, but it could be not related to any archaeological site. The remnants of the Yayoi period were found at 2 meters below the top of the middle surface which was composed of deposits 5 meters thick filling up the former river channel, and at 1.2 m below the surface the fragment of Sue celamic (5th to 12th C.) was containd (Fig. 3). Hence the upper surface had emerged at a certain time before the Yayoi period during the last regression.
The Joni field system, the rectangular lot system established in 7th and 8th C. on many Japanese plains, can be recognized on the upper and middle surfaces at Ogaki.

View full abstract

General relationship between the areal distribution of earthquake-induced landslides and the earthquake magnitude was discussed based on an investigation of more than 100 reports on 37 earthquakes which occurred in Japan after the middle of the last century. Occurrence of superficial slides, which were usually less than 10⁴m³ in volume of debris, was analysed on the maps of small-scale (about 1/500, 000). The larger slides are not suitable for the analysis in order to find the general relationship, because their occurrences are rather rare and usually controlled by particular geologic conditions.
The longest distances from either the epicenter (p) or the surface fault (f) of the earthquake to the outer margin of the area where superficial slides occurred densely (D) or sparsely (d) at the quake were measured, and designated as D_p, D_f, d_p, and d_f, respectively.
Several relations among D_p, D_f, d_p, d_f, and M, the earthquake magnitude, were noticed.
1) For each earthquake the following relations were recognized.
D_f≤D_p and d_f≤d_p
2) Although a considerable variation was observed in the distances measured for a given M, they were, on the whole, correlated positively to the M.
3) The following semilogarithmic relation was roughly recognized between M and D_f*, the maximum of D_f for each M.
log D_f*=a M+b
Where a and b are constants.
Similar trends were observed for D_p*, d_f*, and d_p*, which were the maxima of D_p, d_f, and d_p for each M, respectively. Values of a and b were calculated preliminarily, e.g. a approximated 1 for D_f* and d_f*, and 0.5 for D_p* and d_p*, while b is about 6.1 and 5.4 for D_f and d_f* respectively and about 2 and 1.7 for D_p* and d_p* respectively, when distances were measured in km. But further investigation based on more information would be necessary.
4) A few exceptions were recognized and cause were proposed tentatively.
5) Earthquakes of the magnitude less than 6 seem not to induce so many landslides.
These general relations are usefull, after suitable adjustments for the local variation, for a rough estimation of the area where superficial slides will be induced by an earthquake, of which the magnitude and epicenter can be estimated from informations on active faults.

View full abstract