Observation of a series of more than 50 exposures of the gentle and smooth mountain slope in the Northern Hidaka Mountains, central Hokkaido, revealed that the slope is composed of ill-sorted angular debris which is covered and/or intercalated with Holocene pumice layers. Fabric analysis of the slope deposits, including measurement of vector mean and vector magnitude of the a-axis, which is nearly parallel to the azimuth of the slope, strongly suggests that the deposit is of periglacial origin, probably transported by solifluction. The vector magnitude of the slope deposits in the study area generally is more than 65%, while that of the fluvial, debris-flow, and other non-periglacial slope deposits is below 65%. Dating the pumice layers and Holocene andosol horizon allows us to determine the period of solifluction on the slope as follows: Stage 1: before the fall of Ta-d pumice (9, 000y.B.P.) Stage 2: the fall of Ta-d-short periods (not confirmed) Stage 3: before the fall of To-c2 pumice (3, 000y.B.P.) or Ta-c1 pumice (2, 000y.B.P.) Stage 4: after the fall of To-c2 or Ta-c1 Stage 5: after the andosol horizon (7, 000-3, 000y.B.P.) The periods of solifluction differ with rock types and slope altitude. Granite area: most solifluction ceased in stage 1 or 2 in the area lower than about 1, 000m a.s.l., while the slope was unstable until stage 3 in the area above about 1, 000m a.s.l. Hornfels and shale (Hidaka Group) area: many localities show continuation of the solifluction until stages 3, 4, and 5, while the movement ceased in stage 1 or 2 in some localities. In the granite area, an altitude of about 1, 000m a.s.l. seems to correspond to the lowest horizon where the solifluction occurred in stage 3. The fact that the solifluction continued for a longer time in the hornfels and shale area than in the granite one may explain by the finer debris production in the former.
This study reconstructed the vertical change of slope stability in the southern part of the Yubari Mountains and the western part of the Hidaka Range, Hokkaido, Japan, by using three eolian pumice layers deposited during the Last Glacial and early Holocene; Spfal, 33, 000-32, 000y.B.P.; En-a, 19, 000-16, 000y.B.P.; Ta-d, about 8, 000y.B.P. A pumice layer overlying a mountain slope indicates that the slope has been stabilized since the pumice fell. In this view, the examination of each pumice layer in about 330 outcrops on the mountain slope suggests a belt of slope stability in the study area, as follows: Spfal; crest slope lower than about 250m high. En-a; crest slope lower than about 500m and side slope lower than about 300m high. Ta-d; crest-side slope lower than about 1, 250m high. On the other hand, the lack of a pumice layer and/or the absence of a pumice layer which is usually incorporated in the colluvium suggest slope instability after the fall of the pumice. Several outcrops on the mountain slope suggest that the removal of the pumice layer was mainly caused by periglacial slope processes, associated with the frost shattering of the underlying bedrock. The belt higher than the belt of slope stability, in which each pumice layer was removed, therefore, corresponds to that of the periglacial belt in the period after the fall of each pumice layer. According to this view, the lower limit of the periglacial belt rose from about 250m during the 33, 000-18, 000y.B.P., to about 500m during 18, 000-8, 000y.B.P., to about 1, 250m after 8, 000y.B.P. This vertical change in slope stability is parallel to that of the timber-line and snowline of the study area which has been indicated by other data.
A Paleomagnetic method for detecting fireplaces in Pre-ceramic sites was established, and its availability was tested in detail. At first, a bonfire experiment was carried out with the purpose of evaluating magnetic property changes of soils through a baking process. When soils were baked, the remanent magnetizations changed (1) to become stronger in intensity by one to two orders, (2) to become more stable in both intensity and direction against alternating-field demagnetization, and (3) to be directed more nearly in an identical direction than before being backed. These magnetic property changes may have been caused by chemical change (dehydration) from goethite to hematite or magnetite in soils through the baking process. These differences before and after baking were adopted as criteria to detect fireplaces and to determine their positions in Pre-ceramic sites (Paleomagnetic method). Secondly, the availability of these criteria was tested using soils in two archeological sites, Tamatu-Tanaka site, Hyogo (Yayoi age) and Ichi-no-Kubo site, Oita (early Jomon age-Preceramic age), where fireplaces have been definitely recognized with the naked eye. The results showed that remanent magnetizations of backed soils sampled from the fireplaces in these sites satisfy the above-mentioned criteria, and that those of unbaked soils sampled in the same sites do not. However, differences between the magnetic properties of the baked and the unbaked soils were not so significant as seen in the results of the bonfire experiment. This is attributed to chemical alteration of magnetic particles induced through a chemical process over a long period. Consequently, it is important to utilize all the above-mentioned criteria with significant care, in order to correctly detect fireplaces in Pre-ceramic sites.