As the cause of instability of slopes, 1) Change of slope profile (toe erosion, artificial excavation, etc.) 2) Increase of pore pressure are generally known. However, according to our field investigation and measurement of a crystalline schist landslide, instability of the slope can not be explained by 1)change of slope profile or 2)increase of pore pressure, but we have found "underground erosion" must be the main cause of instability of the slope. "Underground erosion" has hardly been studied in engineering field and geological or geomorphological field up to now. However, "underground erosion" is the main cause of instability of some crystalline schist slopes and it acts a vital role in landslides triggered by vertical subsidence (Sassa et al. 1980a,b)
The authors have studied one of biggest crystalline schist landslides "the Zentoku landslide" in Japan since 1972. The landslide locates in the Shikoku island and its scale is about 1.2 km in length, 2.0 km in width, 25–30° in slope inclination and 40–50 m in depth. The ground consists of clayey schist, sandy schist and quartz schist, and the beddings are nearly parallel to slope. The whole landslide is too big for a precise investigation, then we selected the Taniare district of the landslide as the object of our investigations. Fig. 1 is the map of the Taniare district. "Crushed zone" and "underground valley" in the figure were estimated by seismic explorations Sassa et al. 1980c).
The movement of the Zentoku landslide have been measured from the end of 1973 by extensometers on line X and shear displacement displacement meters (designed by K. Sassa, 1980c) on line Y and pipe strain meters in the boreholes No.1–16. The ground water levels and precipitation have been observed, too.
Fig. 2 is the example of extensometers in 1975, 1976,1978. In the figure the extension/ compression observed by each extensometer is accumulated from the end of the extensometer line. A right/up line shows extension and a right/down line shows compression. Since "Road" seems to have been stable according to the data of extensometers and its retaining wall has been intact, "Road" is dealt as the stable point. Therefore, the vertical distance from the horizontal line of each year to the respective point of the graph indicates the movement at the point/time. The profile of the slope and their slip planes are also shown in Fig. 2. The depth of slip planes was decided by the insert-type strain meter designed by K. Sassa, 1980c) and the pipe strain meters in 13 boreholes along this line. And the sliding blocks were decided from the results of extensometers including Fig. 2. The observation of extensometers for these seven years and Fig. 2 made clear that some tens cm movement took place every year, though there was no change of slope profile. Slip planes finish at "Road" and there have been no toe erosion and no artificial excavation, and no filling, too.
