In the field we carried out the surveying of ricer bed profiles and cross sections in the observed reach which extends about lkm which bas a gradient of 6 degrees, when the floods caused the river bed fluctuation of between 1979 and 1991. Over the period 1979 to 1987, no modification of the channel of the reach by the check dam had been taken. But in 1988, the check dam which was 13 meters in height was built at the end of the observed reach, after that the influence of the check dam appeared in profiles and cross sections of the reach.
As a result of this investigation, we found the following fact. 1)If the check dam existed in the observed reach when enormous volume of sediment supplied from the upper stream, rotational aggradation occurred in the upper reach of the dam, the longitudinal profile extended linearly from the dam crown without being affected by the original channel form and shape, and according as the sediment supply from the upper stream increased, the longitudinal slope
became steeper. When no sediment supplied from the upper stream by the flood after the large deposition, the deposit of the upper reach of the dam removed to the downstream decreasing its longitudinal slope. 2)lf the check dam did not exist, parallel aggradation occurred in the observed reach when enormous volume of sediment supplied from the upper stream, and in erosional process after the large deposition, the longitudinal profile returned to the former profile rapidly.
In the laboratory, we carried out the sediment-feed experiment in a flume 500cm long, 20cm wide and 23cm deep which set the dam of 5cm in height at the end. A beltconveyor 150cm long was used for controlling the sediment feed rate into the flume. Two size material of sand(Imm)and gravel(5.3mm)was used In this study. As a result of this experiment, we got the following fact. I)Rotational aggradation occurred in the upper part of the dam. 2)ln the uniform sand or gravel, the higher the sediment feed was. the steeper the longitudinal slope was. 3)The longitudinal slope made by the uniform gravel was roughly twice as large as the slope made by the uniform sand to any sediment feed rate. 4)ln sand and gravel mixtures, the relation between the concentration and the slope was not so simple as in the uniform material. When sand was added into the flume, keeping the feed rate of gravel constant, In spite of the fact that the concentration of the mixtures was increased, the slope decreased. This is due to Increase the mobility of the mixtures. This mobility depends on the proportion of sand to gravel, but it is not in simple proportion to sand. As mentioned above, without the understanding of characteristics of heterogeneous sediment transport, we can not clarify the function of a check dam. Less