1. The flow of water through a circular bend having a rectangular section very nearly occurs under the law theoretically established. The velocity of flow of water is quicker at the inner radius, and the magnitude of velocity at any point is inversely proportional to the radius of the point. The mean velocity exists practically at the centre line of the bend. The pressure is higher as the radius is greater. At the inner portion a vortex motion seems to exist. 2. Before and after the bend, a pressure rise and fall occurs. Hence to measure the loss of head at the bend, the pressure difference between two points far from the bend, (at least each at a distance of perhaps five times the mean radius of the bend from the beginning and the end of the bend) must be measured. The real loss due to the bend is perhaps much smaller than those reported to be, because in the long run, the energy once apparently consumed will be recoverd. 3. The loss of head due to a bend could be minimized if the centre line is made parabolic. Finally I wish to express my hearty thanks to Prof. A. Inokuty and Prof. K. Takemura, the former for his invaluable aid generally and cordial suggestions on the various apparatus such as the utilization of Ccl_4, and the latter for his initial suggestion on the subject. Moreover I am indebted very much to Mr. Ishikawa, Mr. Otuka, Mr. Nagata and Mr. Kubo, students of our University, for their continued assistance in taking the readings during a long period of time and sacrificing their summer vacation in the interests of our experiments.

View full abstract