]4 AßT. 6.— S. KUSAKABE : 



The influence of the different magnitudes of the constant couple 

 (the resisting mass being the same) upon the yielding was then 

 examined. In successive experiments, the constant couples were in 

 the ratio of 3 : 5 : 7 9 : 11 : 13 : 15. The curves in PL IX. show 

 the result. As a matter of course, all curves are of similar form, but 

 there exists the remarkable difference that of any two curves, the one 

 whose couple was the smaller, approaches to horizontality more 

 quickly than the other. Here it must be observed that the absolute 

 amount of twist is not known, since the so-called zero-reading is 

 nothing but the reading which corresponds, as the case may be, to a 

 certain unknown twisting couple lying between +M <rgi- and —M Q agv. 

 In other words, all curves in PL IX. are not referred to one and the 

 same origin of coordinates. 



Pate of the Yielding : — From the values of the twist and of 

 the corresponding time, the rate of the yielding may be calculated by 

 the formula 



At 



■«")■ 



A* ~ t — V 



where r m represents the amount of twist at the instant t. Plotting 

 the result, obtained from eight observations, as shown in Fig. 16, 

 PL X., we may, at once, perceive that the general relation between 

 the rate of yielding and the time-element much resembles something 

 like a rectangular hyperbola. To test whether this is true or not, 



instead of c and t themselves, their logarithms are plotted in 



Fig. 17 on the same plate. The most probable curve and also the 

 simplest at the same time, would be a straight line. One straight 

 line, whose equation is 



log-^+log 1 = 109^- 



