30 ART. 6.— S. KUSAKABE: 



since the twist consists of two terms, the form of any actual stress- 

 strain curve must be greatly affected in its features by the relative 

 value of the two terms. When the first term, which fulfils Hooke's 

 law, is negligible relatively to the yielding, the curve takes the form 

 given in PI. XIV. On the other hand, if the yielding is negligible 

 as compared with the principal term, then, as a matter of course, the 

 curve shrinks into a straight line. For any other relative values, the 

 curve takes an intermediate form. Figures in PL XV., which have 

 been drawn from the results of actual calculation, show several inter- 

 mediate forms. The inclined straight line in each figure represents 

 the curve to which it shrinks if the yielding be disregarded. In the 

 following pages, all these mathematical deductions will be experi- 

 mentally proved. 



VERIFICATION BY EXPERIMENTS. 



On gradually releasing a twisted piece of rock from the couple, 

 the specimen shows a tendency to persist in its twisted state, so that 

 there remains some residual twist when all the couple is withdrawn. 

 Moreover, if after withdrawing the whole couple, we begin to reapply 

 it, we find that the specimen is more rigid in one direction than in 

 the other. These facts are graphically shown in PI. XVI. for the 

 case of sandstone. 



AVhen a piece of rock is twisted cyclically within any definite 

 limits of couple, its stress-strain curve approaches a closed one of 

 simple and regular form. The on-curve, however, differs entirely 

 from the off-curve, though they are so related to each other that one, 

 when it is turned through two right angles, nearly coincides with the 

 other. These facts will be better understood from the curves traced 

 in Figs 35-4G, Pis. XVII. and XVIII. 



