52 ART. 4. — K. HONDA AND T. TERADA. 



Thus the change of rigidity in these nickel steels is similar, 

 in its general aspect, to that given by the oscillation method. 

 In 50.72 o/o Ni. and 70.32 o/^ Ni., however, the small initial de- 

 crease of rigidity is not observed. Increased twist diminishes the 

 change of rigidity ; increased tension affects the change in a similar 

 manner but in a greater degree. The change as given by Barus's 

 method is generally greater than that by the oscillation method. 

 The difference becomes less as the tension is increased. In 28.74 o/^ 

 Ni., the results by the two different methods fairly agree, and the 

 agreement becomes closer with increased tension. Thus, also in 

 these alloys, relation (o) is not generally satisfied, except in 

 28.74 o/o Ni. The difference, however, becomes less as the tension 

 is increased. 



Beside the above samples, a number of others including iron 

 and nickel steels of different percentages, were tested with the 

 purpose of studying the different behaviours of the different 

 specimens. The samples were as follows : — 



All these specimens were slightly annealed by a Bunsen 

 flame. In the case of Swedish iron and 70.32 o/^ Ni., if we com- 

 pare the following results with those obtained above, we can 

 notice the effect of annealing on the change of rigidity. The 



