64 



K. TSURUTA. 



proportionally to the load applied, and does not present the least sign 

 of hysteresis. 



§ 37. Fourthly nickel. The specimen in hand was a rather 

 poor one, yet inasmuch as experiments on another sample, alleged to 

 be somewhat impure, showed the same general features, the following 

 numbers may perhaps be taken as representing fairly w r ell the thermo- 

 electric effects of stress in nickel. Compare the curve in Fig. 6, 

 PI. IV. 



Up to 5 kilos, the wire showed no sign of elongation, and at the 

 load of 7*5 kilos, it drew very much. The "on" branch certainly 

 deviates from the " off " and is on the negative side of the latter, 

 their juxtaposition being thus opposite to that in iron and steel. 



§ 38. In the fifth place, platinum. The sj ecimen was 0"4 

 m.m. in diameter and could not be loaded with more than 1*5 kilos. 

 Only after a number of loadings and unloading« did the thermo- 

 electric E. M. F. become nearly cyclic and vary perfectly proportion- 

 ally to the load applied. The result is given in the following table, 

 with which the curve in Fig. 7, PI. IV. is to be compared. 



