THERMO-ELECTRIC EFFECTS OF LONG. TENSION TN DIFF. METALS. 



63 



of inflexion, though ;i very indistinct one, while the " off" is convex 

 towards the axis of load ; and in fact, the general form of the curve is 

 such as may induce us to class it among those for hard iron. 



Curves for steel hardened by drawing are almost identical with 

 the above and as was the case with iron present nothing that calls for 

 remarks. 



By a series of experiments I have ascertained that from a smaller 

 cyclic range of loading and unloading up to the greatest possible the 

 form of the curve again remains exactly the same, and also that the 

 effects of vibration are very small, and take place always in the same 

 senses as in iron. 



§ 36. Palladium comes next in the order of my experiments. 

 My measurement completely confirms the remark of Dr. Cohn that it 

 shows a large effect under comparatively low loads. Though I was 

 fortunate enough to have in my possession a sample of it long enough 

 for my arrangements, I could not multiply my measurements as in 

 iron and steel. The wire was 0'4 m.m. in diameter and could be 

 permanently elongated by a load of 1*5 kilos. The following experi- 

 ment is one out of many, all of which were perfectly concordant with 

 each other. 



The graph of the experiment given in Fig. 5, PI. IV, shows that 

 the thermo-electric E. M. F. changes towards the positive side, varies 



