44 



K. TSUllUTA. 



ci ably clear of it after two or three steps of unloading. There were 

 of course some readings which gave points lying above the " on " 

 branch, but I take it to be still open to consider the formation of the 

 loop as one of the distinct features in the thermo-electric curve. 



§ 20. I have remarked that the points of the thermo-electric 

 curve which lie to the right of the minimum points correspond to the 

 more stable conditions of the iron. Examining how the thermo- 

 electric changes set in at these points when we begin to load or unload 

 the wire, we have to distinguish between the two cases represented in 

 the annexed figures. The spots in the figure are meant to be repre- 



Load 



sentative points, and the sloping lines to be the limbs of the thermo- 

 electric curve under consideration. There are added the axes of 

 reference drawn through the spots, with respect to which the limbs 

 lie in the first and third quadrants. Dotted arrow heads show the 

 initial thermo-electric changes produced by loading or unloading the 

 wire as indicated by full arrow heads. Thus, we see that the increase 

 or decrease of the load corresponds to the increase or decrease of 

 E. M. F. 



§ 21. If we now examine this question on the other side of the 

 minimum points when the metal is brought to more stable conditions 

 by being subjected to agitation, it will be seen that so far as only the 

 initial changes are concerned the same relation holds true. Of the four 

 cases represented in the following figures, two (a and c) refer to the 



