A New Thermoelectric Phenomenon. 



125 



The table shows the temperatures at which three different 

 thermoelectric junctions will generate various electromotive 

 forces, the cold junctions being at 0° C. 



The above are based on Professor Tait's results. 



Returning now to the phenomenon under consideration, I may 

 say that, in this paper, I am not following altogether the order of 

 my experiments, but am giving all experiments on one metal 

 together, though I frequently left a metal and returned to it 

 again. 



In platinum there was apparently an anomalous result. Two 

 unequally thick wires and a very fine one were stretched in series, 

 the fine being between the other two. Heating the junction of the 

 fine and coarsest gave '0027 volt ; while the fine and medium gave 

 •0007, both in the satne directioti ; one would expect that in each case 

 it would be from fine to coarse or vice versa. It may however 

 have been due to diiferent amounts of impurity in the specimens. 

 When the medium wire was heated by itself there was no 

 perceptible result, the coarsest gave -0001 volt and the fine gave 

 various amounts up to "0023 volt. 



Fine brass wire heated in a tube behaves similarly to iron 

 when heated in a flame, the effect being very unsteady. It is 

 much less sensitive than gold to being disturbed ; the liighest 

 effect I observed with brass was -015 volt. 



The behaviour of German silver was, in many respects, similar 

 to gold. The effect was greatly increased by repeated heating and 

 cooling, and the changes were generally slow and steady, though 

 occasionally without any apparent cause the changes became 

 great and abrupt, as much as with iron. Like gold too it showed 

 distinct positions of maximum and minimum effects, they being even 

 more marked than in the case of gold. Plate YII. shows the best 



