44: PROCEEDINGS OF THE AMERICAN ACADEMY 



If, now, we pass through the circuit an external current of strength 

 (7, the whole heat absorbed is (O -}- (S') C. 



This should be enough heat to produce a current G times as strong 

 as the proper thermo-electric current, if the thermo-electric current is 

 due to the heat Q -\- S. 



If, then, /be the strength of the thermo-electric current, we should 

 expect that the whole current in the circuit would be: — 



(7+/± CI. 



But experiment shows that the whole current exactly = C -\- J. 

 We must, then, conclude that the current which is equivalent to the 

 heat (9 -}- 'S') C is not (7 times as great as the proper thermo-electric 

 current ; consequently, the thermo-electric current cannot be the 

 equivalent of the heat 9 -j- z^. In other words, the thermo-electric 

 current cannot be the equivalent of the Thomson and Peltier effects. 



All the experiments that have been made on the Peltier and Thom- 

 son effects have been made when these phenomena appeared as the re- 

 sult of a current, not when they appeared as its cause. The heat 

 absorbed by a thermo-electric current itself has never been measured. 

 All measurements of the heat effects have been made by passing an 

 outside current through the circuit, the heat effects due to the thermo- 

 electric current itself being too small to be measured. But we have no 

 right to suppose, a priori, that at the hot junction of a thermo element 

 tlie only heat absorbed in the production of the thermo-electric current 

 is that due to the Peltier effect. If, in the experiment that I attempted, 

 h had satisfied equation (4), then the heat due to the Peltier effect 

 would have been the only heat absorbed at the hot junction by the 

 thermo-electric current. But, as the experiment was a failure, there 

 are no grounds for this assumption. It is, therefore, not surprising that 

 equations (1), (2), and (3), which are based on the assumption that 

 the Peltier effect is the only heat effect at the hot junction which 

 causes the thermo-electric current, should give results inconsistent 

 with experiment. 



The Thomson effect, even more than the Peltier effect, appears to be 

 the result of a current, not the cause. 



All experiments on the Thomson effect are made by passing a strong 

 current along a bar of metal, the ends of whicli are at different tem- 

 peratures. It is then found that the temperatures of fixed points on 

 the bar are different when the current is passing from what they were 

 before it passed. In some metals the temperature is raised, in others 

 diminished, as the current passes from the hot to the cold end of the 



