358 Mr. J. Parker on Thermoelectric Phenomena. 



joined so as to form a circuit without disturbing the equili- 

 brium, as Magnus long ago found by experiment. 



Again, let A, B be two pieces of different metals in contact, 



and let the temperature at the junction be t, and the difference 

 of potential 8. Then, if the free ends be at the same tempera- 

 ture t 0) their potentials will differ by 



This is not generally equal to 8 , and therefore, if the ends be 

 joined, a current will be produced. 



Let us now make a thermoelectric circuit of the two metals 

 A, B, and keep the junctions at the absolute temperatures 

 t, t Q respectively. 



B 



A 



Let R be the resistance of the circuit, and J the intensity 

 of the current, supposed to flow from A to B through the 

 junction of temperature t. 



The " electromotive force " of the circuit is defined to be 

 E^RJ. 



The heat generated in the homogeneous parts of the circuit 

 will be, by Joule's law, RJ 2 in a unit of time. This is exactly 

 balanced by the heat absorbed at the junctions, which in a 

 unit of time is equal to 



j.(n-n +s-s^). 



Hence 



Errll-IIo+Sa-S^ 



= 8-S + A a -A j8 , (9) 



by equations (1) and (8). 



It is a priori evident that the fall of potential in the homo- 

 geneous parts of the circuit must be exactly compensated by 

 the abrupt rises of potential at the junctions, so that equation 

 (9) affords a partial verification of our theory. 



If the temperatures t, t differ by an infinitesimal quantity 

 t, the electromotive force takes the form 



{"+(*.-*) -*}.T (10) 



Now suppose that when one junction is maintained at a 



