384 



BRIDGMAN. 



That is 

 and 



dtjp \df)p 



I-' 



dpdt 



It is obvious that (^) is equal to the entropy of one coulomb of 



dQ = tdS, 



electricity, because 

 and 



dt 



= t 



'd/dj 

 dt\dt 



and 



dQ 

 dp 



= / 



as 



dp 



= t 



'd fdj 



_dp\dt 



If we had some means of determining the work done on electricity as 

 well as the entropy we would be in a position to completely determine 

 its behavior. 



This function ^ gives at once the means of finding on the p, t plane 

 the slope of the lines of constant entropy. I have made this computa- 

 tion for all the metals listed above. As is to be expected, the line 

 assumes every possible slope. It thus appears that there is no simple 

 relation between the ordinary thermodynamic properties of a metal 

 and the thermodynamic properties of the electricity in the metal. 



Conclusion and Summary, 



In this paper measurements are given of the thermal e.m.f. of 

 couples composed of two branches of the same metal, one of the 

 branches being under uniform hydrostatic pressure and the other 

 branch at atmospheric pressure, the junctions between the branches 

 being at 0°C and some other variable temperature. The range of 

 the work covers 20 pure metals and 2 alloys, and all pressures up to 

 12000 kg./cm.2 and all temperatures between 0° and 100°. From 

 these measurements of e.m.f. the Peltier heats and the Thomson heats 

 may be deduced by the ordinary thermodynamic reasoning as a func- 

 tion of pressure and temperature. By the Peltier heat of the couples 

 measured in this work is meant the heat absorbed by unit quantity 

 of electricity in flowing across the junction from uncompressed to 

 compressed metal, and by the Thomson heat of the couple is meant 

 the excess of heat absorbed by unit quantity of electricity in flowing 

 through one degree temperature difference in the compressed metal 

 over the uncompressed metal. With regard to these heats, my position 



