[ 737 ] 



LXXIV. The Electron Theory of Thermoelectric and therm- 

 ionic Effects. By 0. W. Richardson, Professor of Physics, 

 Princeton University *, 



§ 1/TX7HAT follows is mainly a discussion of the 

 Tt criticisms by Dr. Bohrf and Professor H. A. 

 Wilson J, of parts of my recent papers §. These criticisms 

 are especially welcome as it is now very desirable, from 

 several points of view, to determine the precise extent and 

 nature of the theoretical warranty which underlies the dif- 

 ferent formulae considered. There is probably no more 

 satisfactory method of accomplishing this than by means of 

 an open discussion of the questions at issue. 1 have also 

 considered the results of some recent experiments by Pring 

 and Parker || and Fredenbagen ![• 



§ 2. Thermoelectric Phenomena. 



The difference between my formula? and those obtained by 

 Dr. Bohr does not appear to arise from any inherent defect 

 in the method which I employed, but from the fact that I 

 neglected to include effects arising from a possible difference 

 in the rate of transference of kinetic energy by an electric 

 current in different materials. My calculations may easily 

 be extended so as to include these effects in the following- 

 manner. 



Let us take the Thomson effect first. 



The Thomson and Peltier effects are produced by a current 

 flowing uniformly under an applied potential gradient, and, 

 as Dr. Bohr has pointed out, are not necessarily identical 

 with the corresponding effects which are produced by the 

 virtual displacement of an equal number of electrons subject 

 to the equilibrium conditions. In order that we may be 

 sure that a is really the same thing as the Thomson co- 

 efficient, it is necessary to modify a little the cycles such as 

 those considered on p. 603 of my paper. This is to be done 

 in such a way as to make the current flow continuously and 

 reversibly round the cycle, under the influence of an applied 

 electromotive force. The conditions are satisfied if the con- 

 ductor joining A and A' is of sufficiently high resistance, 

 and contains a source of electromotive force, if A and A' 

 are indefinitely large and are contained in chambers bounded 



* Communicated bv the Author. 



t Phil. Mag. vol. xxiii. p. 984 (1912). % lh ^ d - vol. xxiv. p. 196 (1912). 



§ Ibid. vol. xxiii. pp. 263, 594 (1912;. 



j| Ibid, vol; xxiii. (1912). 



i\ VerJi. d. Beutsch. Physik. Ges. 14. Jahrg. p. 384 (1912). 



