738 Prof. 0. W. Richardson on the Electron 



by insulating walls covered on the outside by a conducting 

 lining maintained at the potential V, and if the pistons are 

 sufficiently small and numerous and are made to work fast 

 enough. Since the extra pressure due to the current only 

 introduces second-order terms which vanish in the limit, we 

 are still led to equation (19), except that the meaning of a- 

 and w is different. <r is now equal to the actual specific heat 

 of electricity under the conditions to which it is subject in 

 experimental measurements, and w is the actual internal 

 latent heat of evaporation of the electrons under the condition 

 of continuous flow which holds during the cycle. 



It differs from the corresponding to which occurs in 

 equation (51), p. 619, by an amount which depends on the 

 difference of the rate of transference of kinetic energy by a 

 current on the two sides of the bounding surface. If we 

 keep the symbol id for the w which occurs in equation (20) 

 p. 603, and denote the true static internal latent heat of 

 evaporation which occurs as w in (51), p. 619, by c/>, then we 

 have the relation 



w = <t>-(\-\o), (1) 



where \ is the energy transferred in unit time by unit electric 

 current in the metal, and X is the corresponding quantity 

 for the current outside. As before a is determined by com- 

 parison of equations (20) and (51). In this way we arrive 

 at the equivalent expressions : — 



a -e\y^l +0 m - 6 i-)--d6p ' ' (2) 



^ 1.7-1 # B0J .... w 



e 1.7-1 e# e "d$ j ' • • ( } 



Quite similar considerations apply in determining the 

 Peltier coefficient. If P is the true value of this, then we 

 obtain the equations : — 



e? = w 2 -w 1 -e(V 2 "Y 1 ), (6) 



^^-^-(Xo-XO-KV.-YO, . (7) 



= Jr- Ja + Xr-Xa (8) 



