682 RICE ART. L 



effect of its "induced charge" (i.e., the corresponding positive 

 charge left unneutraUzed by its exit) on it has vanished and no 

 further work is done against its motion on that account; that 

 has already been reckoned in e^ and the movement from the 

 surface to P produces no further disturbance of the surface 

 charge and no practical change in the "electron atmosphere" 

 or "space charge" in the enclosure, which has a very low con- 

 centration. Hence by the same statistical rule 



np ( e\V -Yp\ 



= exp 



( e\Y-YA \ 



n 

 or 



A;/ (log n' - log np) = e{V - Vp). 



Let us now consider two metal bodies not in contact with one 

 another but inside the same enclosure. When in equilibrium 

 the bodies will be at potentials Vi and V2. We then have the 

 following relations 



kt(\og rii — log n/) = e<^i, 



ktilog n/ - log np) = e(Vi - Vp), 



and two similar relations for the other metal. It follows easily 

 that 



ktlogui - 601 - e{Vi — Vp) = U log np 



= kt log n2 — €(f)2 — e(V2 — Vp), 



and therefore 



kt 

 Ti — T2 = "~ (log Wi — log n2) + <^2 — 01. 



B 



This relation is not disturbed by bringing the metals into con- 

 tact; it holds for any relative position of the bodies; when they 

 come into contact the electron concentrations on their contiguous 

 parts adjust themselves to produce a double layer consistent 

 with the discontinuity of potential Vi — V2 across the interfacial 

 boundary. The body with the smaller electron affinity has its 



