Forces in Voltaic mid Thermoelectric Piles. 269 



lightly commented upon. At the same time I remember that 

 whenever direct experimental evidence in favour of this view 

 has been sought for, as it has been industriously by Edlund 

 and by Sundell, it wholly and admittedly fails to support it. 

 So much for the outpost, now for the main attack. 



Besides the above assumption, Ayrton and Perry hold 

 tenaciously to another dogma. They believe (I feel sure 

 they will acquiesce in this) that reversible heat-localities and 

 seats of E.M.F. are quite disconnected ; that Peltier-effects 

 are no sign of a local difference of potential ; conversely, that 

 electricity may rise or fall in potential without disappearance 

 or generation of heat or other form, of energy ; and in general 

 that energy may disappear from, or be created at, any point 

 of a circuit without any work being done on the spot. 



Well, then, if I dissent from all this, what is it that I want 

 to say instead ? I only wish to generalize the above defini- 

 tion (2) so as to make it applicable, not only to the whole 

 circuit, but also to every portion of it. I would remake the 

 statement (2) so as to include all its old meaning and some- 

 thing more: and say, 



2'. The measure of E.M.F. at any point of a circuit is 

 the work done per unit electricity conveyed past that point. 



dE=^ (20 



Surely a harmless modification ; but once grant it, along with 

 the baselessness of Ayrton and Perry's first assumption, and 

 everything goes my way. Reversible heat-localities and 

 seats of E.M.F. are identical in metallic circuits: reversible 

 energy-localities and seats of E.M.F. are identical everywhere. 

 Energy cannot appear or disappear at any point without 

 equivalent work being done on the spot. Thermoelectric 

 forces and Yolta-effects are utterly distinct, having no con- 

 nection with each other. Peltier-heat also is quite uncon- 

 nected with Volta's effect, though wholly bound up with 

 thermoelectricity. The E.M.F. of a cell is due to, and located 

 along with, the reversible chemical and thermal actions from 

 which it is admitted to be calculable, and whence its energy 

 is derived ; and the Yolta-effect sinks into the comparative 

 obscurity and insignificance of a disturbance in the electric 

 condition of metals by reason of their chemical affinity with 

 the particular atmosphere which happens to surround them. 



My view of a thermoelectric current, then, is simply this: — 

 An E.M.F. necessarily exists at every place where reversible 

 heat phenomena, whether of Peltier or of Thomson, are occur- 

 ring ; its magnitude being measurable by the quantity of heat 



