Potential Energy of Liquid Surfaces. 455 



of more than 35°. These last facts (which, I am sure, must 

 have surprised the able experimenter) would be explained, 

 according to equation (1 bis), by the threefold influence of 

 the values of the liquid mass m, of its specific heat k, and of 

 the wetted surface S. 



IV. I now approach another class of phenomena, which is 

 directly connected with equation (1). The nature of the 

 thermoelectric currents developed by the action of heat, 

 whether in one and the same conductor the parts of which do 

 not all possess the same physical properties, or in hetero- 

 geneous conductors of which the surface of contact is heated, 

 allowed me to presume that equation (1) could be transformed 

 into another containing the difference of the electric poten- 

 tials of two conductors on the two sides of their surface of 

 contact, whether these conductors are heterogeneous or only 

 differ in their physical properties. I was fortunate enough to 

 see verified what I had conjectured, and thus to discover the 

 physical cause of a numerous class of phenomena which have 

 much perplexed those who discovered them. 



To effect this transformation, I make use of two proposi- 

 tions borrowed from the theory of thermoelectric currents, 

 and first demonstrated by M. Clausius : — 



1. The difference of electric level x established at the con- 

 tact of two heterogeneous bodies (or differing physically) is 

 proportional to the absolute temperature of the contact. 



2. The quantity of heat dQ which traverses in the time dr 

 the surface of contact of the two bodies is equivalent to 



Axidr, 

 A being the thermal equivalent of the unit of work, x the 

 difference of electric level, and i the intensity of the current. 



Having regard to these two propositions, I can therefore 

 write, designating by X a constant which depends on the 

 nature of the two bodies : — 



x = \t, dQ J = A\itdr. 

 Introducing the value of dQ into equation (1), it becomes 



AXitdr = Aa | y-c/S + -p 2 & dx f > 

 or 



*-<«©■ ....... (2) 



I conclude from this transformed equation that to every 

 variation of temperature of the surface of contact corresponds 

 a variation in the difference of electric level, and consequently 

 a thermoelectric current will be produced if the circuit be 

 closed. 



