84 



SCIENCE. 



[N. S. Vol. XXI. No. 525. 



action at a distance) in carrying unit 

 quantity of positive electricity from point 

 1 to point 2. 



By electromotive force, E^.-,. along a 

 given path from the point 1 to the point 2, 

 I shall mean 



E].2 — + ii-iRi.t, 



where ii.2 is the current from (1) to (2) 

 and i2i.2 is the resistance of the chosen 

 path from (1) to (2). 

 If either i or is zero, 



which is practically the case when we have 

 a battery in open circuit, (1) being one 

 terminal and (2) the other, or when we 

 have under consideration two points on 

 opposite sides of a junction of two metals, 

 but exceedingly near together, even if a 

 current is flowing from one to the other. 



We have already, looking at Fig. 3, com- 

 pared the passage of electricity from metal 

 1 to metal 2 to the evaporation of water 

 in a boiler. Now in this evaporation work 

 of two kinds is done upon the water, in- 

 ternal work and external work. The move- 

 ment of electricity across a junction against 

 a difference of potential corresponds to the 

 external work of evaporation. Is there 

 accompanying this movement anything cor- 

 responding to the internal work of evapora- 

 tion ? If so, the heat absorbed by the elec- 

 tricity in the movement may be as bad a 

 measure for the difference of potential at 

 the junction as the latent heat of evapora- 

 tion would be for the external work of 

 evaporation. 



It is not absurd to imagine that there 

 may be some change of state of electricity 

 besides change of potential. It is possible 

 that we should take account of something 

 like an attraction between electricity and 

 the metals with which it is associated. 

 Helmholtz imagined such an attraction in 

 order to explain the action of a galvanic 

 cell. Indeed, we are familiar with the idea 



that attraction or repulsion exerted on the 

 electric charge which ordinary matter may 

 bear is communicated to the matter itself. 

 When the charge on a pith ball is drawn 

 this way or that, it carries the pith ball 

 along with it. To be sure, this phenomenon 

 and others like it may not indicate any 

 fundamental attraction between ordinary 

 matter and electricity. Perhaps they can 

 all be explained by stresses in the dielectric 

 surrounding or penetrating the ordinary 

 matter ; but M'hatever the true agencies 

 may be, they at least simulate attraction 

 or some physical tie between ordinary mat- 

 ter and electricity. We may, therefore, 

 feel free to make speculative use of such 

 attraction. 



Our problem is to find, if we can, by use 

 of any reasonable hypothesis, an explana- 

 tion of the way in which heat drives an 

 electric current around the circuit of dis- 

 similar metals unequally heated. 



There are two types of mechanical cir- 

 cuits or cycles operated by heat with which 

 we are very familiar, the steam-boiler- 

 engine cycle, in which the circulation may 

 be practically in a horizontal plane, and 

 various convection cycles, commonly used 

 for heating and ventilation, which may be 

 in vertical planes. In the horizontal cycle 

 we must have valves. Circulation is se- 

 cured by heating or cooling a fluid which 

 is free to expand or to contract on one side, 

 but not on the other side, the valves being 

 so contrived as to give the necessary free- 

 dom and the neeessarj' restriction. In the 

 convection cycle we do not necessarily make 

 use of valves. If the heating and cooling 

 are effected at the right parts of the cir- 

 cuit, gravity supplies the differential force 

 necessary to maintain circulation. 



How can the metals of our thermo-elec- 

 tric circuit take the function of valves or 

 the function of gravitation and so deter- 

 mine the flow of electricity at the expense 

 of heat energy? 



