C. JBarus — Expression in Ther mo-electrics. 369 



same in both metals ; and they differ thermoelectrically when 

 this is not the case. If the transfer of charges is brought 

 about in essentially similar ways both in metals and in electro- 

 lytes,* it follows that the available paths of a charge must be 

 in its immediate neighborhood. 



Long ago Wildf pointed out that bad metallic conductors 

 were apt to occupy extreme positions in the thermoelectric 

 scale, relative to good conductors. An analogous but much 

 more definite result was brought out strikingly in the investi- 

 gations on tempered steely already cited, §2. Indeed we 

 found it expedient to refer all our thermoelectric data to an 

 imaginary steel rod whose specific resistance is zero, and an 

 independent consideration! then showed that this rod would 

 be nearly given by steel (better perhaps by iron) cooled down 

 to the absolute zero of temperature. 



7. A somewhat pictorial yet nevertheless interesting devel- 

 opment of the preceding paragraphs is obtained as follows : 

 From its mode of derivation, the thermoelectric condition, e, 

 depends on the state of electric dissociation at the given point 

 of the couple, for the given temperature. Hence it is con- 

 ceivable that a current in the wire, and e, will at all points be 

 reciprocally related. Indeed a serviceable mechanism for 

 Peltier and the Thomson effects is apparently near at hand. 

 To begin with the former : if the two wires of the couple are 

 thermoelectrically active, then they cannot be at the same state 

 of dissociation. If therefore a current crosses the junction, it 

 passes suddenly from a wire in a lower state of dissociation 

 (say) to one in a higher state ; or in other words from a metal 

 having relatively fewer available paths to one having relatively 

 more available paths. The case is thus closely analogous to 

 that of a fluid passing from a state of compression suddenly 

 into a state of dilatation, whereby heat is absorbed because 

 internal work is done. Conversely if the current pass from a 

 metal in a more advanced state of dissociation to one in a lower 

 state, the case is that of a liquid passing from greater to 

 smaller bulk and heat is evolved. Here therefore is a tangible 

 mechanism for the Peltier effect : for if it be admitted that 

 electricity travels in definite charges, its condition in a given 

 space must be quite as atomic! as the matter which transports 

 it, and heat is therefore absorbed or given off in marked degree 

 at the thermoelectric junction for much the same reasons that 



*Cf. J. J. Thomson: Application of dynamics, etc., p. 296, and others. 

 f Wild: cf. Mousson Physik, III, 2d ed., p. 387. 

 % Barus and Strouhal : 1. c, p. 62 et seq. 

 §1. c. pp. 65, 92. 



|| v. Helmholtz commenting on Faraday's law: "... gerade so, ais ware die 

 Elektricitat selbst in Atome getheilt." Berliner Sitzungsberichte, 1883, p. 651. 



