292 Prof. Thomson on the Dynamical Theory of Heat. 



The object to be aimed at in perfecting a thermo-electric dia- 

 gram, and perhaps approximately attained to (conjecturally) in 

 the preceding, is to make the ordinates of the lines (which will 

 in general be curves) corresponding to the different metallic spe- 

 cimens, be exactly proportional to their thermo-electi-ic powers'^, 

 with reference to a standard metal (P3 in the actual diagram). 



135. Judging by the eye from the diagram, as regards the con- 

 vective agency of electricity in unequally heated conductors,! infer 

 that the different metals are probably to be ranked as follows, in 

 order of the values of the specific heat of electricity in them : — 



Specific Heat of Vitreous Electricity. 

 In Cadmium .... Positive. 



... Brass 



Copper 



...-TI;'-^^'^' '^^"'V . . . Positive, zero, or negative. 

 (^ Silver J 



... Platinum .... Probably negative. 



... Iron Negative. 



Zinc probably stands high, certainly above platinum. 



136. A very close analogy subsists between the thermo- 

 djmamical circumstances of an electrical current in a circuit of 

 two metals, and those of a fluid circulating in a closed rectan- 

 gular tube, consisting of two vertical branches connected by two 

 horizontal branches. Thus if, by the application of electromo- 

 tive force in one case, or by the action of pistons in the other, 

 a current be instituted, and if at the same time the temperature 

 be kept uniform throughout the circuit, heat will be evolved and 

 absorbed at the two junctions respectively in the former case, 

 and heat will be evolved in one and absorbed in the other of the 

 vertical branches of the tube in the latter case, in consequence of 

 the variations of pressure experienced by the fluid in moving 

 through those parts of the circuit. If the temperature of one 

 junction of the electrical circuit be raised above that of the other, 

 and if the temperature of one vertical branch of the tube con- 

 taining fluid be raised above that of the other, a current will in 

 each case be occasioned without any other motive appliance. If 

 the current be directed to do work with all its energy, by means 

 of an engine in each case, there will be a conversion of heat into 

 mechanical effect, with perfectly analogous relations as to ab- 

 sorption and evolution of heat in different parts of the circuit, 

 provided the engine worked by the fluid current be arranged to 

 pass the fluid through it without variation of temperature from 

 or to either of the vertical branches of the tube. If a^ and a^ 



* See § 140, below. 



