PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 653 
that the doubts which various writers had thrown on the existence of thermo-electric 
inversions were groundless, I concluded with certainty that the newly conceived 
thermal effect of electricity in unequally heated metals really exists. But the 
theory left it undecided what the absolute nature and amount of this effect may be, 
and only showed how, by observations on thermo-electric currents, its difference in 
different metals may be determined. 
9. I therefore had recourse to direct experiment on the thermal effects of elec- 
tric currents in unequally heated conductors, not to demonstrate the existence of 
the peculiar effect anticipated, but to ascertain its nature, with moreover a view 
of ultimately determining its absolute amount, in some particular metal or metals. 
Before proceeding to describe experiments, by which 1 have now discovered the 
quality of the new effect in several cases, 1 shall, without entering on the mathe- 
matical details of the theory, or the full application of Carnot’s principle, repeat in 
a few words so much of my first communication on the subject to the Royal Society 
of Edinburgh, as to show the reasoning, founded on incontrovertible mechanical 
principles, which made me commence the experimental research with the certainty 
that the property looked for existed, whether I could find it or not. 
10 to 15. General inferences regarding the Electric Convection of Heat from 
Dynamical Principles. 
10. Gumming has discovered that in many cases when one of the junctions of a 
thermo-electric circuit of two metals is kept at a fixed temperature, if that of the 
other be elevated gradually from equality, an electro-motive force is produced, which 
first increases to a maximum, then diminishes, vanishes for a certain temperature of 
the junction, and acts in the contrary direction with gradually increasing strength 
as the temperature is further raised. It is clear that, at exactly that temperature of 
the hot junction for which in any such case the electro-motive force is a maximum, 
the two metals must be thermo-electrically neutral to one another, and must present 
reverse thermo-electric relations for temperatures below and above this point. Hence 
the thermal effect depending on the direction of a current crossing the junction of two 
such metals must be for temperatures above, the reverse of what it is for tempera- 
tures below, the neutral point, and must vanish when the metals are exactly at this 
temperature. 
11. For although Peltier himself supposed the effect he had discovered to depend 
on the conducting powers of the two metals for heat, and remarked as an anomaly 
the case of bismuth and copper, for which his supposition was violated, his own 
experiments show the truth to be, that in a circuit of two metals an absorption 
of heat at the junction where the temperature is higher, and an evolution of heat at 
the other, must be produced by the thermo-electric current which is caused by the 
maintaining of the difference of temperature between the junctions. That this is 
universally true when the temperatures of the two junctions are on the same side of 
4 R 2 
