
V1I.— First Approximation to a Thermo-Electric Diagram. By Professor Tarr. 
(Plates VII., VIII., IX.) 
ety 
(Read December 1, 1873.) 
In the Session of 1867-8 I communicated to this Society a paper on the 
Dissipation of Energy, of which only a very brief abstract was published in the 
Proceedings. The main feature of that paper was the suggestion, as at least a 
valuable working hypothesis, that even in cases of the steady motion of 
heat, electricity, &c., the unexhausted energy is probably as small as possible, 
consistently with the conditions of each form of experiment. 
Applied to the conduction of heat, this hypothesis was shown to lead to the 
result that thermal conductivity is inversely as the absolute temperature, a 
result closely agreeing with the experimental determinations of Forpes. A 
similar result follows (from the hypothesis) for electric conductivity, where it 
has long been known from experiment that the resistance is nearly proportional 
to the absolute temperature. As the latter experimental law, however, is sub- 
ject to numerous exceptions, notably in the case of alloys, it was found neces- 
sary to introduce considerations of molecular change (such as alteration of 
specific heat, &c., with temperature) ; so that I determined to apply Forsers’ 
methods as well as electric testing to other pure metals than iron, and also to 
an alloy such as German silver. The reduction of my observations is still far 
from complete, but I have already stated to the Society that the change of 
thermal conductivity by temperature in German silver is, like that of electric 
conductivity, at least much less than in iron. 
These experimental determinations involved very great difficulties of various 
kinds, so that it was not till 1870 that I had an opportunity of testing experi- 
mentally the working hypothesis above mentioned in its application to the very 
curious phenomena of thermo-electricity. After a few experiments, however, 
I found that (at least within the limits where mercury thermometers can be 
employed) the so-called Specific Heat of Electricity is proportional to the abso- 
lute temperature, precisely the result indicated by the hypothesis. The follow- 
ing note is reprinted from the Proceedings of the Society for Dec. 19, 1870 :— 
“Jn a paper presented to the Society in 1867-8 I deduced from certain 
hypothetical considerations regarding Dissipation of Energy results connected 
with the thermal and electric conductivity of bodies, the electric convection of 
VOb, XXVIL. PART T. 2K 

