
Cre) 
XXX.—Thermal and Electric Conductivity. By Professor Tarr. 
($$ 1-16, Read March 18, 1878—Revised (from a Shorthand Writer’s extended Notes) December 4, 1878.) 
(§§ 17-23, Read June 3, 1878.) 
The following paper contains the results of an inquiry which has occupied 
me at intervals for somewhere about ten years. It was carried out in part at 
the expense of the British Association, and I have already reported results to 
that body in 1869 and 1871. But these provisional reports referred to very 
short ranges of temperature only, and the experiments were made with 
faulty thermometers, for which I had not the corrections which had been care- 
fully determined by WELSH at Kew. 
The inquiry arose from my desire to extend to other metals the very beautiful 
and original method which Principal Forses devised, and which the state of 
his health prevented him from applying to any substance but iron. Forpss’ 
experiments gave a result so very remarkable, and (as it seemed to me) so 
theoretically suggestive, that I wished to extend them to other pure metals, and 
also, in one or two cases at least, to alloys. 
I believe that Principal Forses had at least two reasons for undertaking 
his investigations :—(1.) When he commenced his inquiry, there was no really 
accurate or trustworthy determination of the absolute conductivity of any 
body whatever for heat. (2.) Forses had himself, in 1833* and subsequent 
years, pointed out a very remarkable analogy between the conducting powers 
of metals for electricity and for heat, and had shown that these were almost 
precisely proportional to one another—that is to say, that the list of the 
average relative conductivities of different metals for electricity differed, 
from that of their relative conductivities with regard to heat, certainly not more 
than did the several electric lists furnished by different experimenters, and 
certainly less than the corresponding thermal lists. Hence it was natural to 
suppose that temperature might have a marked effect on thermal conductivity, 
as it was known to have such an effect on electric conductivity. 
The great merit of Forses’ method t is, that it seeks the conductivity in 
terms of its definition, instead of seeking a value of the conductivity which will 
best satisfy the integral of Fourter’s equation formed on the hypothesis of 
uniform conductivity, and of loss of heat from the surface of the bar in direct 
proportion to the temperature-excess above the surrounding air. Although 
eeProcevaioa Bena: + Report B. A., 1852. 
VOL. XXVIII. PART III. 9B 
