718 PROFESSOR TAIT ON THERMAL AND ELECTRIC CONDUCTIVITY. 
Forses’ paper has been printed in the Transactions of this Society,* I may 
make a few additional remarks on the methods he employed. 
He used for the first part of the experiment, what he called the statical 
experiment, a bar of iron, 8 feet long by 14 inch square section. One end of 
this was raised to a high temperature by means of a pot containing melted 
solder, whose temperature was maintained nearly constant for eight or nine 
hours. The rest of the bar was exposed to the air of the laboratory, and of 
course parted with a portion (of the heat: conducted to it) partly by radiation, 
partly by convection. It was found that after about eight hours a stationary 
distribution of temperature was attained, in which the gain of heat in any 
section of the bar by conduction was just neutralized by the surface loss. This 
temperature distribution was then accurately determined. In the second or 
dynamical experiment, a shorter bar, of exactly the same transverse dimensions, 
was employed ; not, however, for the conduction of heat, but for the purpose of 
ascertaining at what rate its heat was lost by radiation and convection at dif- 
ferent temperatures. For this purpose the bar was heated as uniformly as 
possible, once for all, and then allowed to cool in the air, its temperature being 
noted at measured intervals of time. The introduction of the experiments with 
the shorter bar was the main point of great importance in which ForBEs improved 
the experimental part of the determination. And, as regards the subsequent 
calculations, it need only be said, to show the improvement he introduced, 
that had he followed Biot’s mode of procedure he would probably have failed to 
discover that thermal conductivity (in some cases at least) depends on tempera- 
ture. As I have already said, though Fores’ results were confined to iron, 
they were the first of any real value to the absolute measurement of thermal 
conductivity. 
§ 1. Viewed in the light of the results attained, I do not now think so much 
as I was originally disposed to do of one of the chief reasons which led me to 
the present inquiry. But that does not in any way matter to my other chief 
reason ; for, though an attractive hypothesis has been shown to be untenable, at 
all events without very considerable restrictions, some valuable and even curious 
measurements have been made. ForBEs’ results for iron have been, in all but 
one particular, closely reproduced by myself, but their most striking peculiarity, 
the falling off of conductivity by increase of temperature is, so far as I yet know, 
confined to the single metal which he experimented on. I had fancied that as 
the numerical results given by him seemed closely consistent with a conductivity 
varying inversely as the absolute temperature, such might be generally the 
case. Inquiring into possible physical reasons for this, I saw that if it were 
assumed that, in the steady linear propagation of heat, the amount of available 
* Trans, R. 8. E., 1860-61, and 1864-5. 

