
PROFESSOR TAIT ON THERMAL AND ELECTRIC CONDUCTIVITY. 723 
highest temperature applied at one end, after 10 hours exposure, there was 
scarcely any perceptible heating at the further end. The same bar served 
first for the statical experiment, and then was heated again for the cooling 
experiment. 
§ 7. In procuring the thermometers, on whose accurate indications the 
whole value of the experimental work depends, I availed myself of the assist- 
ance of Dr Batrour STEWART, who was then director of the Observatory at 
Kew. Two sets of thermometers were made under his supervision, one set 
with long range and short degrees, the other with short range and long degrees, 
and all were tested by him. I had wished as far as possible to carry out 
ForsEs’ idea that it was better to use thermometers, even if they did not show 
the zero point, which even at high temperatures exposed only a small quantity 
of mercury in the stem, than to have a long column exposed to the air, with its 
temperature of course very different at different parts. Dr BaLrour STEWaRrt, 
however, told me that, so far as he knew, it was impossible to accurately 
graduate thermometers under these conditions ; and he advised me to take the 
thermometers as he could make them and guarantee them, namely, mercurial 
ones, made of proper glass, carefully divided by graduating instruments at Kew. 
As this is a point of vital importance, I append in a foot-note an extract from 
Dr Stewart's letter.* 
I have already spoken of the circumstance that when the bulbs of some of 
these thermometers had been heated several times to over 200° C., and especially 
when heated more than once to nearly 300° C., their indications began to be per- 
manently altered in the way of increase ; and in some of them which had been 
exposed in the holes or bores, closest to the source of heat, where they had 
been often raised to a temperature of 300° centigrade, it was found that the 
permanent alteration of zero was as much as 5 degrees. As it appeared that the 
probable nature of the distortion was a permanent shrinkage of the bulb, I 
calculated what should on that supposition be the behaviour of the instrument at 
different temperatures, and by comparing its indications step by step with those 
of another of the thermometers which had not been distorted by violent heating, 
I found the results of calculation verified. The altered instrument loses slightly 
* Katract from a letter, Dr Stewart to Prof. Tarr. 
“‘ Kew Opservatory, 8th December 1868. 
.... “We have come to the conclusion that each instrument ought to go down as low as the 
freezing point. 
“Tt 1s possible, no doubt, starting with an instrument that includes the freezing point in order to 
determine the graduation constants, and afterwards taking out some mercury, to produce instruments 
that begin to register only at high temperatures. But there is an element of uncertainty introduced in 
taking out the mercury, which may not only cause a constant error, but an error of scale value... . . 
(Signed) “B, Stewart.” 
