RADIATION IN ABSOLUTE MEASURE. 
435 
great indeed. Now it is, perhaps, scarcely realised that the time taken for a high 
vacuum (yg- M, one-tenth of one-millionth of an atmosphere, for example) to equalise 
itself throughout the narrow glass-tubes used for connections with the Sprengel 
pump is very considerable, and may amount to 20 minutes, half-an-hour, an hour, or 
two, according to circumstances. It follows from this that the indications of the 
McLeod gauge are certain to be behind time in the case of a high, but variable, 
vacuum ; and the introduction of any considerable length of tubing that can possibly 
be avoided is very disadvantageous. The difficulty here referred to has proved a more 
fruitful source of trouble, and seems to me less likely to be successfully dealt with, 
than any other connected with these experiments ; though it is just possible that the 
use of platinum-iridium wire or platinum-silver may prove of considerable advantage. 
It has been already explained that the temperature at which the radiation is taking 
place is,determined in these experiments by finding the resistance of the wire at the 
moment, and by knowing from separate experiments the resistance of the wire at 
different temperatures. The variation of electric resistance of platinum wire with 
temperature is very different in different specimens. It seems likely that small 
impurities in the platinum, in the way of admixture with it of minute quantities of the 
iridium class of metals, may be the cause of this variation. I, myself, have found that 
a very minute quantity of tungsten combined with German silver makes a vast differ¬ 
ence as to the variation of resistance with temperature of that alloy.* The great 
variableness as to this quality in platinum wires is abundantly attested by many 
experimenters,! and my own experiments fully confirm those of others. I have, 
therefore, in every case taken a portion of the radiation wire itself, and determined its 
resistance through as wide a range of temperature as was attainable in the way to be 
described presently. The results of these experiments are laid down in the form of a 
curve having temperatures as abscissas and resistances as ordinates; and from this 
curve, which turns out to be almost a straight line with different inclinations in 
different specimens, the temperature of the wire in any experiment is taken off, its 
resistance being known. 
The apparatus used in determining the resistance of platinum wires at high 
temperatures is shown in fig. 4, a, b, c. In my earlier experiments, and at lower 
temperatures, I employed the vapours of liquids with high boiling points as first 
proposed by Andrews, \ some of the liquids suggested by Drs. Ramsay and Young § 
proving convenient. The boiling points of liquids, however, would not give tempera¬ 
tures nearly high enough for the purpose in hand, and I was about to construct a 
* “On the Electric Resistance of anew Alloy named Platinoid,” ‘Roy. Soc. Proc.,’ vol. 38, 1885, 
p. 340. 
t Siemens: Bakerian Lecture for 1871, but first published in the ‘Transactions of the Society of 
Telegraph Engineers ’ for 1874. Schleiermacher : ‘Wiedemann, Annalen,’ vol. 26, 1885, p. 295. 
H. L. Callendar : ‘ Roy. Soc. Proc.,’ vol. 41, 1887, p. 231. 
+ ‘ Comptes Rendus,’ vol. 81, 1875, p. 279. 
§ ‘ Chem. Soc. Journ.,’ 1885. 
3 K 2 
