230 
MR. H. L. CALLENDAR OR THE MEASUREMENT OF TEMPERATURE. 
The resistance of the platinum-silver wire was found to be increased about 5 per 
cent, by annealing for three hours at a red heat till it reached a steady state ; wffiile its 
mean coefficient between U° and 100° fell from ’000279 before annealing to ’000236 
after. For this reason it seems desirable that the wire used in the construction of 
standard cods should be well annealed. 
The resistance-variation of this specimen of platinum-silver between 0° and 650° C. 
was well enough represented by the formula 
P v /R 0 = l + -000232 7t + 0-000000033^. 
Comparisons at High Temperatures. 
The coil used in the experiments recorded in Table P.-i., p. 202, was tried in the 
porcelain tube alone, without the non tube, at the highest temperature the gas-furnace 
would produce with the air blast. The insulation was too bad and the thermo-electric 
effects too large to allow of accurate measurements : and so, wffien the apparatus cooled, 
a copper tube was inserted inside the porcelain. On again heating the tube the 
thermo-electric effects had disappeared, but the resistance was found to change 
enormously. When cold, the resistance of the wire was doubled, and it had become 
quite brittle. The copper tube came out as bright as if it had been heated in 
hydrogen ; metallic copper was deposited all over the cooler parts of the electrodes 
and terra-cotta; the porcelain tube was bent with the heat. It is, therefore, 
necessary to try some other method at high temperatures. The apparatus is not yet 
complete, but promises well. I hope soon to be able to extend the comparisons 
accurately to 1000° at least, 
Note. 
[The experiments on surface condensation, mentioned on page 181, were not sufficiently 
precise to be applied to the case of the air thermometer; they only served to confirm 
the results of recent experiments, to show that the phenomenon is dependent in great 
measure on the past history, and to emphasise the importance of eliminating aqueous 
vapour and C0 3 . It has not been thought worth while to use Equation (E), p. 179, in 
the reduction of the observations, partly because they are not yet sufficiently accurate, 
and partly because it is doubtful whether Equation (E) itself is strictly applicable 
beyond the limits of the experiments on which it is founded. The corrections it 
involves are of the same order as the errors of observation, and anyone who thinks it 
worth while can easily apply them.—July 11, 1887.] 
