186 
MR, H. L. CALLENDER ON THE PRACTICAL 
Table A-i.—C ontaining some of the Results of Observation with the experimental 
Air Thermometer No. 1. 
Number of 
observation. 
Temperature by 
air thermometer. 
Resistance observed, 
reduced. 
Resistance calculated 
by formula (a). 
Resistance calculated 
by exponential 
formula. 
1 
88-0 
1-304 
1-304 
1-304 
2 
96-9 
1-335 
1-335 
1-335 
3 
49-9 
1173 
1173 
1-172 
4 
205-5 
1-707 
1-711 
1-708 
5 
210-2 
1-723 
1-727 
1-724 
6 
280-3 
1-960 
1-970 
1-958 
7 
138-7 
1-480 
1-481 
1-481 
8 
179-1 
1-620 
1-620 
1-619 
Observations.—Series II. 
Owing to the success of the experimental instrument, a slightly improved form was 
devised and ordered of a London glass-worker about the beginning of December, 
1885, for which I wound a platinum spiral of the same wire at the Cavendish 
Laboratory; the first spiral was broken and another sent, and other casualties 
intervened, so that the instrument did not arrive until February, 1886, and then, 
owing probably to some want of clearness in the specification, I had to alter it in 
some essential details before it could be used. It was cleaned and calibrated in the 
usual manner, and finally, on February 13, it was filled with dry air, at a temperature 
of about 350°, and a series of observations, recorded in Table A-il, were taken with 
it in a large mercury vapour bath; the fixed points being determined with the aid of 
a hypsometer and melting ice. 
On February 17 another series of observations was begun, but the zero pressure 
had increased in four days from 34'10 cm. to 36‘09 cm., owing to leakage, the 
platinum wires not having been fused air-tight through the glass. In an attempt to 
remedy this somewhat serious defect, the instrument was unfortunately broken. This 
was much to be regretted, as less reliance can be placed on observations taken while 
the temperature is varying; all the observations are given, however, as they show 
how perfectly the method works. The time observations show the rate of variation 
of the temperature. A mercury thermometer in the same enclosure often differed 
from 20 to 50 degrees, in its indication, from the air thermometer. 
Values of R are calculated by both formulae ( b ) and (e) for comparison. The 
agreement with the formula (e) is in some cases so close that it must be fortuitous to 
some extent; it is, however, satisfactory to see that single observations are so 
consistent, and give so smooth a curve. 
