182 
MR. H. L. CALLENDAR ON THE PRACTICAL 
of Pi/Pq, uncorrected, for surface condensation, to be 1 ‘3693, which very nearly agrees 
with that actually observed (see Table A, III.) [For another explanation see 
Appendix, p. 222.] 
It appears, however, that when the ah' film has been very completely removed bv 
exhaustion at a high temperature it increases gradually for some time before it has 
recovered its normal state. 
To illustrate this, the following Table gives the changes of zero pressure observed 
with two of my air thermometers. 
Table. %. 
Date. 
Zero pressure. 
Oxygen 
thermometer. 
Nitrogen. 
thermometer. 
1886. Jan. 21 . . 
69-31 
69-54 
„ 22 . . 
69-29 
69-51 
„ 23 . . 
69-29 
69-49 
„ 25 . . 
69-20 
69-38 
„ 25 . . 
69-22 
69-41 
Remarks. 
Four days after filling by exhaustion at 300° C. 
After heating to 100° C. 
After heating to 100° C. 
Such changes of zero pressure are a most important source of uncertainty, especially 
if the instrument be used at high temperatures. 
We have said a great deal about the air thermometer, because the chief experi¬ 
mental difficulties arose from it ; the other essential part of the apparatus is the 
platinum spiral, which we will now proceed to consider. 
The Platinum Spiral. 
The spirals used in this investigation were all made from the same reel of pure 
platinum wire supplied by Messrs. Johnson, Matthey, and Co. Its purity is attested 
by the following facts : (l) the high temperature coefficient of its resistance, ‘00346, 
instead of about ‘0030, as with ordinary platinum ; (2) it was very infusible and 
unalterable : after being kept for about two hours at a temperature of about 1200° C., 
its resistance had not altered by 1 in 5000 at 0° C. ; (3) no impurities could be 
detected by chemical analysis. 
Details of the wire. —Diameter, about ‘017 centim. ; resistance, about 5 ohms per 
metre ; mass, ‘00488 gramme per centim. About 1 metre of the wire is measured off 
and its ends fused on to tails of thicker platinum wire (diameter, ‘073 centim.) in the 
oxy-hydrogen blowpipe. (It is less trouble to solder with copper, but the spiral 
cannot then be boiled with acids in the bulb of the ah' thermometer.) The wire is 
then annealed by passing it slowly through the solid Bunsen flame ol a Fletcher 
