184 
MR. H. L. CALLENDAR ON THE PRACTICAL 
and the resistance box. They were always very tight and clean, and cannot have 
introduced appreciable errors. The plugs of the resistance box were very large and 
truly ground, and their contacts good. 
It was suggested that the conductivity of the glass at high temperatures might 
prove a source of error in the resistance measurements. A rough experiment was 
therefore tried. A cubic millimetre of the hard glass at a dull red heat had a 
resistance of about a megohm, and about ten times less at a bright red heat. This is 
less than 1 in 10,000, and may be neglected. Moreover, since only one of the 
platinum terminals was actually fused through the glass in this case, the bad contact 
further diminished any possible error from this source. Current heating of the wire 
would produce a small and nearly constant error; since, however, the current was 
never greater than ‘01 ampere, the effect must have been almost negligible. From 
experiments on the cooling by air contact of this particular kind of platinum wire, I 
am in a position to estimate the excess of temperature due to a current of '01 ampere 
at about ‘016° Centigrade. 
Reduction of the Resistance. 
There are some small corrections to be applied. 
The resistance of the copper wires connecting the coil to the resistance box is 
generally about '0 I ohm, and varies very slightly with the temperature. The whole 
correction amounts to about 1%, and may be measured on each occasion by means of 
double electrodes. 
The temperature of the resistance box varies slightly. A special experiment gave 
•00040 as the value of the temperature coefficient. The box is correct in B.A. units 
at 14'2° Centigrade. The temperature is therefore taken by a mercury thermometer 
kept inside the box, and the observed resistance reduced to 14‘2° Centigrade. This 
correction is also very small, and can be applied with tolerable certainty. 
Finally, to render the results comparable with each other, each observed resistance 
of the platinum coil is reduced to what it would have been for a coil having the 
resistance of a unit at 0° Centigrade, by dividing by the observed value of the 
resistance of the coil at 0° Centigrade. 
Explanation of the Tables of Comparison of Platinum and Air Thermometers. 
The letters at the head of each column have the same meaning as explained in 
section entitled Formuhe (see page 176). The column headed Thermometers gives the 
observed temperatures B of the barometer, v of the air near the U gauge, Hg of the 
mercury in the right limb of the manometer, R of the air in the resistance box ; 
v, Hg, R, being taken with the same mercury thermometer.* Each line contains the 
* [Tliis is bad, and may introduce small errors : latterly a separate thermometer was set apart for 
each purpose.] 
