CONTINUOUS ELECTRIC CALORIMETRY. 
87 
to absolute electrical measurement. It should be noticed that the result above 
deduced from the first series of observations by Mr. King agrees with that deduced 
from the observations with the electrical calorimeter by comparison with Howland 
and with Reynolds and Moorby, namely, 1 '4332, within the limits of probable error 
of the several methods.^ 
Part III. —Thermometry. 
(18.) The Compensated Resistance Box. 
Nearly all the temperature measurements in this investigation were made with a 
special form of resistance box, which contains some devices which have not as yet- 
been adopted in any other instrument of its class, or described in any scientific 
periodical. The most important feature in its construction was the system adopted 
for compensating the resistances to eliminate the effect of change of temperature. In 
the usual form of box the temperature of the coils is taken by means of a mercury 
thermometer, and the correction for change of temperature applied from a knowledge 
of the temperature-coefficient. The principal objection to this method is that the 
mercury thermometer cannot follow the temperature changes of the coils with 
sufficient exactness, and that the temperature is generally far from uniform through¬ 
out the box. In the method of compensation which I patented in 1887 (Complete 
Specification No. 14,509) the temperature correction is automatically eliminated by 
combining with each of the resistance coils proper, which are made of platinum- 
silver or some other alloy possessing a small temperature-coefficient, a compensating 
coil of copper or platinum having a large coefficient. The resistance of the compen¬ 
sating coil is adjusted so that its change of resistance per degree is equal to that of 
the coil it is intended to compensate, while its actual resistance at any temperature 
is much smaller, the ratio of the resistances being inversely as the temperature- 
coefficients. Each compensator is placed in the box in close proximity to the coil it 
is intended to compensate, so as to be always at the same temperature, but coil and 
compensator are connected on opposite sides of the bridge-wire, so that the balance 
depends only on their difference, which remains constant for any change of tempera¬ 
ture, provided that the adjustment has been properly effected. The advantage of the 
method lies in the fact that this adjustment can be made with extreme accuracy, by 
testing coil and compensator together over the required range of temperature before 
they are connected in their places in the box. But the method has not come into 
general use, partly on account of the labour involved in the adjustment of the coils, 
and also in part owing to the discovery shortly afterwards of manganin and other 
alloys of small temperature-coefficient, which are fairly satisfactory for ordinary work 
though inferior, in my opinion, to the compensation method for work of precision 
The objections to manganin, for instance, are—(1) That it cannot be perfectly 
* The details of the construction and comparison of the Clark and cadmium cells employed are 
sufficiently described in the ‘ Proc. Roy. Soc.,’ 1897, vol. 62, p. 117, and by Dr. Barnes, Section 3a, p. 159. 
