CONTINUOUS ELECTRIC CALORIMETRY. 
85 
the mean radius of the small coil in his current balance. The chief difficulties were 
due to change of resistance of the coils with change of temperature. These changes 
were considerably reduced by employing copper resistances for the adjustment, and 
by arranging the details so as to secure the greatest rapidity of observation. 
Readings were taken in each position with the large and small coils severally 
reversed, and interchanged, and replaced, the complete series each night including 
36 independent readings of the ratio. The distance between the two positions of the 
small coils on either side of the centre is the most important measurement. This 
was determined by two shoulders turned on the thick brass tube, the distance 
between which was measured with the callipers. It was necessary to make all these 
comparisons at night between the hours of 2 and 5 a.m. in order to avoid disturbance 
from the electrical railway, two lines of which passed within about a quarter of a mile 
of the building. The values of the ratio obtained from observations on three different 
nights, when there was no magnetic disturbance and the conditions were otherwise 
satisfactory, showed extreme differences, amounting to nearly 1 part in 5,000. These 
may have been partly due to temperature, as the small coils had been boiled in 
paraffin, which must have affected the expansion. 
(16.) Ratio of the Currents in the Coils. 
In the usual method of employing the electrodynamometer, the same current is 
passed through the large coils and small coils in series. This method has the 
advantage of simplicity, but it is essential, in order to obtain steady deflections and 
secure the maximum accuracy of reading, that there should be no appreciable heating 
of the small coils by the current. The latter condition, however, cannot be satisfied 
if the currents are equal, unless the large coils are wound with very fine wire, which 
is for many reasons objectionable. In the present case, the windings were designed 
to secure approximately equal current heating when the currents are in the ratio of 
1 to 10 in the small and large coils respectively. The ratio of the currents was 
adjusted to this value at each observation by having a standard ohm in series with 
the small coils, and a standard tenth-ohm in series with the large coils. 
The whole arrangement formed a Wheatstone bridge, the balance of which was 
adjusted by means of some copper coils and a high-resistance shunt in series with 
the small coils. The standard coils were of manganin immersed in oil, and the 
ratio remained extremely constant, owing to the equality and smallness of the current 
heating, and the perfection of the insulation. Thus, although the division of the 
current involved an additional adjustment of the ratio of the resistances, no 
appreciable error was thereby introduced, and great advantages were secured by 
the equable distribution of the heat developed, and the steadiness of the observed 
deflections. 
