JAN. 19, 1923 WHITE: ELECTRIC HEATING OF CALORIMETERS 27 
is 5 times as great; not so if the 5 amperes is used to give the same heat, 
at lower voltage. The limits of toleration of air loss for each case are 
shown in the second and third sections of table 1. For a given size of 
wire the proportional error is constant for a given coil resistance, 
regardless of current or voltage. The error from lack of symmetry in 
the thermal contacts at the two ends of the free portion of the leads 
may also become serious. For 1 cm. difference, 5 amperes and 110 
volts, this error is nearly 40 per million for No. 24 wire, which there- 
fore seems the very smallest that should be used for 100 per million 
precision with coils of as low resistance, 22 ohms, as is required for 
this current and voltage. Another advantage in using slender leads 
is that they can be removed from the calorimeter with but little change 
in the total heat capacity. This is assuming that the heater itself 
remains, so that any uncertainty there may be as to its heat capacity 
does not affect the application of the result of the calibration. 
The conductivity of the potential leads may disturb the symmetrical 
distribution of the heat generated in the current leads. This can be 
avoided by making the potential leads symmetrical with respect to 
calorimeter and jacket, as by having one lead run from the neutral 
point into the calorimeter, there turning around (still insulated elec- 
trically) to run entirely across the gap and out through the jacket. 
Thus far the upper part of the heater case and the leads within it 
have been taken as at the calorimeter temperature, except for a slight 
correction due to the effect of heat coming in from the free part of the 
leads. If the upper part of the case or the leads are heated above 
the calorimeter temperature by the heating coil itself, the only way of 
avoiding an uncertain correction appears to be to run the leads a 
sufficient distance in close contact with some conducting body which, 
by immersion in the water or attachment to the calorimeter wall, or 
in some other way, is kept at the calorimeter temperature. If only the 
leads are heated and not the top of the case, it may be sufficient merely 
to have the leads run a considerable distance before leaving the case. 
The required distances can be calculated by the formulas of the 
section on Conduction through the Jacket, in the paper already referred 
to. For example, the following results were thus obtained for No. 24 
wire, running 5 cm. in the case and 6 cm. through the air to the jacket. 
For this wire » in air has been taken as 0.28. For the wire in close 
thermal contact with a plate on one side only, as it would be if wound 
around a central mica strip in a flat case, » can be taken as 0.040 
Then if @ is the temperature of the inner end, next the heater coil, of 
the 5 cm. of wire, referred to the calorimeter-air-jacket temperature as 
