It will be seen that the mean galvanometer throw, when the condenser was immersed in liquid oxygen, was 10'903 centime., and the mean throw, when raised into the gaseous oxygen, was 7*146 centims.
One matter which we felt it important to examine, was whether there was any correction needed for the change in the dielectric constant of the glass separators with temperature.
Since these glass separators had a total surface of nearly 1 per cent, of the area of the metal plates, the condenser may be regarded as consisting of two condensers joined in parallel, one consisting of a glass dielectric condenser having an effective surface of 1, and the other a condenser having a liquid or gaseous oxygen dielectric having an effective area of 99. In the course of these experiments we have therefore examined the effect of low temperature upon the dielectric constants of glass, paraffined paper, and mica. Wo find that on cooling these bodies to —182° C. they experience a marked reduction in dielectric constant. The dielectric constant of a certain specimen of crown glass was reduced by 21*4 per cent, by cooling to the temperature of liquid air or to —185° C. The dielectric constant of paraffined paper was reduced by 28'4 per cent, under the same circumstances.* We are engaged in a systematic examination of the influence of very low temperatures on the dielectric constants and specific resistances of the principal dielectric bodies. The crown glass used as separators in the construction of our small condenser had a specific inductive capacity of about 6*0 at ordinary temperature, and this at the low temperature would be reduced to nearly 5*0. Hence in estimating the capacity of the condenser, as constructed, there comes in as we have seen a correction from the presence of the glass. We selected glass in the first instance rather than ebonite or sulphur, as we thought it probable we should use the same condenser in determining other dielectric constants, and we wished to construct the separators of a material which was very rigid and not easily acted upon by oils or other liquids.
Taking the formula above given, we can deduce from the observed results the required constant, for, wTe have
0 K + O-Oo 1030 o' 1*05 1019 ’ K = 1*491
- By another method we have found th a t for the glass of a glass test-tube the
dielectric constant was decreased 22*2 per cent, by cooling to the tem perature of liquid air. U nder the same circumstances a certain specimen of mica decreased only 3'01 per cent, in dielectric constant.
