Dielectric Constants of certain Frozen Electrolytes, 303 



suffices to reduce the galvanometer reading to the value of the 

 dielectric constant of the dielectric used referred to that of gaseous 

 air as unity. 



The galvanometer used in these experiments was a Holden-Pitkin 

 galvanometer with movable coil. The scale was placed 125 cm. from 

 the mirror, and the sensibility was such that an electromotive force 

 of 1*434 volts, acting through a megohm, gave a scale deflection of 

 3'6 cm. 



In making the measurements a careful watch was kept to see if 

 the charge current of the condenser was always equal to the dis- 

 charge current ; and the moment the charge and discharge galvano- 

 meter deflections differed by anything like a millimetre or two of 

 scale deflection (the scale being 125 cm. from the galvano- 

 meter) the capacity measurements were no longer trusted. More- 

 over, at intervals, a large resistance of several thousand ohms was 

 inserted in the galvanometer circuit to ascertain if there was any 

 variation of deflection. The absence of any change in deflection 

 was also taken as the proof that no sensible leakage current was 

 passing through the dielectric, and polarisation practically absent. 



From the construction of the condenser it will be seen that a con- 

 siderable vertical movement of the inner cone only changes the 

 perpendicular distance between the surfaces of the cones by about 

 one- tenth of the vertical rise. When a liquid is placed between the 

 cones and frozen there is a change of distance due to the lift of the 

 inner cone. We have, however, convinced ourselves by experiment 

 that the change in the capacity of the condenser which may occur 

 owing to change in the bulk, either expansion or contraction, of the 

 dielectric cannot make more than 5 per cent, change in total capacity 

 measured. This is not more than may easily occur through unavoid- 

 able degrees of impurity in the dielectrics used. We have not 

 thought it necessary to apply a correction for the change in dimen- 

 sions of the dielectric thickness on cooling.* 



In reducing the observations a correction has been made for the 

 capacity of the connecting wires and for that of the vibrating 

 contact maker,f and in the following tables the reduced and cor- 



* There is also a correction due to the fact that some of the lines of electrostatic 

 induction start from the inside of the inner cone and terminate on the carrier or 

 other portions of the outer cone. In other words the whole of the capacity of the 

 system is not comprised between the opposed surfaces of the cones. This, how- 

 ever, is not a large correction, as we have found that even if it is omitted we obtain 

 results for the values of certain well-known dielectric constants closely agreeing 

 with those of other observers. Our object in these experiments has not been so 

 much to obtain great numerical accuracy of value as to determine the general mode 

 of variation of the dielectric constant with temperature. 



f See ' Roy. Soc. Proc.,' vol. 61, p. 12, Dewar and Fleming, " On the Dielectric 

 Constant of Ice and Alcohol at very low Temperatures." 



