500 Dr. H. H. Poole on the Electrical 



were not very concordant, but indicated clearly a rise of 

 conductivity from about 2*5 at very small gradients to about 

 9 at 0*2. This means a slope of the logarithmic isotherms 

 of about 2 7, i. e., somewhat greater than that of glass or 

 mica. 



A film of shellac was also cast by the same method as 

 that employed with paraffin-wax, but owing to the viscosity 

 of the melted shellac it was impossible to obtain a thin film 

 by this method. The film used was 2*0 mm. thick and 

 failed under a pressure of 8000 volts. As this only corre- 

 sponds to a gradient of 0*04 it seems probable that a flaw 

 existed. No current was detected through this film indi- 

 cating that the conductivity did not exceed -2 X 10 ~ 3 , only 

 about iooq part of that of the film obtained from alcoholic 

 solution. 



Celluloid. 



An old piece of Kodak film 0*080 mm. thick from which 

 the emulsion had been removed was used, brass contact- 

 pieces and a brass guard-ring being gummed to its faces. 

 Tests were carried out at 1I 0, 3 C, the pressure being raised 

 by small steps until piercing occurred. This happened at 

 1250 volts (corresponding to a gradient of 0'16) after the 

 voltage had been steady for long enough to enable readings 

 to be obtained. The readings, including the last, were very 

 concordant, the logarithmic isotherm being a straight line 

 represented by the equation log K=1*33 + 3'8X, so that the 

 conductivity increases much faster with gradient than that 

 of any other dielectric tested. It is interesting to note 

 that the conductivity of celluloid obeys a similar law to 

 those of the inorganic dielectrics mica and glass, in which 

 the conduction might possibly be of an electrolytic nature*. 



Tabulation of Results. 



The various results are tabulated below. Here t is the 

 temperature centigrade. A and B are constants in the 

 formula log K = A-|-BX, where K is the conductivity, the 

 unit being 10 -12 (ohm) -1 (cm.) -1 , and X the potential 

 gradient in megavolis per centimetre. X m is the value of X 

 which pierced the specimen. In cases where the specimen 

 was not pierced the highest gradient attained is given as a 

 minor limit. R is the resistivity in ohm centimetres at very 

 small values of X as obtained by extrapolation. 



* Evidence against this view has since teen obtained. See 'Nature/ 

 July 7, 1921, p. 584. 



