Jan., 1916] Electrical Behavior of Porcelain and Glass 87 



creepage over the surface occurred the leakage current would be 

 diverted to earth and the ammeter would register only the 

 current passing through the cylinder specimen, i. e., the Number 

 One type. It was not convenient to make this cylinder of 

 the same dimensions as the one previously used, but the 

 characteristics of the current time curve were the same as in the 

 previous cases. This would indicate that the characteristic 

 curves of all were of the same kind and were due to the volume 

 conduction of the specimens rather than leakage over the 

 surface. 



Discussion of Results. 



A few preliminary trials indicated that it was not possible 

 to fix the resistance of the porcelain or glass bodies by the 

 simple application of the formula suggested earlier in this 

 paper. The current strength was found to be a function of the 

 material, dimensions, temperature and potential applied, also 

 it depended upon the length of time of application of the e. m. f. 

 The apparent resistance could have wide variations in value 

 which seemed to be limited only by the time of application of 

 the e. m. f. We are accustomed to think of the resistance of a 

 conductor as having a definite and fixed value determined by 

 the dimensions and temperature of the conductor. Our legal 

 definition of the ohm is of this kind. In such a sense these 

 ceramic bodies at temperatures where they become appreciably 

 conducting have no definite resistance. 



The usual method of study was to obtain a predetermined 

 temperature and maintain it at a constant value through a 

 single experiment. A potential was applied which was left 

 constant and the current noted at time intervals. Figure 4 

 shows a typical curve. This was made on glass, with the 

 guard ring arrangement, hence current values are for those 

 which must pass through the glass. 



The current decreases initially very rapidly with time and 

 then at a much diminished rate. The units shown as ordinates 

 are 10-^ amperes and on this scale the values may be shown 

 over a long period. The first value shown in Figure 4 is one 

 taken one minute after the potential was first applied. The 

 earlier values are much larger and cannot be shown on this 

 scale. The general procedure was to take current readings 

 every 15 seconds during the early stage for say five minutes, 



