SHUDDEMAGEN. — RESIDUAL CHARGES IN DIELECTRICS. 519 



Several of the current curves, if plotted out for each interval of the 

 charging time, would show a depression such as has been shown in the 

 case of the cover glass condenser (Figure 12). It is barely possible 

 that this peculiar result may be genuine and indicate a "backward 

 surge " of the extra dielectric polarization which is conditioned by the 

 molecules of the dielectric. But it is more likely that it is due to ex- 

 perimental error in the estimation of the charging time and perhaps in 

 the reading of the ballistic throws. The peculiarity occurs sometimes 

 in the charging interval corresponding to the second thickness of the 

 type metal strip and sometimes in that of the third thickness. In 

 either case the experiment is extremely delicate and one would expect 

 a slight shifting to occur. 



Conclusion. 



The results of this research, as shown graphically in the "current 

 curves of the Figures," prove clearly that the current which forms 

 residual charge, or, in other words, the "absorption current," is far 

 from negligible when the charging interval is very small. Not only 

 is the current very large, but the residual charge which it forms within 

 0.0017 of a second after charging begins, is of the order of several per 

 cent of the " free charge." Glass and paraffined paper condensers show 

 the greatest residual charge formation for short charging times. In 

 each of the two mica condensers which were tested the residual charge 

 which is formed in 0.0017 seconds is only one-half of 1 per cent of the 

 " free charge." But, on the other hand, the mica condensers exhibit 

 an absorption current which decreases but slowly with the time, so 

 that for long-continued charging they may take up much more residual 

 charge than the paraffined paper condensers, whose absorption current 

 is very large at first but decreases much more rapidly as the time in- 

 creases. The glass condenser shows both a high residual forming 

 current immediately after the beginning of the charge and a rather 

 slow decrease as the time increases. To give a striking example of its 

 high initial value, we may note that during the charging interval from 

 0.00007 to 0.00020 seconds its average value is such that if it con- 

 tinued uniformly for one second, the condenser would get a total 

 residual charge equal to one hundred times the total "free charge." 



It thus appears that the conception of " free charge " is not a very 

 convenient one, for various investigators have shown that the law 

 of superposition holds true, at least to a very close approximation, and 

 this law gives the corollary that if a condenser has been charged for a 

 long time with a constant potential difference and is then discharged, 



