INDUCTANCE AND CAPACITY 357 



The active portion of any condenser intended for use as a 

 standard must be firmly confined between clamps, so that its 

 geometry and, consequently, the capacity of the condenser may 

 be definite. Condensers without clamps are greatly affected by 

 temperature, and when taken through a cyclic variation of 

 temperature (for instance, 17, 30, 17), do not return to their 

 initial capacities. This permanent alteration may be as much 

 as 3 or 4 parts in 10,000. 



Condensers on Direct-current Circuits. In the use of conden- 

 sers with direct currents, difficulties arise from "absorption" 

 and its related effects. It is found that the discharge of any 

 condenser having a solid dielectric consists of two portions 

 a sudden rush of current at the instant of closing the circuit, 

 due to the free charge, and a small, gradually decreasing current, 

 due to the liberation of the "absorbed" charge. This latter 

 current complicates the various methods of measurement when 

 direct currents are employed (see "Direct-deflection Method," 

 page 369). If high voltages be used, the absorbed charge con- 

 tinues to be given up for a long time. 



When the condenser is charged, the first rush of current 

 consists of two portions one furnishing the free charge, the 

 second a diminishing current furnishing the absorbed charge. 

 This latter current, for a short time, about 0.01 sec., may be rela- 

 tively large. If the charging circuit be broken too soon, before 

 the dielectric is "saturated," the absorption goes on, and if there 

 is a delay in discharging the condenser, the free charge will be 

 diminished below its proper value. Thus the apparent capacity 

 depends on the previous history of the condenser, on the 

 time of charging, on the length of time between disconnecting 

 from the battery and discharging and on the time of discharge. 



An arbitrary measure of the absorption may be obtained by 

 subjecting the condenser to a definite series of operations, for 

 example, by charging for 1 sec., insulating for 30 sec., discharging 

 instantaneously through a ballistic galvanometer, insulating 

 for 30 sec., discharging again, and so on, until five residual de- 

 flections have been obtained. The measure of the absorption 

 is the total quantity in the five residuals expressed as a fraction 

 of the free charge. The absorption curves in Fig. 212 were 

 obtained in this manner. 



