308 M. R. Kohlrausch's Theory of the 



§2. 



The question, to what other source than the loss of electricity 

 in the air is the quick diminution of the tension of the knob of 

 the Leyden jar at the commencement to be ascribed, finds a 

 kind of answer in the fact, that after complete discharge has 

 taken place, a short time having elapsed, a charge more or less 

 considerable, of the same kind of electricity as that with which the 

 jar was originally charged, makes its appearance. From this we 

 may conclude that a portion of the electricity first communicated 

 has concealed itself somewhere, so as not to contribute to the 

 tension of the knob, and was prevented at the same time from 

 being a sharer in the discharge ; to this circumstance we might 

 be inclined to refer the more speedy diminution of the tension 

 than that which could be referred to the loss in the air. 



Besides this, the following efi"ect is observed. If a charge 

 has remained in the jar for a considerable time, and a consider- 

 able portion of the electricity be taken from the jar, so that the 

 tension becomes feebler by sudden springs, then immediately 

 afterwards this tension is observed to sink very slowly, or some- 

 times to stand for a time and then sink, or, which is mostly the 

 case, first rise for a time, then stand, and then sink finally. 



Combining the facts of the last two paragraphs, we are com- 

 pelled to divide the entire quantity of electricity, Q, in the jar 

 into two portions, one of which, L, can be discharged, and which 

 on this account may be called the disposable charge, and the 

 other r, which is prevented from sharing in the discharge, and 

 first makes its appearance after the removal or diminution of the 



If by observation we have 



Lo= 1 •4968 and L<=0-5266, 

 then 



«=0-0001945; 

 and as, on tlie other hand, 



we can readily calculate the value of the charges for every t between the 

 beginning and the end, and lay down the ciu-ve represented by the dotted 

 line. In this way we learn that for the times t the cur^'e must possess the 

 following ordinates 0. 



