MEASUREMENT OF A QUANTITY OF ELECTRICITY. 557 



The means of five series of experiments gave the following 

 numbers : 



i-i94- IO ~ 6 36*06. io 3 30*20. io 9 



1-300 4 J '94 3 2 ' 26 



1-568 49-70 31-70 



1-480 44-35 29-96 



1-586 49-66 31-25 



Mean . . 3i'o7.io 9 



This experiment presents the greatest difficulties. The law of 

 distribution assumed for the contact of the small ball and of the 

 sphere would only be exact provided the two conductors were 

 withdrawn from any extraneous action, a condition which cannot 

 be realised. 



The rectangular shade was 144 cm. high by u6cm. in length 

 and 87 cm. in breadth. These dimensions are so large that the 

 influence of the sides need not be allowed for, especially in mea- 

 suring charges (803). 



The greatest source of error is probably due to phenomena of 

 absorption and to residues of the Leyden jar. Kohlrausch,* in a 

 preliminary investigation, had endeavoured to determine the law 

 of the residue and the loss in air for the jar used for the experi- 

 ment. For a charge q of the jar, the charge q disposable after a 

 time /, is equal to the excess of the initial charge over the loss by 

 air a t and the residue b t . Kohlrausch found for b t the expression 



in which g t = g Q -a tt and in which the values of the coefficients are 

 p = 0*04494, = 0-1834, m = 0-4255, this latter only depending on 

 the dimensions of the jar. 



We could also allow for the variations of the charge due to 

 the loss and the residue, first during an interval of 40 seconds, 

 which elapsed between the two electrometric measurements, and 

 then during a second interval of 3 between the period of this 



* KOHLRAUSCH. Pogg. Ann., Vol. xci., p. 56. 1854. 



