250 Mr. A. T. Mukerjee on a Method of Measuring 



Let E c -, E K , E Y , and E„ be the probable errors of the 

 corresponding quantities. Then the probable error of the 

 result E c ' is given by the following equation : 



Taking the approximate numerical values K = 162, V = 110, 

 ii = l, E K = + 0*6 (determined experimentally as shown below) ; 

 E y? being the combined errors due to the voltmeter used and 

 to the leakage during the interval between the charging 

 and the sharing of the charge, may be taken at not more 

 than 0'5 volt; E Vi the error in measuring the final voltage by 

 the electrometer, is certainly not greater than ±0*002 : 



(E c -) 2 = (-006) 2 + C007) 2 + (-002) 2 = C009) 2 . 



Determination of K, the capacity of the quadrant system 

 up to the connecter :■ — 



The system was insulated and charged by a standard 

 cadmium cell and the charge shared with a capacity of 92*1 

 E.S.U. on a standardised sliding condenser having a very 

 accurate Reichsanstalt certificate. It was necessary to 

 attach a mercury cup to the connecting wire of the standard 

 condenser. This introduced a slight correction to the values 

 read off from the calibration curve of the condenser. This 

 has been called x and determined as shown below. The de- 

 flexions of the electrometer before and after sharing are 

 as follows : — 



340-0 338-3 336*2 335*0 333*0 331*5 

 99*2 99-0 98*5 97*8 96*7 96*5 



We get from the above 



tl '= 2*427, 2*417, 2*413, 2424, 2*444, 2 435 



K 



= 2*427 (mean) (1) 



The standard condenser was then shifted to a position 

 corresponding to a value of 31*0 E.S.U. on the curve and 

 the measurements repeated with the following results : — 



330*0 



328*6 327*8 



326*0 



324*3 



323*0 



177-0 



1760 175*2 



174*7 



174*0 



173*0 



