238 



Journal of the Asiatic Society of Bengal. [June, 1914. 



In the earlier series the electrometer needle was of the original 

 paper type, and the key was separate, its capacity together with 

 connections being about 21 E.S.U. In the later measurements 

 an aluminium needle and the later type of key were used, the 

 capacity of the latter being about 8 E.S.U. 



Measurements of capacity by the method of mixture prob- 

 ably do not reach as a rule a higher standard of accuracy than 

 that of the earlier series quoted above. For instance, the figures 

 given by F. C Brown, 1 in a recent paper on " A practical elec- 

 trical method of measuring the distance between parallel con- 

 ducting planes" show that the values of consecutive measure- 

 ments varied by several electrostatic units : 



45-1 

 36-1 

 329 



31-2 

 357 

 300 



96-6 



100-9 



99-1 



Mean 



38-0 



323 



98-9 



s 



(b) Variation of the effective capacity of the quadrant 

 with the potential of the needle. 



In February 1898 Prof. Clifton at the Clarendon Laboratory, 

 Oxford, first noticed that the effective capacity of the quad- 

 rants of an electrometer depended very largely on the charge 

 given to its needle. By increasing the charge on the needle of 

 his modification of the Kelvin type of instrument, from a value 

 which gave moderate sensitiveness to one at which the limit ol 

 stability was almost reached, Clifton found that the effective 

 capacity of the quadrant system was increased practically five- 

 fold. The capacity of the parallel-plate air-condenser, which 

 was kept constant during the measurements, and the potential 

 given to the needle, were not measured, but a summary of his 

 results shows the effect very clearly: 



Date, 1898. 



February 2nd 



2nd 



4th 

 5th 



* * 



9 f 



Capacity of quadrants 

 (Air condenser = !)• 



51 

 2-54 



1 -62 



1-30 



Owing no doubt to the very high values of the charge on the 

 needle, the sensibility was not proportional to its potential, 

 and the true connection of the latter with the variation of the 



capacity was not apparent. 



The measurements summarized in the followin 



Table 



2nd 



