DowiJNG — Sensitive Valve Method J or Measuretnent of Capacity. 179 



graduated from 100 to 1,200 micro-microfarads, and in one experiment B 

 was made 200,000 ohms. 'I'he relation between the galvanometer deflection 

 and capacity was almost linear, as shown in the curve (fig. 3), the condenser 

 graduations being unreliable for the smaller values. 



A very sensitive moving-coil galvanometer^ was then substituted for the 

 Unipivot instrument, the universal shunt being retained, and the connexions 

 altered to those shown in fig. 4. This had the advantage of being more 

 compact, as only one storage battery was required. Preliminary observations 

 were made, as before, with the graduated air-condenser, and the expected 

 high sensitivity of the apparatus was fully obtained. For example, with a 



Fio. 4. 



capacity of 100 micro-microfarads (lO"'" farad) an alteration of 1 per cent, in 

 the capacity produced a deflection of 300 scale-divisions, the galvanometer 

 shunt being at 10.* With this degree of sensitiveness the deflections were 

 fairly steady, but, with a lower value of the shunt, the spot of light commenced 

 to execute irregular movements, rendering accurate readings impossible. 



1 Elliott's "Century" galvanometer: 1,000-ohni coil; 2-5 x lO'" amp. per scale- 

 division. 



■ In view of the fact that the resistance of the plate-circuit is very high, and that the 

 galvanometer and shunt were in circuit with 10,000 ohms in the potentiometer-circuit, 

 the actual shunt values were not very different from those marked on the shunt itself. 



