578 On a Source of Error in the Ballistic Galvanometer. 

 the whole of this is dissipated in its coils. Hence 



f 



Jo 



i 2 dt=~ (hi 2 + EQ), 



which gives 



50«Gfl 



(LP + EQ) (4 a) 



for the percentage error. 



In an actual experiment carried out on these lines it was so 

 arranged that the second term in the bracket was small com- 

 pared with the first. The condenser C had a capacity of 

 one-third of a microfarad ; the circuit KLB had a self- 



■02 -03 



I in Amperes. 



oe 



inductance of one-third of a henry ; the cell had an electro- 

 motive force of two volts ; and the resistance of the galvano- 

 meter was 6600 ohms, while its coil constant was 3300 C.G.S. 



