214 Lord Rayleigh. On the Bridge Method in its [Feb. iy, 



To take a numerical example, let 6 3 = ; and suppose a t = ^ b t . 

 Then, according to (33), a, = T ^ 5,. Also by (20), (21), 



The corresponding minimum value of (32), equal to (34), is 



But with this value of a t the gain by allowing a- t to be finite is 

 great. If a, = 0, 



and the value of (32), equal to (36), is | 6,. 



We see from (36) that when a, = there is little to be gained by 

 further reduction of a\. But when o-j is suitably chosen the gaii 

 may be worth having. Thus, in (34), if a\ = y^ 61, wo have yV&i* 

 Corresponding to this a, = + y 5 6 t nearly, and 



These are not unreasonable proportions, and we see that the use 

 Oj may be advantageous, even when the subject of measurement is 

 mere resistance. It will be remarked too that, except as regai 

 62, /], the sign of <t~ is immaterial. 



When the branches b, d consist of electromagnets, and still moi 

 when they consist of condensers, &! may be very small. If we sa| 

 pose it to be zero, (30) becomes 



From (37) we see that the increase of Oj is favourable, especially 

 if the sign be the same as of 6 a . Even if a? = 0, (37) now assuming 

 the form 



)TT- (4) 



