28 Mr. R. H. M. Bosanquet on a 



of the resistance, a process capable of being executed with a 

 very high degree of accuracy. 



The channel of the circle in the instrument I have con- 

 structed is exactly 1 metre in circumference, as measured by 

 a thin wire lying close to the floor of the channel. In the 

 channel 100 turns* of a very fine covered wire were wound, 

 which was made to have a resistance of 10 ohms per metre (it 

 is really somewhat more). The wire forms one layer and 

 about half another. The thickness of the layer is very small, 

 and it is assumed that the average length of each turn is 

 exactly 1 metre. The resistance of the whole was compared 

 with an old set of B.A. standards, and found to be 1175 B.A. 

 units. For the present I use this number ; for purposes of 

 great accuracy it would be reduced to ohms. 



The ordinary formula for the tangent-galvanometer gives 



for the current C, 



TIP 

 C = ^tan0(inC.G.S.) (i.) 



Let V be the number of volts between the terminals, 

 the number of ohms, 

 then 



B"~O.10 9 ~10.O' ' ' ' ' K } 

 substituting in (i.), 



HP 

 V = lO.O.=£-tan0 (iii.) 



In the present case, putting 



U 



2tt 



H=-181, k=~, L = 10,000, = 1175, 



we find 



log V= 1-73136 + log tan 0; 



whence a table can be constructed giving the number of volts 

 corresponding to a given deflection. 



For small deflections the number of degrees is somewhat 

 greater than the number of volts. 



Applying this standard galvanometer to determine the 

 E.M.F. of two bichromate cells, whose resistance can be prac- 

 tically neglected, I found from 2*36 to 2*0 volts per cell, 

 according to the condition of the cells. 



The behaviour of a trough of nine Daniell's cells, when ex- 



* The turns were counted three times — once in winding on, once in 

 winding off, and once in winding on again. If any discrepancy occurred 

 the process was begun again. 



