TRANSIENT CURRENTS. 329 



Suppose first that the intensity I is constant in the principal 

 circuit, and let / , /, and i' be the three intensities observed in 

 the galvanometer G, we have 



whence 



r 



(15) 



r t n -t t 



The conditions are the same as in the first case (931). 



If the intensity i Q is kept constant in the branch circuit, the N/ 

 intensities I , I, I' of the principal current give the ratios 



;_! a =I * =I > a _ 



Q g + a g+r + a g+r' + a' 

 from which is deduced 



(is)' ?'f^- 



This latter method, which was described by Bosscha,* can only 

 give exact results provided the differences I 1 and I' 1 are 

 large enough that is to say, if the resistances are of the same 

 order as that of the galvanometer. 



When the resistance of the galvanometer, which serves for the 

 readings, is very large, it may be placed as a shunt on a constant 

 part of the branch circuit, and the formulae do not alter. 



936. TRANSIENT CURRENTS. An instantaneous current, such 0~v 

 as those produced in induction, may also be used provided the 

 duration of the current is very small compared with that of the 

 oscillations of the galvanometer needle ; in the methods based on 

 the comparison of two currents it is sufficient to replace the de- 

 flections of the needle by the angle of swing. W. Weber, f for 

 instance, displaced a magnet between fixed limits in the interior of 

 a coil. The resistances to be compared, r and r, will be introduced 

 in the circuit as before (931) ; but the experiment may also be 

 arranged differently. The circuit is first closed by a galvanometer 

 of resistance y, and then the galvanometer is shunted by the re- 

 sistances r and r' in succession, then by the same resistance in 



* BOSSCHA. Pogg. Ann., Vol. ex., p. 452. 1860. 



t W. WEBER. Electrodyn. Maasbestim., p. 209. 1863. 



