266 MEASUREMENT OF CURRENTS. 



The following experiment of Mr. Latimer Clark* illustrates very 

 clearly the influence of the motion of the needle. A condenser is 

 discharged through two identical galvanometers shunted in respect 

 of each other on the same circuit ; both give the same deflection, 

 which is half that observed in each of them when it alone receives 

 the discharge. If the experiment is again made after fixing one 

 of the needles, the deflection of the free needle is much less. In 

 the first case the two needles experience the same swing and 

 expend the same work ; in the second case there is electromagnetic 

 action in only one of the circuits, and only for this alone is there an 

 apparent increase of the resistance (535). The current, therefore, 

 does not divide equally between the two branches. 



With differential galvanometers where the needle is stationary, 

 shunts can always be used, even for instantaneous currents. 



885. The ballistic method involves two kinds of errors! which 

 it seems difficult to eliminate completely, and which in certain cases 

 may have a considerable influence. 



In the first case the current which traverses the galvanometer 

 acts during a very short time, but with great energy. This action, 

 which is perpendicular to the axis of the needle, is, it may be, of a 

 nature to produce a temporary change in its magnetic moment so 

 great that the swing is no longer equal to that which would 

 correspond to the original moment. 



In the second place, we apply to the angle of throw the cor- 

 rection for the logarithmic decrement deduced from observations 

 of continuous oscillations. This method would be perfectly correct 

 if the damping were due merely to currents induced by the motion 

 of the needle ; but when a considerable part of the effect is due to 

 resistance of the air, we may inquire whether this resistance acts in 

 the same way on a movable body oscillating regularly as it does in 

 this same movable body originally at rest, and which moves under 

 the influence of a shock, and particularly if it is near its position 

 of equilibrium. The resistance of the air should be greater in the 

 second case than in the first, hence the correction as made would be 

 inadequate. It is true that the effect of this same cause would be 

 to increase the time of oscillation, and therefore to introduce a new 

 error in the contrary direction to the first, and which would partially 

 destroy it. 



* L. CLARK. Journal of 'the Society of 'Telegraph Engineers ' f Vol. n.,p. 16. 1873. 

 t L. RAYLEIGH. Experiments to Determine the Value of Unit of Resistance. 

 Phil. Trans., 1882, Partii., p. 619. 



