568 Mr. M. Solomon on the 



1?? =1-096, ^=1-099, —=1-088. ~ = 1096, — =1096, 

 164 149 137 125 114 



111=1005. 

 104 



Square root of mean of above values = 1*047. 



Average decrement for observations (a) = * / — =1 



0465. 



22 /m 



(*>=a/™=1W5. 



Mean decrement, #=1-047. 

 \=log e 5=00159. 

 Eatio, £/\=79-5. 



Altogether six results were obtained for the galvanometer 

 on open circuit, and a corresponding set for closed circuit. 

 The experiments were made by first adjusting the controlling 

 field until a suitable period was obtained ; this was done by 

 moving the controlling magnet, wbich rested on the bem-h on 

 which the galvanometer stood, up to or away from the in- 

 strument. The period and decrement were then observed 

 with the galvanometer on open circuit ; the terminals of the 

 galvanometer were then short-circuited, and a fresh obser- 

 vation of the period and decrement made. The initial deflexion 

 was in all cases given to the needle by bringing a weak 

 magnet into its neighbourhood, and the needle was started 

 swinging by the sudden removal of this magnet. The results 

 obtained are tabulated below (Table IV.) and plotted as a 

 curve in fig. 4. 



These results show that the ratio t/X is constant for all the 

 periods observed — from 2*5 to 16*5 seconds — when the galva- 

 nometer coils are on open circuit. When, however, the 

 terminals are short-circuited, we see, in the first place, that 

 the damping is increased, showing that eddy-current damping 

 has been introduced, and secondly that the ratio t/X is no longer 

 constant, but increases with the period, i. e. with a decrease in 

 the strength of the controlling field, becoming, however, con- 

 stant at the higher periods. It is evident, therefore, that the 

 alteration in the ratio is due to the eddy-current damping: and, 

 moreover, of the three ways in which it has been pointed out 

 the eddy-current damping may be affected, only the first is 

 consistent with a decrease in the value of t/X as the period 

 decreases ; for both the self-induction of the galvanometer- 

 coils and change in their resistance due to a rise in tempe- 

 rature caused by the induced currents would cause the ratio to 

 increase as the period decreased. We must therefore conclude 

 that the change in the ratio t/X is due to an alteration in the 



