170 Perkins — Velocity of the Propagation of Magnetism. 



this distance 

 360 ' d'i 



"d" and the angular difference "S," we have 



-, where n~ the frequency, and in this case was 



60 cycles per second. 



In comparing the curves of magnetic leakage and velocity it 

 is evident that as the flux in the bar diminishes the velocity 

 increases with a consequent lessening of the lag per unit length. 

 This is what would be expected, for with a decrease of flux the 

 eddy currents must grow smaller, and there is less opposition to 

 the advance of the magnetism. Beyond nine inches the velocity 



eoo 























70 o 



















1 



1 





6,00 























soo 























Ve/oc 

 in fee 

 secon 

 4-00 



fjper 





















300 













































ZOO 







iOO 















CUflVE 



or MAG 



SHOW/ 



NG VEL 

 \f /A/ 3 



.oc/ry 



T££L 



p. 





2. 



3 



/rrches / 



1 



neoscre 

 S 



:/ from c 



1 



BAH. 1 



'enter of exc/f//-7a ce//. 



\e \, 1 



increases with great rapidity and presumably becomes equal to 

 the velocity of light when the flux density is infinitely small, or 

 when the core is a non-conductor. To test this latter case I 

 made an observation with an air core, and with coils A and B 

 about 6 inches apart. In this case the induced e.m.f. was too 

 small to appreciably affect the electrometer, so a very sensitive 

 D'Arsonval ballistic galvanometer was used. Instead of an 

 alternating current a constant voltage was impressed on the 

 exciting coil and the throw of the galvanometer observed at 

 both " make " and " break." Now if B is so adjusted that it 

 gives exactly the same throw as A, then when they are in series 

 and opposed there should be no throw provided the resistance 

 and self-induction of both are made equal. This condition was 

 fulfilled by adding an external coil in series with A, thus mak- 



