OF ARTS AND SCIENCES. 335 



The minimum positive velocity has, of course, the greatest signifi- 

 cance, being most free from the effects of direct action ; and we may 

 finally conclude that the " principal velocity " of propagation of the 

 magnetic wave over a half-inch rod, unaffected by foreign influence, is 

 in the neighborhood of twenty-five feet per second. The " principal 

 velocity " of the electric pulse over the French Atlantic cable is about 

 13,700,000 feet per second. 



Besides the experiments already alluded to, there was a preliminary 

 series made with a Siemens armature machine, — the armature being 

 converted into a powerful electro-magnet, and caused to revolve by 

 means of a band and pulley-wheel. A series of magnetic waves was 

 induced in the fixed horseshoe between whose poles the armature re- 

 volved ; and, on screwing an iron rod firmly into one of its extremities, 

 the waves spread more or less over the bar. A small induction coii 

 was placed around the bar at a distance of sixteen inches, and con- 

 nected with the galvanometer through a commutator which acted at 

 the moment when the armature reached the position of greatest attrac- 

 tion. It is evident that, when the machine was set revolving very 

 slowly, the impulses were all added together. It was found, however, 

 as expected, that, as the velocity increased, the deflection of the gal- 

 vanometer diminished, and was finally completely reversed, showing 

 that a retardation of more than 90° of phase must have taken place. 

 The least velocity which could cause a reversal ^was usually about 

 fifteen revolutions per second, — although the number varied greatly 

 at different times. It is interesting to observe that this velocity and 

 distance correspond to a retardation of phase, in our table, of 108°. 



It was thought that by this method the uncertainty of the time of 

 magnetization of a primary coil would be got rid of, since the exact 

 time of each impulse could be known ; but the imperfection of the 

 apparatus and discordance of the results caused this method to be 

 abandoned. 



The phenomena of the propagation of the magnetic wave in a bar 



