however, a glow radiating' in the direction of tlie lines of magnetic, 

 force was produced, which lasted as long as the magnet was on, and 

 died away rapidly, but not instantaneously, when the magnet was 

 taken oft". In this case the discharge seems to be nuicli easier along 

 the lines of magnetic force. 



The following experiment shows that this <'ftect is not (routined to 

 the glow discharge, but is also o[»crative when the discharge passes 

 entirely through the gas, A scpiare tube ABCD (Fig. <») is placed out- 

 side the luinuiry EFdII, the lower part of the discharge tube CD being 

 situated between the ])oles L M of an electro-magnet. l>y altering the 

 length of spark of the Winishurst macliine, the electro-motive intensity 



Fig. 6. 



acting on flic secondary can be so adjusted that no <}ischarge i^asses 

 round the tube ABCD when the magnet is off", whilst a blight dis- 

 charge occurs as long as the magnet is on. The two effects of the 

 magnet on the discharge, viz, tlu^ stopi>age of the discharge across the 

 lines of magnetic force, and its acceleration along them, maybe prettily 

 illustrated b.\' i)lacing in this exjteriment an exhausted bulb X inside 

 the i)rimary; then the spark length can be adjusted so that when the 

 magnet is off the discharge passes in the bulb, and not in the scpiare 

 tube, while when the magnet is on the discharge passes in the S(piare 

 tube, and not in the bulb. 



The experiments on the effect of the magnetic lield on the discharge 

 were tried with air, carbonic acid, and oxygen, but 1 could not detect 

 any differen<;e in the behavioi of the gas«^s. 



The explanation of the longitudinal effect of magnetic force seems 

 more ol)scure than that of the transverse effect: it is ])ossible how- 

 ever that l)otli may be due to the sanu^. cause, for if the feeble dis- 

 charge which we suppose precedes the main discharg(^ branches away 

 at all from the line ol" main discharge, the action of the magnetic force 

 when it is al(»ng the discliarge will tend to bring these branches into 

 the main line of discharge; and tlius there will be a greater sujjply of 

 dissociated molecules along tlie main line of discharge, and therefore 

 an easier i)ath for the sidjsecpient disclnnges when the magnetic force 

 is acting than \y\nm it is absent, 



It is perhaps not necessary to assume tiiat the mechanical acticui of 



