the Electric Discharge in highly Rarefied Gaseous Media. 519 



shaped like a large comma, escapes from it. When the soft 

 iron is magnetized, this comma is distinctly seen to revolve, to- 

 gether with the halo from which it proceeds, in one direction or 

 the other, according to the direction of the magnetization. At 

 the same time the violet sheath which surrounds the ring is seen 

 to revolve in the same direction as the rose-coloured halo, 

 although they are separated by a perfectly dark space. On 

 changing the direction of the discharge, a violet envelope appears 

 at the negative electrode, but does not completely cover the 

 summit of the soft-iron rod unless the gas is very much rare- 

 tied ; while at the positive electrode brilliant points may be seen, 

 separated from each other by a rose-coloured glow which sur- 

 rounds the whole of the ring, and from which there proceed a 

 few regular stratifications concentric internally with the ring. 

 When the gas is not very much rarefied, a luminous jet appears 

 starting from the ring and extending to the top of the central 

 rod of soft iron (from which it is separated only by a narrow 

 dark space), and exhibiting a movement of rotation one way or 

 the other, like the hand of a watch, according to the direction 

 of the magnetization. In this case only part of the top of the 

 soft-iron rod is covered by the violet sheath, and this luminous 

 segment turns with the brilliant jet. 



I made a very great many experiments, under the conditions 

 I have just described, with atmospheric air, nitrogen, and hy- 

 drogen, both dry and more or less charged with vapour. I will 

 now give a summary description of these experiments, remark- 

 ing here, in the first place, that whatever was the nature of the 

 gas and its degree of elasticity, and whether it was dry or im- 

 pregnated with vapour, the velocity of rotation was always much 

 greater when the ring was made the positive electrode than when 

 it was negative, and that this rotation, which became more and 

 more rapid as the tension diminished, ceased to be appreciable 

 under a much lower tension in the latter case than in the 

 former. 



In my first experiments I used a large globe 25 centims. in 

 diameter, the diameter of the ring being 20 centims., and that 

 of the central rod of soft iron 3 centims. This globe had two 

 tubulures; one of them served for the introduction of the soft- 

 iron rod, the top of which reached to the centre of the globe, 

 while the lower end projected from the tubulure, so that it could 

 be placed on the polar surface of the electromagnet; the other 

 tubulure was closed by a stopcock through which the gas and 

 vapour could be introduced, and which carried an insulated con- 

 ductor that supported the ring and allowed of its being put into 

 the electric circuit. The discharge passed accordingly between 

 the summit of the soft-iron rod and the metallic ring. 



2M2 



