the Electric Discharge in highly Rarefied Gaseous Media. 523 



several jets of much greater width than what are seen with 

 aqueous vapour. Nevertheless, if the diameter of the ring is 

 too great (more than 15 centims., for example), the subdivision of 

 the jet cannot be obtained except with difficulty, unless the in- 

 tensity of the discharge and that of the magnetization are very 

 considerable. 



In the following experiment the rarefied gas was hydrogen, 

 and several successive quantities of alcohol-vapour were intro- 

 duced into it. The pressure of the pure and dry gas at the be- 

 ginning was 5 millims. At this pressure, as we shall see imme- 

 diately, hydrogen transmits the discharge solely under the form 

 of a luminous sheet. The pressure was then augmented entirely 

 by the alcohol-vapour, and the following results were obtained : — 



Pressure. 



Number of turns in one minute. 



Ring positive. 



Ring negative. 



niilliras. 

 7 

 10 

 12 

 15 

 18 

 22 

 27 

 36 

 38 



Luminous sheet. 

 80 

 64 

 48 

 40 

 30 

 24 

 12 

 12 



92 

 52 

 48 

 38 

 32 

 25 

 18 

 10 

 10 



The division into a greater or less number of separate jet 

 took place when the ring was made the positive electrode. 



When pure and dry hydrogen is the medium through which 

 the discharges are passed, it is very difficult to obtain the phe- 

 nomena of rotation. At comparatively high pressures, such as 

 128 millims., jets are certainly obtained, but they are not suffi- 

 ciently continuous for a magnet to be capable of affecting them. 

 At 90 millims. I obtained a small jet in the form of a bluish- 

 white thread, which, when the ring was positive, revolved at the 

 rate of 35 turns in a minute ; but it soon broke up into a mul- 

 titude of little irregular jets, and the rotation was then no longer 

 perceptible. As far as 40 millims. the action of the magnet was 

 but slightly marked ; at 30 millims. the negative ring became 

 covered with small violet sheaths, separated by uniform intervals, 

 which, at the instant of making the magnet, seemed to exhibit a 

 tendency to motion in one direction or the other, according to 

 the direction of the magnetization. The same thing happens 

 in the case of the small, uniformly but closely distributed bril- 

 liant points wherewith the ring becomes covered when it is 

 positive. At 5 millims., and still more at 3 and 2, the ring 



