138 MU. C. E. S. PHILLIPS OX THE ACTION OF MAGNETISED ELECTEODES 
shown in fig. I, the large mass of metal exposed within the hulb, made it difficult to 
ensure that the gases experimented with Avere in a pure state. The iron electrodes 
continued always to give off a little contaminating gas at very low pressures, eA'en after 
a Aveek’s pumping. When the characteristic spectrum of the gas used was A'isihle 
AA^hile the discharge passed freely (the IuiIId having been evacuated and filled Avith the 
gas a.lternately six times), ol)serA"ations on the appearance of the ring at still loAver 
])ressures Avere begun. In the case of oxygen obtained from steel cylinders, and also 
from potassic chlorate, the ring appeared to he someAAdiat Avliiter than in air. With 
hydrogen and carljon dioxide, hoAvever, no change of any kind Avas noticeable. 
Mercury Ampour AA^a.s present in every case. As the object of the iiiA^estigation ATas 
to determine the process l)y AAdiich the luminous ring formed, it Avas not thought 
advisable to devote too inucli time to more detailed ex2)eriments upon tlie gases. 
The Spectrum. 
It Avas considered sufficient to ascertain AAdiether the composition of the light 
einltted l)y the ring AA^as similar to that of the faint Ijlue haze Ausible in tlie bulb 
Avhile the induction-coil discharge AA'as passing. This Avas found to l)e the case OA'er a 
range from al)Out \4365 to X5I5I. Great difficulty, hoAA'ever, AArns found in distin¬ 
guishing the lines OAving to their feeble intensity, nor AA'as it possible to obtain 
satisfactory photographs of the ring itself. 
The Ring in Rotation. 
The AA'ant of uniformity in the density of the gloAAung nebulous stream constituting 
the ring enalded its rotation to be plainly made out. The folloAving experiment, 
lioAveAmr, })laced tlie fact beyond a doubt. 
A bulb, similar to that shoAvn in fig. 3, AS'as exhausted and a discharge passed 
Fig. 3. 
through it. It Avill be seen that a hnninous ring forming under the Influence of the 
magnets Avould, in this case, i)artially envelop the AA'lngs of the light mica A'ane, Y, 
so that, if the ring reAxWe, the Amne might also turn, and by the direction of its 
