through Rarefied Gases. 411 



were so brought in equatorial position between the two approxi- 

 mated armatures that the bulb containing the electrode chosen 

 as the electro-negative one touched them from the outside (to 

 take one particular case), the electric cuiTent being then led 

 through the tube, and the magnetic light becoming concen- 

 trated into a sickle-shaped disc. 



53. The colour of the magnetic light appears to depend upon 

 the nature of the gas in the tube. This light is in most cases 

 violet, of a more or less red tinge ; it is often accompanied by an 

 intense green light (35); frequently it is of a beautiful blue; 

 and under otherwise unfavourable circumstances I have obsei'ved 

 an intense yellow flame of magnetic light. 



Hitherto we have assumed the use of platinum electrodes. In 

 the phsenomena under consideration, the nature of the metal of 

 the electrode appears nevertheless to exert no influence. I have 

 already mentioned the apparently perfect similarity in the beha- 

 viour of strongly coppered, silvered, and gilded platinum elec- 

 trodes (37). A sewing-needle as negative electrode gave the 

 same appearance, as also electrodes of lead and tin. (The 

 portion of the electrode fused into the glass can of course be 

 nothing but platinum. The metal under investigation was fast- 

 ened appropriately to the platinum.) 



54. From these experiments we are, it seems, justified in con- 

 cluding that the diamagnetic or paramagnetic nature of the par- 

 ticle torn from the electrode need not be taken into account in 

 considering the magnetic light. If, on the other hand, we were 

 to assume that these particles, retaining the electricity of the 

 electrode, gave rise by their motion to elementary currents, and 

 further assuming their motion to be a spiral one, we should cer- 

 tainly have a sort of explanation for the pha^nomena observed. 

 Such an explanation, however, is not as yet sufficiently jus- 

 tifiable. 



55. Magnetic light under the influence of magnetic excite- 

 ment gives no trace oi polarization. 



56. In consequence of our want of knowledge about magnetic 

 light, and of the total want of analogous phsenomena, I per- 

 formed many experiments in order to obtain magnetic light under 

 other circumstances. All such attempts, however, were fruit- 

 less ; it may suffice to mention one of them here. 



I directed my attention especially to that remarkable fluores- 

 cent light, for the more exact knowledge of which we arc in- 

 debted to Mr. Stokes. In order to produce this light I em- 

 ployed a narrow Geissler tube, which showed an intense blue 

 liglit on tlie passage of the current ; this was fastened to the 

 middU; of a wider tube containing an aqueous solution of a;sculine. 

 On exciting the curicnt, the fluorescent light in the water formed 



