THE PROPAGATION OF ELECTRICITY. 1^5 



on auotlicr portion of the current, provided it be not too remote from the neg- 

 ative electrode. 



The consequence of the explar.atiou just given, and which it is easy to verify 

 by experiment, is, that the portion of the gas which transmits the discharge 

 must, when it is subjected to the action of the magnet, become a more imperfect 

 conductor, and that consequently the electric current must, on the whole, en- 

 counter a greater resistance in its passage along the interior of the tube when 

 one part of the tube is approached by the electro-magnet than it encountered 

 previously. 



Thus the tube of one metre being filled with rarefied hydrogen, and the ap- 

 paratus of derivation placed in the circuit,* we obtain the following results : 



Pressure. Intensity of the derived current. 



Without magnetization. Magnetization at the Magnetization at the 



positive electrode. negative electrode. 



4™™ 33° 30° 20° 



8°^ni 30° 30° 10° 



With the tube 50 centimetres in length, filled with nitrogen, we have : 



Pressure. Intensity of the derived current. 



Without magnetization. Magnetization at the Magnetization at tfie 



positive electrode. negative electrode. 



21"" 57° 52° 42° 



4"'" 37° 27° 17° 



6"^™ 25° 20° 12° 



The effects are more marked when the tubes are placed equatorially between 

 two soft-iron armatures of the electro-magnet, Avhich are immediately in contact 

 with tlie walls of the tube, than when they are placed axially on the poles 

 themselves. We see that there is a much greater increase of resistance when the 

 magnetism acts on the portion of the current near the negative electrode than 

 when it acts on the portion near the positive electrode. The reason of this 

 difference is, that the first portion which, as we have seen in the preceding 

 paragraph, is endued with a much greater share of conductibility, must natu- 

 rally experience a more considerable diminution of that property by the con- 

 densation of the gaseous matter produced by the action of the magnet, than is 

 experienced by the second portion, where the gas is less rarefied. The direction 

 of the magnetization has no influence on the results : it has no other effect than 

 to elevate or depress the current, which, when the magnet does not act, is simply 

 horizontal. 



Among the experiments which I have made regarding the influence exerted 

 by the exterior action of magnetism on rarefied gases enclosed in tubes, I will 

 further cite those in which the tube is convoluted into a flat spiral terminated 

 by two prolongations perpendicular to the plane of the spiral which serve to in- 

 troduce and rarefy the gas, as well as to give a passage to the discharges ; the 

 tube of the spiral and its prolongations is a little less than one centimetre in 

 diameter, and its total development nearly eighty centimetres. It is necessary 

 that the gas should be rarefied at least as much as 2™™ for the discharges to 

 pass when nitrogen or atmospheric air is employed. With hydrogen, a press- 

 ure of 5 or 6™™ suffices for their transmission. But whatever the gas or its 

 degree of rarefaction, it is only after the lapse of some minutes from its being 

 placed in the circuit that the discharge begins to pass. It is evidently neces- 

 sary that it should be sometimes charged with static electricity for the resist- 

 ance to the establishment of the continuous stream to be surmounted. But that 



* It should not be forgotten that here the derived cuiTent is proportional to the principal 

 current, so that its intensity may be regarded as being quite approximately the measure of 

 that of the discharge which traverses the tube. 



