Magnetism on Gases traversed bij Electric Discharges. 213 



In order to study this special point more closely, we made 

 use of a large glass tube with two compartments divided the one 

 from the other by a stopcock, also of glass, with a large opening. 

 The length of this tube is 51 centimetres, its diameter 65 milli- 

 metres, the length of each of the two compartments 225 millims.; 

 they are joined together by a tubulure 60 millims. long., 10 mil- 

 lims. in diameter; and the opening of the glass stopcock has the 

 same section as the latter. The stopcock, working with perfect 

 smoothness, closes hermetically. The tube carries at its two 

 extremities brass mountings, with stopcocks and knobbed elec- 

 trodes, also of brass. The distance from one electrode to the 

 other is 41 centims. The tube is placed transversely between 

 the two poles of the electromagnet in such a manner that the 

 centre of one of the compartments is on the axis of the magnet, 

 while the other is completely withdrawn from the action of the 

 magnetism. The electromagnet used in these experiments was 

 that described by M. de la Rive in his memoir on the magnetic 

 rotatory polarization of liquids*. We usually magnetized it 

 with the current produced by 40 Bunsen couples. In order that 

 the magnetic action might be as intense as possible on one of 

 the two moieties of the electric discharge, the two poles of the 

 electromagnet were brought into immediate contact with the 

 sides of that compartment of the glass tube which was to be 

 submitted to its action. This compartment was put in commu- 

 nication, by means of a small leaden tube, with a very sensitive 

 manometer, consisting of a double barometer and a cathetometer 

 which permitted the estimation of hundredths of a millimetre. 

 The other compartment was connected by a system of leaden 

 tubes with an ordinary air-pump, a mercurial air-pump, and an 

 apparatus for introducing and desiccating the gases. The in- 

 duced current was furnished by a Ruhmkorff coil of medium 

 size excited by the current of a Grove pile of four pairs. 



In making the experiment, .we commenced by producing a 

 vacuum throughout the apparatus, then introduced the gas which 

 was to be operated on, again exhausted the apparatus, reintro- 

 duced the gas, repeating the process several times, till the con- 

 tained gas was sufficiently pure ; the gas in the tube was then 

 brought to a well-determined pressure, which could be easily 

 done, within 5 or 10 hundredths of a millimetre, either with the 

 mercurial pump, or with the ordinary air-pump, by suitably re- 

 gulating the working of the stopcocks. 



Communication with the lead pipes being closed, the level of 

 the mercury is taken very exactly; then the cock which regu- 

 lates the communication between the glass tube and the mano- 



* Archives des Sci. Phys. et Nat. vol. xxxviii. p. 209 ; Phil. Mag. S. 4. 

 vol. xl. p. 394. 



