1891.] Luminous Discharge of Electricity through a Gas. 85 



the line of sight is in the direction of the discharge tnbe. It follows 

 from this, by Doppler's principle, that the particles when emitting 

 light are not travelling in the direction of the discharge at the rate of 

 more than a mile a second, proving at any rate that the luminosity 

 does not consist of a wind of luminous particles travelling with the 

 velocity of the discharge. 



Wheatstone's observations only give an inferior limit to the velocity 

 of the discharge. Nothing was observed in these which indicated 

 that the velocity of discharge was finite. A method which would 

 enable us to measure this velocity would also at the same time show 

 whether the discharge always started from the positive or negative 

 end of the tube, and so enable us to trace the course of the dis- 

 charge. 



In the following experiments I have endeavoured to measure this 

 velocity, and also to ascertain whether the main discharge starts from 

 the anode or the cathode. The long tubes used in my experiments 

 were practically filled by the positive column ; thus in the tube 50 feet 

 long the positive column extends to within an inch or two of the 

 cathode. All the experiments described below relate to the behaviour 

 of the positive column. 



Pliicker (Poggendorff's ' Annalen,' vol. 107, 1859, p. 89) concludes 

 from the action of a magnet on the discharge that it starts from the 

 anode. This conclusion does not seem to have met with much accept- 

 ance ; my experiments, however, fully bear it out, as I find that, except 

 under exceptional circumstances, which will be described later, the 

 luminosity of the positive column begins close to the anode and 

 travels away from it. 



The experiments for measuring the velocity of the luminous column 

 were after several preliminary trials finally arranged in the following 

 way. ABCDEFG . . . . L (fig. 1) is a glass tube about 15 metres long 

 and 5 mm. in diameter, which, with the exception of two horizontal 

 pieces of BC and GH, is covered with lamp black; this tube is ex- 

 hausted, and a current sent through it from a coil giving sparks 6 or 

 7 inches long in air ; the light from the uncovered portions falls on a 

 rotating mirror MN, placed at a distance of about 6 metres from BC ; 

 the light from GH falls on the rotating mirror directly, that from BC 

 after reflection from the plain mirror P. The images of the bright 

 portions of the tnbe after reflection from the revolving mirror are 

 viewed through a telescope, and the mirrors are so arranged that when 

 the revolving mirror is stationary the images of the bright por- 

 tions of the tube appear as portions of the same horizontal straight 

 line. The terminals of the long vacuum tube are pushed through 

 mercury up the vertical tubes AB, KL. This arrangement was adopted 

 because by running sulphuric acid up these tubes the terminals could 

 readily be changed from pointed platinum wires to flat liquid sur- 



