90 Prof. J. J. Thomson. On the Propagation of the [Jan. 15, 



When the mirror was running at full speed, its rate was pretty 

 constant and, as determined by the note given out in a telephone 

 (taken for the sake of avoiding the noise made by the Gramme and 

 mirror to an adjoining room) when the circuit was broken once in 

 each revolution of the mirror, nnd also by the velocity of the band 

 driving the mirror, was about 480 per second. The distance between 

 the places BC and GU is 7 metres, so that if v is the rate at which 

 the luminosity of the positive column travels, 



hence v = 1 2000 x 2*- x 480 x 700 x 



I'o 



= 1-6 x 10 10 , 



or rather more than half the velocity of light ; but, as I explained 

 before, this must be regarded as an approximation, rather than as an 

 accurate determination. It is sufficient, however, to show that the 

 luminosity of the positive column travels through a vacuum tube 

 with a velocity comparable to that of light. 



The preceding results hold when there is a short air break in the 

 circuit, but, if the air break is increased until the coil can only spark 

 through the tube with difficulty, the luminosity seems inclined to 

 start from the air-break electrode, and the direction in which it 

 travels is not always reversed by reversing the coil. 



The fact that the main portion of the luminous discharge in a long 

 vacuum tube has its origin at the positive electrode may appear at 

 first sight inconsistent with the result that glow discharge takes 

 place more easily, that is, with a smaller value of the electromotive 

 intensity, at the negative than at the positive electrode. Thus 

 Faraday states that the discharge from a sphere takes place more 

 easily when the sphere is negatively than when it is positively elec- 

 trified. 



Again, ultra-violet light can produce a discharge from a nega- 

 tively but not from a positively electrified piece of metal. Thus 

 Lenard and Wolf (Wiedemann's ' Annalen,' vol. 37, 1889, p. 443) 

 have proved that we can produce a cathode by allowing ultra-violet 

 light to fall on a negatively electrified plate, while no discharge 

 occurs if the plate is positively electrified, and Hallwachs (Wiede- 

 mann's 'Annalen,' vol. 34, 1888, p. 731) and Righi have shown that 

 when ultra-violet light falls on an unelectrified piece of metal the 

 metal becomes positively charged, i.e., the light converts it into a 

 cathode. 



These considerations do not, however, seem to affect the question 

 we are considering when the electromotive force is sufficiently great 



