88 Prof. J. J. Thomson. On tli?. Propagation of the [Jan. 15, 



tween the break of the coil and the .position of the mirror was 

 attempted by artificial means. 



When the observations were made in this way, the observer at the 

 telescope saw, on an average about once iu four minutes, sharp bright 

 images of the portions BC and Gil of the tube, not sensibly 

 broadened, but no longer quite in the same straight line ; the relative 

 displacement of these lines was reversed on reversing the coil, and also 

 on reversing the direction of rotation of the mirror. These bright 

 images are not the only ones observed through the telescope; ill- 

 defined and widened images were much more frequent; sometimes 

 these were widened out so as to fill the whole field of view with a 

 luminous haze, at others, the images appeared as broad bands, the 

 boundaries of these bands not being in the same straight line ; these 

 images indicate a discharge lasting for a very much longer time than 

 that which produced the bright sharp images which were the object 

 of our attention. When these sharp images were very bright, it 

 could be distinctly seen that they were striated. 



The displacement of the images from the same straight line is due 

 to the finite velocity with which the luminosity is propagated ; for if 

 the mirror can turn through an appreciable angle while the luminosity 

 ti'avels from BC to GH, or vice versa, these images of BC and GH, as 

 seen in the telescope after reflection from the revolving mirror, will 

 no longer be in the same straight line, but if the mirror is turning so 

 that, on looking through the telescope, the images seem to come in at 

 the top and go out at the botton of the field of view, the image of 

 that part of the tube at which the luminosity appears first will be 

 raised above that of the other. If we know the rate of rotation of the 

 mirror, the vertical displacement of the images and the distance 

 between BC and GH, the rate of propagation of the luminosity may be 

 calculated. The observations were made in the following way : The 

 tube having previously been properly dried and exhausted, sc that the 

 discharge would pass freely through it, one observer took his seat at the 

 telescope, the room was then darkened, and the coil set in action by 

 another observer, the observer at the telescope not knowing which of 

 the electrodes was positive and which negative ; the mirror was then 

 set in rotation, and about once in four minutes, on an average, the 

 observer at the telescope saw the bright sharp images alluded to 

 above. When the observer had seen two or three of these, he stated 

 whether they were very bright, fairly bright, or indistinct; which 

 image was the higher, and by how much. The distance between the 

 images was estimated in terms of the apparent distance between the 

 divisions of a vertical millimetre scale placed at GH when seen 

 through the telescope ; the scale was not seen at the same time as the 

 image of the tube, so that the observation cannot claim any great 

 accuracy ; different observers agree, however, under the same circnm- 



