THOMSON'S METHOD OF CHEMICAL ANALYSIS 53 



the top of a light-tight metal case and as soon as the fluorescent 

 screen at A shows that the desired conditions have been 

 obtained the plate is lowered into the field of the rays and 

 a photograph taken. The exposures depend almost entirely on 

 the diameter of the canal-ray tube and vary from three minutes 

 to three hours. By the use of a long plate, as many as three 

 photos could be taken before it was necessary to destroy the 

 vacuum in the apparatus and introduce another plate. As it 

 is usually desirable, for reasons which will be explained, to 



w 



Fig. 2. 



have as low a pressure as possible in the " camera," one or 

 two Dewar charcoal tubes are attached to it and are immersed 

 in liquid air while the photograph is being taken. As gas 

 can only enter through the long and fine canal-ray tube the 

 pressure in the camera may be very much lower than that in 

 the bulb. 



The illustration facing p. 48, which is from a flashlight 

 photograph taken by Mr. Hayles, of the Cavendish Laboratory, 

 conveys a good idea of the actual appearance of an apparatus 

 set up by the writer with which a great many results were 

 obtained. On the extreme right can be seen part of the gas 



