528 



THE POPULAR SCIENCE MONTHLY 



them, A is exhausted until the vacuum is low enough to give the cathode 

 rays; the discharge is then sent through A, and the cathode rays bom- 

 bard the solid. The result of this is that in a few seconds so much gas, 

 mainly C0 2 and hydrogen, is driven out of it that the pressure gets too 

 high for the cathode rays to be formed, and unless some precautions to 

 lower the pressure were taken the bombardment would stop. To avoid 

 this, a tube containing charcoal cooled by liquid air is connected with 

 A, and this absorbs the C0 2 and enough of the hydrogen to keep the 

 vacuum in the cathode ray state. To see what new gases are given off 

 in consequence of the bombardment, a photograph is taken while the 

 connection between A and B is cut off. After this is finished, and when 

 the bombardment has gone on for about four hours, the tap is turned and 

 a little of the gas from A is allowed to go into B ; another photograph is 

 taken, and those lines in the second photograph which are not in the 

 first represent those gases which are liberated by the bombardment, and 

 which have escaped being absorbed by the charcoal. I have here a slide 

 (Fig. 5) representing the result of bombarding nickel. There are two 



Pig. 5. 



photographs, one (a) before turning the tap and the other (/?) after; 

 in the second you see the three line very distinctly, while it is absent 

 from the first, showing that the gas giving the three line has been lib- 

 erated by the bombardment. I have got similar results to these when, 

 instead of nickel, iron, copper, lead, zinc have been bombarded. I have 

 tried two specimens of meteorites kindly lent to me from the Minera- 

 logical Museum, Cambridge, and found there the three line. Nearly 

 every substance I have tried gives, the first time it is bombarded, the 

 helium line as well as this line due to X 3 ; if, however, the same sub- 

 stance is bombarded a second time, the helium line is in general absent 

 (occasionally it is still to be detected, though exceedingly faint) ; and 



