36 



much more nearly the same for all substances, and that 

 such differences as existed (e.g., carbon generally gave more 

 than lead) appeared to be proper, in view of the conditions 

 stated above. As the results we had obtained seemed to be 

 sufficient to give a criterion between the material and the 

 ether-pulse theories, we thought it right to publish the details 

 of the work as far as we had carried it. 



We have now made a more thorough investigation of the 

 nature and amount of emergence radiation. We believe that 

 we are in a position to connect together all the effects atten- 

 dant on the absorption of the 7 rays, and the consequent pro- 

 duction of i8 rays, to a first approximation at least. Included 

 in these effects are some which have come to light during this 

 work, the neutral-pair theory having led us to their discovery. 



It is not a very easy thing to obtain an exact measure of 

 the emergence radiation. When a stream of y rays is shot 

 into an ionization-chamber through a plate forming one of 

 the chamber walls, a very large proportion of the ionization 

 produced is actually due to the emergence rays of the plate. 

 But if the plate is taken away, the place of these rays is sup- 

 plied, to a varying extent, by secondary rays, made in the 

 air or emergent from the last substance traversed by the rays. 

 A normal stream of y rays always contains jS rays : if these 

 are stopped by a screen, fresh )8 rays emerge from the other 

 side of the screen. It is impossible, therefore, to measure the 

 emergence rays by subtracting the value of the current when 

 the above-mentioned plate is not in position from the value 

 when it forms part of the chamber wall. Such an operation 

 might conceivably show a negative emergence radiation. 



It is, in fact, necessary to remove the ^ rays from the 

 stream of y rays by some means which is more effective than 

 a screen. A powerful magnetic field can considerably purify 

 the y-ray stream for a short distance along the path. Owing 

 to the action of the air, ^ rays will reappear again in mea- 

 surable amount after the rays have traversed a few centi- 

 metres ; nevertheless, we have found the method to be fairly 

 satisfactory so far, and will doubtless be able to improve it 

 when better information is available for the calculation of 

 the remaining errors. The method has already been employed 

 by G. Kucera (Bulletin International de I'Academie cles Scien- 

 ces de Boheme, 1905"), but it was finally abandoned, be- 

 cause the magnetic lines of induction penetrated the ioniza- 

 tion-chamber, and affected the paths of the secondary ^ rays, 

 and therefore the magnitude of the current. We therefore 

 placed a thick iron screen (2*5 cm.) between the magnet and 

 the chamber ; the y rays passed through a hole in the iron. 



