188 



calculated above may be placed in the position shown by C, 

 fig. iii. The curve D is now drawn to represent the differ- 

 ence between the curves B and C . We may now see liovv 

 closely this remainder corresponds to what might be expected 

 if the soft bundle of original y rays produced a correspond- 

 ing softer secondar}-. The maximum in the case of this re- 

 mainder curve is further to the left than in the case of curve 

 C, and after reaching its maximum value the curve falls away 

 much more rapidly than does C . It corresponds, in fact, 

 very well with what we might reasonably expect to be pro- 

 duced by the soft bundle of y rays coming from the Ra. 

 Again, although for any given thickness of radiator the ratio 

 of the ordinates to the curves C and D is by no means con- 

 stant, it is of the order indicated in fig. v. by the curve A" . 



If, now, the curves C and D represent the effects due 

 respectively to the hard and soft y rays from the Ra, it 

 should be possible to show experimentally that the quality 

 of the radiation from a Zn radiator depends to some extent 

 upon the thickness of the radiator. No serious attempt has 

 so far been made to carrj- out this investigation. However, 

 we may check the result in another way. The quality of the 

 radiation from Zn should show some change if we absorb 

 some of the soft y rays coming from the Ra by means of a 

 Pb screen or plug before allowing the radiation to fall upon 

 the Zn radiator. As we have already seen, it is possible in 

 this way to reduce the softer radiation without cutting down 

 the hard to such an extent as would be necessary if a plug 

 of material such as Zn were used. 



Using 17 mm. of Al as radiator, with no plug, the read- 

 ing with no dome was to the reading with a 6 mm. dome as 

 100:31. With a 9 mm. plug, however, inserted in the coni- 

 cal hole just over the Ra the reading with no dome was to 

 the reading with a fi mm. dome as 100:39. A similar effect 

 was also obtained with a Zn radiator. 



We see, then, that hardening the primary radiation has 

 the effect of hardening the secondary, and this is in agree- 

 ment with the suggestion put forward, viz., that in the case 

 of Zn, Al, and such substances the secondary emergence y 

 radiation consists of two bundles corresponding to, in fact 

 derived immediately from, the two corresponding bundles con- 

 stituting the original y radiation. 



In fig. ix. the dotted curve C has been plotted to a larger 

 scale, and the curve D obtained previously in fig. iii., and 

 shown in that fig. curve B, has been plotted with the same 

 maximum value as C for comparison with it. It will be 

 noticed that the curve D reach ea its maximum for a smaller 

 thickness of radiator than does C . This we might expect 



