306 



vary somewhat with the state of the bulb ; and since these 

 variations are comparable with those which are met with on 

 changing the nature of the screens, we are not now in a posi- 

 tion to discuss smaller variations in detail. We must con- 

 tent ourselves with quoting a few results in order to show 

 the want of symmetry, which is a persistent effect. When, 

 for example, two tinfoils were used as screen (weight per 

 square cm. of each, 0056), we obtained the following figures, 

 which represent movements of the scale in mm. during 10 

 sees. : — 



Radiator. 



Emergence Current 



Incidence Current 



With four tinfoils the figures were : 



Radiator. 



Emergence Current 



Incidence Current 



Again, using a copper screen '002 



Radiator. Pt Sn 



Emergence Current ... 86 140 

 Incidence Current ... 65 104 



Putting together a number of results for Cu screens of 

 different thicknesses we obtain the logarithmic curves of ab- 

 sorption shown in the accompanying figures. It should be 

 observed that some of the results thus shown were obtained 

 at different times, so that too much must not be built upon 

 a comparison between them ; only the relative positions of 

 the emergence and incidence curves of each substance are suf- 

 ficiently correct, and the form of each curve as showing the 

 homogeneity or otherwise of the various radiations. One 

 figure shows the emergence {E) and incidence (/) curves for 

 Pt, Cu, and Fe ; the other the corresponding curves for Sn, 

 Al, and celluloid. 



The experiments described in this paper show that a very 

 marked want of symmetry occurs iu the case of secondary 

 X-rays, the emergence rays being generally greater than the 

 incidence. This is another instance of the close parallelism 

 between X- and y rays. On a material theory of X- and 

 y rays the effect is easily explained, and is to be classed with 

 the scattering to which /3 and also, as lately shown clearly by 

 Geiger, a rays are subject. But if the X- and y rays consist 

 of energy bundles of very small volume, as suggested by J. 

 J. Thomson, then these bundles must be capable of deflec- 

 tion in going through atoms — that is to say, swung out of 

 their paths by the electrical forces to be found within the 

 atoms, just as neutral pairs would be in virtue of their elec- 



