Mr. Norman Campbell on Delta Rays. 293 



would be if there were no reflexion : it will make the curve 

 steeper than it should be. The second action will confuse 

 the two branches of the curve, for they can be separated 

 entirely only if on one branch all the rays from one electrode 

 are received by the other even for the lowest potential dif- 

 ferences. As the potential difference is increased the current 

 will be changed, not only by the stoppage of rays leaving 

 the electrode with which the branch is concerned, but also 

 by the retention of more of the rays emitted by the other 

 electrode. This action should make the rays appear less 

 easily stopped by small potentials and make the curve flatter 

 than it should be. I think, however, that there can be no 

 doubt that the former effect, when the field and reflexion act 

 in the same direction, is the more important. 



Accordingly, though reflexion may modify considerably 

 the form of the curve, there appears to be no reason why it 

 should render invalid conclusions obtained by the comparison 

 of different curves in all of which the geometrical arrange- 

 ments and the material of the electrodes are the same, unless 

 the average speed of the particles varies. In order to escape 

 from the conclusion that the speed of the delta rays is inde- 

 pendent of that of the alpha rays for the same metal by the 

 consideration of reflexion, it must be supposed that the varia- 

 tion of reflexion with the speed is such as exactly to com- 

 pensate the variation due to the direct effect of the change 

 of speed. Such an assumption would appear somewhat 

 forced, but it must be noted that the two effects act in com- 

 pensating directions, von Baeyer shows that, within the 

 range of speeds from 5 to 10 volts, with which we are largely 

 concerned, reflexion increases with the speed. Now if, as we 

 have concluded, reflexion tends on the whole to make the 

 curve steeper, and the direct effect of an increase of speed 

 undoubtedly tends to make the curve flatter, the two effects 

 might compensate each other. But the closeness of the 

 agreement shown in Table II., when considered in the light 

 on the enormous reflexion found by v. Baeyer, appears to me 

 to render such an explanation unsatisfactory. 



It appears that reflexion might well be called in to explain 

 the divergence between the estimates of the difference of the 

 emergence and incidence radiation noted in § 12. For the 

 existence of reflexion would doubtless tend to diminish the 

 apparent inequality, as may be easily seen by a consideration 

 of the analogous optical case. When V — reflexion has a 

 chance to exert its full effect ; but when V = 40 and saturation 

 is attained, reflexion has ceased to have any influence. 



But the question of reflexion is most troublesome when 



