39 



radiation, is too small in all cases, because the screen, when 

 placed on the bottom of the chamber, stops a certain amount 

 of yS radiation, which is made in the air just underneath the 

 plate, and is out of reach of the magnet, as well as some fast 

 )3 radiation which comes from lower down, and is strong 

 enough to escape from the magnetic field. This is reckoned 

 in a, but not in h ; so that on this account all the emergence 

 radiations are too small by a certain constant amount. We 

 have not yet succeeded in determining this constant with any 

 accuracy. We believe it to be mainly due to the fast ^3 rays, 

 which emerge with hard y rays from the upper surfaces of 

 the magnet poles, of the lead block, and of the plug. It can- 

 not be much less when the plug is removed, and must there- 

 fore be of more relative importance to the results for hard rays 

 than for soft, since the total observed effects are smaller in 

 the former case. Its magnitude is, perhaps, indicated with 

 some accuracy by the fact that when a lead plug of 3 cm. 

 thickness was used the emergence radiation of tellurium 

 seemed to be nearly zero : the substance had stopped almost 

 as much as it generated. It can easily be seen from what fol- 

 lows that tellurium should probably have less emergence radia- 

 tion than any other substance, yet it ought to approach half 

 the value for carbon ; and this would imply that the value of 

 the constant was about 300 for the hard rays, and perhaps 

 rather mere for the soft. On the other hand, c — a, the inci- 

 dence radiation, is too large, because the plate that is placed 

 on top of the chamber at q,q not only gives rise to the inci- 

 dence radiation to be measured, but also turns back to a 

 greater or less extent the ^3 rays striking it from below. This 

 effect' increases with the atomic weight, and must be of some 

 importance in the case or Sn and Pb. As it is clear that 

 these corrections will increase the want of symmetry, already 

 obvious enough, and as we are hardly in a position as yet to 

 make the corrections with accuracy, and as we hope to dimi- 

 nish our experimental errors in the future, we have for the 

 present left these figures uncorrected. 



In the first of these papers we showed that the incidence 

 radiation should be somewhat less than p times the emergence 

 radiation where p is the reflection constant of the substance 

 in question for ^3 rays. The above table does indeed show that 

 the ratio of the two radiations increases with the atomic 

 weight, and therefore with p \ but the quantitative comparison 

 appears poor. But it is to be remembered that (1) corrections 

 yet to be made will alter the figures somewhat, making, for 

 example, the emergence radiation of Pb greater, and the 

 incidence radiation smaller ; (2) the quantity p is somewhat 

 indefinite. It is true that McClelland has made careful mea- 



