168 



shown by the line c,f (fig. i.). As any radiator is increased 

 in thickness it not only sends out its own secondary, but 

 absorbs that coming from the sides of the conical hole. It 

 will be shown later how proper correction can be made for 

 this effect. 



A third set of experiments was carried out with Al as 

 radiator ; the effects were of a similar nature as those from 

 Zn, the maximum value being about the same in amount 

 as for Zn, and was obtained with a thickness of about 4'5 

 cm. ; but it was found that an appreciable difference was 

 made when a plug of Al was placed at h in place of d as 

 shown in fig. i., even though the total thickness of Al was 

 kept the same in both cases. It will thus be seen that some 

 secondary radiation can, after emerging obliquely from a 

 depth of several cm. of Al, produce an effect in the ioni- 

 zation chamber. The present form of apparatus was there- 

 fore not suitable for obtaining from Al all the results so far 

 obtained for Pb and Zn ; in these latter it was found that 

 similar errors were very small. It may, however, safely 

 be said for Al as for Zn that the effect produced is consider- 

 ably greater than for Pb. 



We have so far neglected all consideration of the sec- 

 ondary radiation which comes from the sides of the conical 

 hole ; this it will now be shown is by no means small, but 

 the correction which it introduced was found not to be suffi- 

 cient to interfere with the deduction just made. 

 Consequently we may say that Zn and Al both give out 

 more secondary emergence y radiation than does Pb, pro- 

 vided of course that the quality of the radiation in the two 

 cases is not so different as to balance the effect. 



In order to find the amount and nature of the radia- 

 tion which proceeds from the sides of the conical hole, Pb 

 domes were placed as at Z?,/?, or Pb cylinders as at ?*,?', fig. i. 

 When the latter were used, their equivalent thickness was 

 found in terms of corresponding domes. A cylinder was 

 found to be approximately equivalent to a dome 1*5 times 

 its thickness. 



The results of experiments upon the radiation from the 

 sides of the cone are shown in fig. iv., curve .4, in which 

 thicknesses of dome are shown horizontally and the corre- 

 sponding leaks per 30 sec. vertically. 



The curve A can be represented approximately by the 

 expression 500 + 570(^-i*25rf. 



That the radiation comes from the sides of the cone can 

 be shown by placing at p, fig. i., Pb plates in which circular 

 holes are cut so as not to intercept the main stream of v 

 radiation coming from the Ra. 



