CONTEMPORARY ADVANCES IN PHYSICS 153 



protect the ionization-chaml)er? the air, the ice, the water, the leaden 

 walls themselves? Of the atmospheric air I have already admitted 

 this; the quantity of radioactive gas commingled with it can, however, 

 be measured, is generally very small, and does not seem to bother 

 observers when they are not too close to the ground. As for the water : 

 Millikan has always carefully chosen "snow-fed" lakes, high in the 

 mountains where the water is derived not from springs which have 

 seeped through soil, but from snow which has fallen onto bare rock 

 from on high. Ice of glaciers, it is to be presumed, has an origin equally 

 uncorrupted. The same cannot be said for the water of Lake Con- 

 stance wherein Regener's data were taken, but he by special tests 

 proved the effect of its radioactivity to be slight. 



As for the lead (or whatever other metal or metals may be used in 

 making the ionization-chamber) this too may be contaminated with 

 radioactive atoms. Some people have mounted wire netting all 

 around the interior of the chamber at some distance from the walls, so 

 that such rays as come from the wall and are soft (in the sense opposite 

 to that in which I spoke of gamma- rays as hard) may be largely ab- 

 sorbed between the wall and the net; voltages are so arranged that the 

 ions produced in this space are not counted by the observer. This 

 device is not always used, nor when used is it fully effective. A some- 

 what experienced man might expect to be able to allow for the radio- 

 activity of the wall, through knowledge acquired by varying the 

 density of the air and studying the ionization as function of density; 

 this has been done, but we shall see that the results have added new 

 mysteries to the problem. 



The walls, then, must be expected to cause a permanent ionization 

 in the air-chamber, constant and independent of the outside world. 

 Moreover electrical leaks must occur, whereby the charged electrode 

 gradually loses its charge by conduction through the not-quite-perfect 

 msulators on which it is mounted. How can we hope to distinguish 

 the joint effect of these from that of the cosmic rays? or, to put the 

 same question more properly, how can we hope to find what part of the 

 ionization, if any, is not due to either of these? 



This is indeed a serious question, inasmuch as it is a problem not of 

 distinguishing a known from an unknown, but two unknowns from one 

 another. Of course, if one makes the gratuitous assumption that the 

 "cosmic rays" cannot penetrate x feet of rock, then one may take the 

 apparatus into a cave or hollow under more than x feet of rock, and 

 measure the rate at which the charged electrode is discharged; and 

 then, after deducting the allowance to be made for the rays from radio- 

 active substances in the rock which penetrate into the chamber, one 



