636 BELL SYSTEM TECHNICAL JOURNAL 



Struck centrally by alpha-corpuscles as fast as these. (That many of 

 the nuclei causing scintillations did not have so great a range is easily 

 accounted for; it is due to the fact that most of the impacts are sensibly 

 "off-centre," the struck particles flying off obliquely with less energy 

 than they would have derived from a "square hit.") But at thirty 

 centimeters of "air-equivalent," they cease entirely; this sustains us 

 assuming that if with any other gas or any solid there are scintillations 

 when the screen is so much shielded, they cannot be due to admixtures 

 of hydrogen. 



Curve C was obtained with oxygen ; what there is of it is ascribed to 

 commingled hydrogen; in any case, it does not extend beyond the 

 critical point at which, were there any flashes still to be seen, they could 

 safely be attributed to something else. 



Curve A is more sensational : very definitely it extends beyond the 

 critical length ; very definitely there are corpuscles able to make their 

 way through deeper strata of matter than either the primary alpha- 

 particles or such nuclei of stray hydrogen atoms (so both theory and 

 experiment assure us) as these might find to strike. This curve was 

 obtained with air. Since with pure oxygen there was no sign of such 

 extraordinary corpuscles, it is to be presumed that they were due to the 

 other of the major gases of the air — an inference which the study of 

 pure nitrogen made sure. 



The most astonishing of all the curves is D. It stretches far beyond 

 the critical point; flashes appeared on the screen when even as much as 

 the equivalent of ninety centimeters of air lay between it and the sub- 

 stance which the alpha-rays were striking, which was aluminium in the 

 form of a thin leaf. Thus, when foil of aluminium is subject to the 

 impacts of these rays, it throws out corpuscles three times as penetra- 

 tive as the very fastest which a critic might possibly discredit by ascrib- 

 ing to occluded hydrogen. 



Are these, then, fragments of disrupted nuclei of aluminium or ni- 

 trogen? and are they protons? 



In principle the second question is answerable by itself; it is sufficient 

 to deflect the corpuscles by electric and magnetic fields, and measure 

 their deflections; the value of their charge-to- mass ratio (which if they 

 are protons is about .00054 of the value for an electron) could then be 

 computed, and incidentally their speed also, which itself would be 

 well worth determining in a way more direct than by inference from 

 the range. But though such measurements have many times been 

 made on other kinds of particles, and the technique is very well de- 

 veloped, the application to those of this especial kind is difficult be- 

 cause they are so few. Say that the apparatus is so built that they 



