226 ANNUAL REPOET SMITHSONIAN INSTITUTION, 19 2 8 



process, of some, at least, of the common elements out of the 

 primordial positive and negative electrons. 



Fourth. — The evidence herewith obtained is, however, not merely 

 qualitative but, if Dirac's formula is dependable, fairly accurately 

 quantitative. For we analyzed very carefully our cosmic-ray curve 

 empirically before we called on any theoretical considerations 

 whatever to explain it, and we reported in scientific papers 

 that our observed curve demanded three cosmic-ray bands 

 of absorption coefficients 0.35, 0.08, and 0.04 per meter of water, 

 respectively. It was after this work had been done, reported in 

 seminars, written up, and prepared for publication in essentially the 

 form in which it has appeared -* that we set about computing from 

 the foregoing considerations what the theoretical absorption co- 

 efficients would be if our observed cosmic raj^s were produced (1) 

 by the formation in one single act of helium out of hydrogen, (2) 

 by the similar formation of oxygen out of hydrogen, and (3) by 

 the formation of silicon out of h3^drogen. The results of this com- 

 putation with the aid of Dirac's formula came out 0.30, 0.075, 

 and 0.043, well within the limits of the resolving power of 

 our curve of the observed values. Further, there are only a few 

 elements so abundant that their formation needs to be consid- 

 ered as the possible source of the observed cosmic rays. For the 

 spectroscopy of the heavens shows a very great abundance every- 

 where of the gases hydrogen, helium, and nebulium. But Bowen -^ 

 has just identified the last as nitrogen and oxygen so that these 

 gaseous elements, hydrogen, helium, nitrogen, and oxygen seem to 

 be extraordinarily widely spread through space. As to the elements 

 found in solids the meteorites have 96 per cent of their mass in the 

 four elements, oxygen, magnesium, silicon, and iron. Magnesium 

 and silicon are close together in atomic weight, 24 and 28, respec- 

 tivel}'^, so that their formation would constitute but one. band, the 

 mean energ}^ of which lies approximately at ju, = 0.04. It was this 

 joint band that we above called for convenience the silicon band. 

 Similarly the atomic weights of nitrogen and oxygen are respectively 

 14 and 16, and the mean absorption coefficient 0.08 corresponds to 

 this joint band. The helium band, most significant of all, corre- 

 sponds to /A=0.30. There is no other abundant element except iron, 

 and the formation of this out of hydrogen gives a cosmic ray for 

 which /A= 0.021. The existence of such a radiation helps rather than 

 interferes with the fit of our theoretical and experimental curves, 

 but on account of the lack of resolving power in the lower end of our 

 jurve it furnishes no trustworthy evidence that this particular act is 

 the one by which iron is most commonly formed. This uncertainty 



=" Millikan and Cameron. Phys. Rev., June, 1928. 



=» Bowen, Astrophys, Jour. 57, 1, 1928. 



