EVOLUTION OF THE UNIVERSE. 595 
I have thus far reached, merely by a process of exclusion, the con- 
clusion that the cosmic rays are the wireless signals of the building in 
interstellar space of some at least of the heavier elements out of the lighter. 
The evidence, however, is very much better than that. It is reasonably 
quantitative in the case of the main cosmic ray band. Here again theory 
and experiment support one another. For, if atom-building takes place 
at all, there is one such atom-building act that is more fundamental than 
all others, and that also must take place more frequently than all others, 
namely, the formation of helium out of hydrogen, because we have 
abundant evidence that all the elements are actually built out of hydrogen 
and helium, and that helium is built out of four atoms of hydrogen, so that 
the hydrogen-to-helium transformation should take place much more 
frequently than any other. The energy of this transformation is com- 
puted from Einstein’s equation and Aston’s measurements at 25,000,000 
volts: just what I stated above to be the energy of the large cosmic ray 
band that carries the great bulk of the cosmic ray energy entering the ° 
atmosphere. But the way in which I arrived at that figure requires some 
explanation. Most simply stated, the method used was to compare directly 
in the waters of high altitude lakes the penetrating power of the cosmic 
trays there found with the penetrating power of the hardest known mono- 
chromatic gamma rays, namely, those of thorium ©”, which have an 
energy of about 2,500,000 volt-electrons. This I have done directly, the 
observed ratio being between 6 and 12. 
The comparison cannot be made directly with great precision, because 
the cosmic rays are not homogeneous, but when the inhomogeneities are 
sifted out into bands (and this does not need to be done with great 
precision, almost any reasonable kind of sifting being satisfactory, since 
the softest band carries so large a fraction of the total energy) the 
penetrating power of this softest band comes out very close to five times 
that of thorium ©”. Further, at altitudes between 6 km. and 9 km., 
where the harder components should exert almost no influence, according 
to the observations of both Hess and Kolhérster in their manned balloon 
flights, the directly measured penetrating power of the cosmic rays was 
six times that of the hardest gamma rays. This checks most satis- 
factorily with our analysis of our curve. The best formula we now have 
(the Klein-Nishina) connecting penetrating power and energy then makes 
this energy come out ten times that of thorium ©”, or about 25,000,000 
volts. This Klein-Nishina formula has been directly proved by Bowen 
and myself, as well as by others, to fit the facts up to rays of the hardness 
of those of thorium ©” reasonably well, and the extrapolation from there 
up to the softest cosmic ray band is not likely to be very badly in error. 
Indeed, this whole procedure may be looked upon merely as the extra- 
polation, to a not unreasonable distance, of an experimental curve, and 
thus as largely independent of the Klein-Nishina formula or indeed of any 
theory. At any rate, the foregoing is, in my judgment, the only 
quantitative test of the energy of any portion of the cosmic ray spectrum 
that has any sort of significance. For to use the Klein-Nishina formula 
to compute the energies of the rays, not five times but 200 times as 
penetrating as those of thorium C”, as some who have sought to make the 
cosmic rays proton-annihilation rays have done, seems to me to be extra- 
QQ2 
