EVOLUTION OF THE UNIVERSE. 591 
Tf, then, the cosmic rays are forming now, there is no place of origin 
left except outside the stars, in the interstellar spaces,where both temperature 
and pressure are exceedingly low. This is not so unlikely a place of origin 
as it was a few years ago, before Bowen" at the Norman Bridge Laboratory 
had solved the seventy-year-old riddle of the nebulium lines and proved 
that the common elements oxygen, nitrogen, carbon, and sulphur, as well 
as hydrogen and helium, exist out there, giving rise to these nebulium 
radiations in regions which may be lightyears away from the exciting-stars. 
(2) The second most significant cosmic ray fact is that which, after 
some less precise tests by Cameron and myself in 1926, was brought to 
light most unambiguously just a year ago when, by taking very careful 
and very exact readings with the same sensitive electroscope, first at 
Pasadena, lat. 34° N., then at Churchill, Manitoba, lat. 59° N., the nearest 
settlement on earth to the north magnetic pole, I proved that the incoming 
cosmic rays are not influenced at all by the earth’s magnetic field, and 
drew from that fact the inevitable conclusion that when these rays enter 
the earth’s atmosphere they have not previously traversed any amount 
of matter which is comparable with the thickness of that atmosphere, or 
else they would of necessity be mixed with secondary beta rays which 
would be directed into the earth most abundantly along the earth’s 
magnetic lines and therefore enter it in the neighbourhood of the magnetic 
poles. These experiments furnish another independent and, I think, 
very beautiful proof that the rays must originate in interstellar space, for 
if they came from even the remotest exteriors of stars, they would have 
to be appreciably mixed with magnetically deflectable beta rays. The 
fact that they are not so mixed seems to me to hit the annihilation theory 
a second fatal blow, for there is no sort of reason for assuming that 
annihilation takes place only in interstellar space. Such an assumption 
would destroy the whole purpose of the annihilation hypothesis, which up 
to the present has been to furnish the requisite energies to the stars. 
Before I leave this point (No. 2), let me combine the experimental fact 
of uniformity of distribution (No. 1) with this other fact that the rays 
enter the atmosphere unmixed with deflectable beta rays. The uniformity 
of distribution could indeed be reconciled with the annihilation hypothesis 
if it could be assumed that, in looking out from the earth into the celestial 
dome, one were always looking through a quantity of matter equivalent 
to, say, a hundred metres of water, which is a thickness sufficient to absorb 
99 per cent. of all incoming cosmic rays. That one would actually 
encounter one-thousandth part as much matter as this in going out to 
infinity through interstellar space seems to be contrary to all the other 
astronomical evidence. But let us forget this and see what the cosmic 
rays alone have to say about this point. If annihilation is going on in all 
matter independently of temperature, then so far as the cosmic rays coming 
to the earth’s atmosphere are concerned, every element of the celestial 
dome would be like every other element, even if the sun were in one of 
these elements. This is the only way the annihilation hypothesis of the 
origin of cosmic rays can be reconciled with their uniformity of distribution. 
But in this case the whole of the cosmic ray,beam entering the earth would 
8 Bowen, Astroph. Jour., 67, 1; 1928, 
