592 DISCUSSION ON THE 
be a beam in complete equilibrium with its whole train of secondaries, 
including the deflectable beta rays, and these latter would of necessity 
spiral about the earth’s magnetic lines and thus enter the earth only near 
the earth’s magnetic poles, a behaviour which clashes with fact No. 2. 
Since, then, no trace of such an effect is actually found, the cosmic rays 
alone deny the validity of this hypothesis and with it of this form of the 
annihilation hypothesis as to their origin. It seems to me, therefore, that 
as an explanation of facts about cosmic rays the annihilation hypothesis 
fails at every point at which one can test it. I shall comment upon still 
another of its failures in the next section. 
(3) The third well-established and most significant fact of cosmic 
rays 1s that they have a banded structure. Cameron and I first brought this 
fact sharply to ght in 1925,” but I do not think the evidence is yet clearly 
understood or its significance fully appreciated. The evidence goes back 
to the high balloon flight’® with self-recording electroscopes which Bowen 
and I made in 1922, a flight in which we reached an altitude of 154 km. 
and got 92 per cent. of the distance to the top of the atmosphere as measured 
by the weight of the earth’s atmosphere left beneath our instruments. 
In other words, remembering that the atmosphere is the equivalent of 10 metres 
of water, we were within 80 centimetres of the top. Now, 80 centimetres of 
water allows at least 3 per cent. of rays as soft even as those of thorium 0” 
to pass through it. We made this high flight on purpose to determine 
whether the rate of increase of ionisation within a closed electroscope 
continued exponentially to the top of the atmosphere as Kolhérster’s 
earlier balloon flight to a height of 9 km. had indicated was the case up 
to that altitude. Further, Cameron and I have since, with precision 
instruments, obtained quite as high ionisation readings up to 5 km. as 
those given by Kolhérster, thus checking sufficiently his exponential 
curve in the lower stretches of the atmosphere ; but the 154 km. balloon 
fight not only failed to do this, but it showed definitely and altogether 
unambiguously, since all possible sources of error would have increased, 
not decreased, our readings, that with increasing altitude the ionisation 
fails to maintain its rate of increase but drops back toward lower values, 
just as it should do if a band of about 25,000,000-volt pure photons enters 
the atmosphere and requires a considerable distance of penetration into 
it before it gets into equilibrium with its secondaries. This furnishes a 
third bit of independent evidence that the cosmic rays do enter the 
atmosphere as practically pure photons, and hence that they originate in 
interstellar space. 
The high balloon flight shows much more than this. It proves con- 
clusively that neither gamma rays of energy 2,500,000 volts, like those 
from thorium C”, nor rays of intermediate penetrating power up to that 
of the softest cosmic rays, which pass through about five times as great a 
thickness of water and have an energy of about 25,000,000 volts, enter 
the earth’s atmosphere in amounts appreciable in comparison with that 
of this softest cosmic ray band; for any such abundant rays would have 
® Millikan and Cameron, Phys. Rev., 28, 851; 1926: see also Phys. Rev., 87, 244 ; 
1931. : 
10 Millikan and Bowen, Carnegie Institution Year-book, 24, 385; 1922: also 
Phys. Rev., 22, 198 ; 1923, and 27, 353; 1926. 
