139 
126, 127 J Highly Penetrating Radiation 
takes account of the relativity-correction and leads to very different results. 
Gray has found that the available experimental evidence on the whole sup 
ports the new formula. And the observed penetrating power of the radiation 
agrees excellently with that which it assigns to radiation of 940-million volts 
energy, such as would be produced by the annihilation of protons and electrons'^. 
Rosseland j has suggested that bombardment by this radiation may be the 
cause of the observed bright lines in stellar spectra; I had previously § 
suggested a similar origin for the luminosity of the irregular nebulae. 
Finally we may notice that, as a gramme of matter contains 9 x 10 20 ergs 
of energy, our estimate of 0 = 43 ergs per gramme per second for the rate at 
which energy is generated by the annihilation of matter throughout the 
universe, assigns to the matter of the universe an average expectation of life 
of 2 x 10 19 seconds, which is only about 600,000 million years. But our whole 
estimate is of so rough a character that this figure can claim at most to be 
accurate as regards order of magnitude. 
Summary. 
127. The general impression produced by our analysis of observational 
facts is that of a universe slowly but inexorably dissolving into radiation. 
The process may not properly be compared to that of ice dissolving into water, 
a process whose speed is governed by the rate at which heat is supplied from 
outside, nor to gunpowder dissolving into smoke, a process whose rapidity is 
checked only by mechanical conditions extraneous to the process itself; it 
should rather be compared to the dissolving of uranium into lead, a process 
which pursues its steady course uninfluenced by all external factors. 
One reservation only need be made. We have found reasons for supposing 
that the complete ionisation of stellar matter inhibits the process of trans 
formation into energy, and it seems likely that partial ionisation may produce 
a partial inhibition of this process. The chemist who is given a sample of rock 
in which lead and uranium are mixed can estimate the age of the rock by 
measuring the proportion of uranium which remains. The astronomer who is 
given a sample of stellar matter cannot estimate its age in a similar way; he 
must first know for how long it has been protected from dissolution by being 
in a state of complete or partial ionisation. The ages which we have calculated 
for the stars may need to be substantially increased for this reason, and what 
appear to be the youngest of the stars may conceivably be constituted of atoms 
which are as old as any in the universe but have been protected by ionisation 
from dissolution throughout the greater part of their lives. 
t See a general discussion by Sir E. Rutherford, Proc. Roy. Soc. 122a (1929), p. 15. 
+ Astro. Journ. May, 1926. 
§ Nature, December 12, 1925.