BEYOND THE MILKY WAY. 563 
On the face of it, this looks as though the whole universe were uniformly expanding, 
like the surface of a balloon while it is being inflated, with a speed such that it doubles 
its size every 1,400 million years. 
One of the great puzzles of astronomy at the present moment is whether these 
apparent motions of recession are real or not. They are deduced from spectroscopic 
observations ; the nebular spectra show displacements to the red, which, interpreted 
in the most obvious way as Doppler effects, give the speeds already mentioned. Yet 
every spectroscopist knows that many factors besides motions of recession are capable 
of reddening light. 
There is one strong theoretical argument in favour of regarding the apparent 
speeds as real. LHinstein’s original cosmology regarded the universe as being as full 
of matter as a universe of its size could possibly be without violating the theory of 
relativity. Recently, Lemaitre, of Louvain, has shown that a universe of this type 
would not be static—there would be an unstable quality about it. The condensation 
of the primeval gas into distinct nebule, and the imprisonment of a large part of the 
free energy of the universe in these nebule would cause it to start expanding, in 
which case it would continue to expand, its radius finally increasing exponentially 
with the time, until it ended up as an empty universe—finite matter spread through 
infinite space. Throughout the motion, the relative speed of recession of any two 
nebule would be exactly proportional to their distance apart, so that, at first glance 
at least, this theory seems exactly to fit the observed facts. It not only provides a 
suggestion as to why the nebulz may be receding. It goes much further and predicts 
that they must be receding ; if Einstein’s relativity cosmology is sound, the nebulz 
have no alternative—the properties of the space in which they exist compel them to 
scatter. 
Yet various circumstances suggest a need for caution. For one, the speeds of 
the nebulz are not strictly proportional to their distances, and it is not easy to explain 
the discrepancies, which do not look like mere random motions, Again, the very 
magnitude of the apparent speeds casts doubt on their reality ; they would reduce 
the whole existence of the universe to a mere flash—at any rate in comparison with 
what we have recently believed. If they are real, Eddington has calculated that the 
universe must have started from a radius of about 1,200 million light-years, and that 
its total mass must be about 2-3x10°° grammes, which is the mass of 1-4 107 
protons and an equal number of electrons. So far as we can tell from the masses of 
the extra-galactic nebule, the present average density of matter in space appears to 
be not less than 10-° grammes per c.c., which, with the same amount of matter, 
would assign a radius of 13,200 million light-years to the universe—only eleven times 
its initial value. If, then, the motions are real, the universe is only at the beginning 
of its career ; it cannot have doubled many times since it started. And as it appears 
at present to be doubling in size every 1,400 million years, the few doublings which 
these figures permit cannot have occupied more than 10,000 million years at most. 
General calculations on the ages of astronomical bodies point to far longer periods 
of time than this. The mere act of condensation of the nebulze—the occurrence which 
seems most likely to have started the expansion—was probably a matter of hundreds 
of thousands of millions of years. Perhaps there is no real difficulty here; it might 
well take this long to get the doubling process really going—indeed, Lemaitre has 
calculated that it would take some such time. The real difficulty is that the stars 
carry intrinsic evidence of having lived through far longer periods than this. Both 
single stars and binary systems show an approximation to equipartition of energy 
which must have taken far longer than this for its establishment. 
Spectroscopic binaries provide further evidence. These consist of pairs of stars 
revolving round one another. Observation reveals a complete sequence; it begins 
with systems which appear to have just broken into two as the result of rotation— 
pairs of stars describing circular orbits, and almost in contact—and ends with systems 
in which the two stars are far apart and describing elliptical orbits. Theory suggests 
that this observational sequence exactly depicts the evolution of a star which has 
broken into two as the result of excessive rotation. The outstanding importance of 
this sequence to our present discussion lies in the fact that the stars at the beginning 
of the sequence are undoubtedly many times more massive than those at the end. It 
seems likely that those which are now at the end must have begun at the beginning 
and lost the greater part of their mass in the form of radiation, and to do this would 
take millions of millions of years. Considerations such as these make it very difficult 
to believe that the universe can be such an ephemeral concern as the apparent speeds 
002 
