© 1948 SCIENTIFIC AMERICAN, INC

GALAXIES IN FLIGHT
If the island universes are indeed racing away

fron1 one another, the fact n1ay shed light on

the primordial fOrInation of nuclei and aton1S

by George Gamow

N THE YEAR 1929 the Mount Wilson

I
iruitful study, not only in astronomy but
astronomer Edwin Hubble made a also in geology, physics and chemistry.
ver y remarkable discovery. He found It gave us a new start fur investigating
that the giant accumulations of stars the age of the universe and the creation
known as galaxies, which are scattered in of the stuff of which it is made. If our far­
great multitude through the vast expanses flung cosmos came originally frum a dense
of the universe as far as the best telescopes hot core of material concentrated in one
can see, seem to be running away from place, then we can reasonabl y assume that
one another at fabulously high speeds. this tightly packed core must have con­
From this observed fact originated the sisted in the beginning of elementary
famous theory of the expanding universe. building blocks, most of them probably
Although the theory is still not finally neutrons. out of which all the chemical
proved, it seeded a whole generation of elements later were made. I shall discuss
briefly some recent studies of this phase
GREAT NEBULA in Andromeda is of the expanding-universe theory which
full-size g ala xy closest to Millo, Way. have been made by Ralph Alpher, Hans
Photograph b�- Lick Observatory_ Bethe and George Gamow. The main sub-

© 1948 SCIENTIFIC AMERICAN, INC

 ject of this article, however. is the basic ilUclei of spiral ones than in the spiral
theory itself, and how it stands up today, anns. It was only several years ago that
19 years after Hubble's discovery. Walter Baade (If the Mount Wilson Ob­
The idea of stellar galaxies is a com· servatory succeeded in resolving these in­
paratively recent discovery in astronomy. terior stars by using special photographic
The celestial shapes that we now recog­ plates and carrying out the exposures with
nize as galaxies had been observed for a great care. His pictures revealed an un­
long time as faint nebulosities of various expected fact: the stars forming the nu­
regular forms, but they were generally clear regions of spiral galaxies, and all
believed to be simply luminous clouds of stars of the armless galaxies, have very
gas floating in the spaces between the different physical characteristics from
stars of the Milky Way. Observations with those in the spiral arms. The meaning of
more powerful telescopes, however, reo this difference in stellm' population is not
solved these "nebulosities" and showed clear, but there is no doubt that it has an
that they were not clouds but huge col­ extremely important bearing on stellar
lections of extremely faint stars. These and galactic evolution.
giant stellar aggregates were far beyond The galaxies are scattered more or less
the outer limits of our own stellar system, uniformly through space as far as our
the Milky Way; in fact, it soon became telescopes can probe. The average dis­
clear that they formed systems very simi· tance between neighboring nebulae is
lar in shape and structure to the Milky about two million light-years. The limit
Way galaxy itself. of our vision with the 100-inch telescope.
The nearest and most familiar external the largest with which observations have
galaxy is the great nebula in Andromeda, yet been made, is about 500 million light.
which 'can be seen with the naked eye years. Hence in the observable region of
SPIRAL nebula in Canes Venatici
as a faint, spindle-shaped speck of light space there are some 100 million galaxies.
is seen along a line perpendicular
in the upper part (from the Northern The new 200-inch telescope on Mount
to its long axis. Photograph by the
Hemisphere) of the constellation of An­ Palomar, which will double the distance
60-inch reflector at Mount Wilson.
dromeda. Photographs made with large we can see into space, may reveal an
telescopes show that this galaxy has a enormously larger number. Most galaxies
rather complicated structure consisting are isolationist, dwelling in remote and
of an elliptical center, or "galactic nu­ solitary splendor, but we find a number
cleus," and "spiral arms" flung into the that group themselves together to form
surrounding space from the central body. more or less compact clusters. In the con­
The photographs also show two nearly stellation of Corona Borealis, for example.
spherical nebulosities close by, probably there is a cluster containing some 400
satellites of the central system. galaxies. Our Milky Way is a member of
Among the myriads of stars in the arms a small cluster which embraces, among
of the Andromeda Nebula are many pul­ others, the Andromeda Nebula and the
sating ones, of the type called Cepheid two galaxies known as the Magellanic
variables. They brighten and fade in a Clouds, which are of a relatively rare
regular rhythm, and their pulsation pe· type that has no well-defined shape.
riod provides a method of determining The distances of all but the nearest
their absolute brightness. By comparing galaxies are so great that even the most
their apparent brightness (which depends powerful telescopes fail to resolve them
on their distance from us) with their cal­ into individual stars. Astronomers' calcu­
culated absolute brightness, Hubble was lations of their distances depend entirely
able to prove that the Andromeda Nebula on their apparent brightness. Hubble.
EDGE of galaxy N.G.C. 4594 in Vir­ is some 680,000 light-years from the studying a group of about 100 well·known
Milky Way. To a hypothetical observer in neighboring galaxies, established the fact
go faces the Milky Way. Below:
the Andromeda galaxy, the Milky Way that on the average they were of about
"harred" spiral N.G.C. 5850 suggests
would look much the same as the Androm­ the same size and the same intrinsic lu­
galaxy in earlier stage of formation.
eda system looks to us, except that the minosity. Using this standard, we can esti­
spiral arms of the Milky Way are some­ mate the distances of remote groups of
what more open. Our sun, with its family galaxies by comparing their mean appar­
of planets, would be seen through a tele­ ent brightness with that of nearby galaxies
scope within the Andromeda Nebula as a whose distances are known. Such measure­
rather faint star near the end of one of ments give the value of 7.5 million light­
the spiral arms, some 30,000 light-years years for the distance of one of the nearest
from the Milky Way center. groups of galaxies in Virgo. Similar galac­
tic groups in the constellations of Coma

T
HE GALAXIES generally are shaped Berenices, Corona Borealis and Bootes
like a discus. The Andromeda system are respectively 30 million, 100 million
looks like an elongated spindle to us be­ and 180 million light.years away.
cause it is tilted to our line of sight, but

N
there are many other galaxies that we see OW WHAT was it that gave Hubble
from the top or straight on edge. Most the notion that the galaxies are run­
galaxies have the same sort of spiral arms ning away from one another and that the
as the Milky Way and Andromeda, but universe is expanding? His basic dis­
there are also some armless ones. Indi· covery was made with that indispensable
vidual stars are much more difficult to tool of the astronomer, the spectrograph,
distinguish in armless galaxies and in the which analyzes the color components of

© 1948 SCIENTIFIC AMERICAN, INC

the light coming from stars. Studying the You must not conclude from this that sion" that started its headlong expansion.
spectra of distant galaxies, he noticed a we stand at the center of the universe and To get this figure, we must know the ex·
curious fact: all the lines in their spectra. that all the rest of it is running away from act values for the distances and the reces·
regardless of the wavelength or color of us. Picture a slowly inflated rubber bal· sion velocities of distant galaxies. This is
the line. were displaced toward the red loon with a large number of dots painted less simple than it sounds. The velocities.
end of the spectrum. Furthermore. the on its surface. An observer on one of the as we have seen. can be computed from
amount of this "red shift" was always di· spots would be under the impression that the observed red shift. and the distances.
rectly proportional to the distance of the the other dots were racing away from him presumably. from the galaxies' apparent
galaxy from us. The most natural expla· in all directions, and so indeed they would brightness. But there is a catch: the ap·
nation of this shift was that the source of be, but the same thing would be true no parent brightness of the stars is affected
the light was moving away. This is the so· . matter which dot he was OIl. In the case not only by their distance but also by the
called Doppler effect, of which the classic of the galaxies, we are dealing with the fact that the light coming from them i,
redder, and therefore carries less energy.
than if the light source were stationary.
To illustrate this, suppose for a moment
I�I ;11 I ill that you .are shot at by a gan ,;ster operat·

{ .- ing a submachine gun from the back win·

dow of a speeding car. Since the vehicle
, , r II I 'fli is receding, the bullets move more slowly
toward you than they would from a sta·
N.G.C. 221: 125 MILES PER SECOND 900,000 LlGHT·YEARS
tionary gun, and they strike your bullet·

III il i;lI proof jacket with less energy. A receding

light source produces exactly the same
.,�- effect; its emitted light quanta strike the
eye with less energy and therefore look
I j I II I �fl redder than they should. An astronomer
N.G.C. 4473: 1,400 MILES PER SECOND 6,000,000 LlGHT·YEARS must make the same correction for the
weakening of light intensity as a ballistics
expert would make in estimating the muzo
zle speed of the bullets.
•
•

There is a further complication. If the

submachine gun shoots, say, one bullet
per second, its bullets will strike you at
N.G.C. 379: 3,400 MILES PER SECOND 23,000,000 LlGHT.YEARS
longer and longer intervals as the gun

. . . recedes, for each successive bullet will

.. .... . have farther to travel. Similarly. light

quanta from receding stars enter the ob·

. server's eye with less frequency, and this

fact calls for another correction of th{'
URSA MAJOR I�EBULA: 9,600 MILES PER SECOND 85,000,000 LlGHT·YEARS observed brightness.
• Applying both corrections, and taking
the most accurate possible observations.
. Hubble calculated that the universe be·
.�
gan to expand less than one billion yean­

. ,
ago. This result stands in contradiction to
:eological evidence, which indicates that
GEMINI NEBULA: 14,300 MILES PER SECOND 135,000,000 LlGHT·YEARS
the age of the solid earth crust, estimated
RED SHIFT of various galaxies is illustrated hy comparing their spec­ quite reliably from radioactive decay in
tra with that of helium (above and below each nebular spectrum). the rocks, must be at least two billion
Shift is demonstrated hy length of arrows pointing to spectral lines of years. Since numerous pieces of evidence
calcium. Letters N.G.C. hefore galaxies stand for New General Catalogue. in various sciences support the two billion­
year estimate. Hubble was forced to reo
consider the expansion theory and con·
and most familiar example is the change effect of a uniform expansion throughout sider the possibility that the red shift wa'
in pitch of a locomotive whistle as the all of space. due not to the normal Doppler effect but
train approaches us and then speeds If you pick an arbitrary point in space, to some unknown physical factor which
away. A light wave. like a sound wave, say the earth, and divide the distance of caused light to lose part of its energy duro
appears to shift to a longer wavelength a given galaxy by its recession velocity. ing its long trip through intergalactic
when it reaches us from a receding source. you get a figure which represents the space.
And the speed with which the source is length of time that the galaxy has been Such a conclusion would ruin many
'
moving away is directly proportional to receding from that point. The strange and beautiful scientific developments tha t
the shift in wavelength. Since the red wonderful consequence of Hubble's obser· have flowed from the hypothesis of the ex·
shift of the galaxies also varied as their vations is that the figure will be the same panding universe. It would confront physi·
distance from us. Hubble concluded that no matter what pair of galaxies you pick. cists with the difficult task of explaining
the speed of the receding stars was propor· Thus it works out that at a fixed, calcula· the red shift in non.Dopplerian terms­
tional to their distance; the farther awav ble time in the past all the galaxies now so which would seem to contradict everything
they moved from one another, the faste� widely scattered were packed tightly to· we know at present about light. Fortu·
they traveled. The red shilt of the most gether in one place. And the time figure nately. there is a simple way out of the
distant galaxies that have thus far been you arrive at is the age of the universe, dilemma which is usually overlooked by
observe::! is 13 per cent, which suggests measured from that instant when the orig' the proponents of the "stop-the.expan.
that they are receding from us at the ter· inal condensed lump of universal matter sion" point of view. The point is that Hub·
rific velocity of 25,000 miles per second. was torn apart by the primordial "explo. ble's method of estimating the distance_

23

© 1948 SCIENTIFIC AMERICAN, INC

of faraway galaxies assumes that at the individual galaxies. Already they have tude with the age of the universe as esti·
moment when they emitted their light they demonstrated quite clearly the danger of mated from the red shift and other evi·
were just as bright as the galaxies we see building any conclusions on the hypotbe· dence. Thus we have fairly good reason
closer at hand. It must be remembered. sis of constancy of galactic brightness. to suppose that the radioactive elements
however, that the light we see [rom tbe Having made tbis fiery defense of the were formed at the beginning of the uni·
distant galaxies was emitted at a fantas· right of our universe to expand. let us verse.
tically distant time in tbe past; the light consider the physical consequence of the Actually, the picture presented by the
now coming to us from the Coma Bere· expansion tbeory suggested at the begin. expanding universe theory, which assumes
nices cluster, for example. started on its ning of this article. What pbysical process that in its original state all matter was
way some 40 million years ago. and the was responsible for the present relative squeezed together in one solid mass of
most distant �alaxies used by Hubble in quantities of the various chemical ele· extremely high density and temperature,
his studies are seen as they were almo.t ments that make up the universe? Why, gives us exactly the right conditions for
half a billion years ago! building up all the known elements in tbe
Do we have the right to assume that the periodic system. As I have mentioned.
galaxies, which are evolving like every· Alpher, Bethe and Gamow have attempted
thing else in the universe, have kept til"ir to reconstruct in some detail the processes
luminosity constant over such long period" by which the various elements may have
of time? In view of the known facts about been created during the early evolutionary
the evolutionary life of individual star •. stages of the expanding universe.
which maintain their luminosity by the ex·
penditure of nuclear energy. such an a,· UR STUDIES indicate that, under the
""mption would be very strange indeed. O tremendous temperatures and densi·
Actually. we can remove the entire diffi· ties prevailing in the nucleus of the uni·
culty in Hubble's time scale by remember·

1
verse during the stage of its maximum
ing that the nuclear processes that fuel contraction, primordial matter must have
the stars are not endlessly self.perpetu. consisted entirely of free neutrons moving
ating but are accompanied by a gradual much too fast to stick together and form
dissipation of the originally available en· stable nuclei. As the universe started to
ergy. The assumption that an average expand, this primordial gas began to cool.
galaxy loses a mere five per cent of its A.TOMIC WEIGHT
When its temperature dropped to about
luminosity in the course of 500 million ELEMENTS are distributcd In rela­ one billion degrees, neutron condensation
years would bring the age of the universe tive amounts according to their began. The neutrons collected in aggre·
to the two billion.year figure demanded by atomic weight. Author's theory coin­ gates of val·ying numbers of particles. It
other astronomical, geological and physi. cides with clcmcnts observcd in stars . is known that neutron aggregates are in·
381 evidence.
•
trinsically unstable unless about half of
for exampIe, are oxygen. iron and sil icon their particles catTy a positive electric

T
HIS CONCLUSION finds :'trona con· so abundant; and gold, silver and mercury charge. Hence they must have emitted
(
firmation in recent work by J oe Steb. so rare? electron, until they achieved a state of
bins and A. E. Whitford at the Mount We know that. except for the lightest electrical equilibrium. The electrons fell
Wilson Observatory, who have stlldied the elements (such as hydrogen, helium. nitro· into orbits around the nuclei and formed
apparent luminosities of distant galaxies gen and carbon, involved in the sun's nu­ electronic envelopes around them; thub
on special plates sensitive to red light. To clear cycle). transformation of one atomic atoms were created.
everyone's surprise, they found these gal. nucleus into another requires tremendous I shall not attempt here to go into a
axies much brighter in the red part of the temperatures such as do not exist at the detailed description of the rather involved
spectrum than tbey had previously ap· present time even in the hot interiors of mathematical theory of atom·buildino-
o but
peared to be on ordinary photographic the stars. Consequently there can not have shall simply present a graph which om. �
plates, which are sensitive mostly to the been any revolutionary change in the rela· pares the abundance curves of the chemi·
blue rays. It looked at first as if this phe· tive abundance of the various elements cal elements as observed and as calculated
nomenon was due to the same kind of since the expansion of the universe began. by our theory. The theoretical curve cor·
optical scattering which makes the sun On the other hand, there has been some responds pretty well with the observed
look red during dust storms; light from change, for a number of atoms are radio· values; the fluctuations of the empirical
the galaxies, it was thought, was reddened active and have gradually decayed into curve are due to minor periodic variations
by the clouds of fine intergalactic dust more stable elements. of nuclear properties and can be explained
through which it passed. Calculations Considering the latter case first, we by a more detailed form of the theory.
showed, however, that to account for the note, for example. that the lighter isotope According to our calculations, the for·
observed reddening would take a fantastic of uranium, U-235 (atomic bomb stuff). mation of elements must bave started five
quantity of dust-IOO times as mucb as constitutes only .7 per cent of a given minutes after the maximum compression
the total amount of matter in the galaxies amount o( uranium found in nature; the of the universe. It was fully acc'lmplished.
themselves. Sucb an assumption would rest is the heavier isotope U-238. The half· in all essentials. about 10 minutes later.
come into serious conflict witb many facts life of U·235 is only. 7 billion years. while By that lime the density of matter had
and theories about the structure of the that of U·238 is 4·.5 billion years. If we dropped below the minimum necessary
universe. make the reasonable assumptiun that at for nuclear.building processes. All the ele·
It therefore seems more reasonable to the original formation of the univer,e both ments were created in that critical 10
�
tiuppose that the distant galaxies look red· i"otopes were pruduced in about equal minutes, and their relative abundance in
der simply because they actually were amounts. t he age of the universe figure. the universe has remained essentially con·
redder when they emitted the light which up to about four billion years. Similar cal· "tant throughout the two or three billion
is now reaching our telescopes. This could culations based on the naturally radio· vears of subsequent expansion.
be explained if we assumed that young active isutope of potassium (relative -
galaxies contain more red stars than more abundance-.Ol per cent; half·life-.4bil· George Gamow, professor oj physics
mature ones. Further studies by Stebbins linn years) yields the figure of 1.6 billion at George Washington University, is
and Whitford should yield important in· years. While there is a discrepancy, both author oj Birth and Death of the Sun
formation about the evolutionary life of figures agree roughly in order of magni. alld other popular scientific books.

24
© 1948 SCIENTIFIC AMERICAN, INC
FOUR GALAXIES of different types recede from the galaxies have the same average size and intrinsic lumi.
Milky Way in a mighty company. They are N.G.C. 3185, nosity. Groupings of galaxies are not uncommon. The
N.G.C. 3187, N.G.C. 3190 and N.G.C. 3193. The evidence Milky Way galaxy and the Andromeda Nebula, plus
that these aggregates are in the same cluster is that the several smaller aggregates, form a system of their own.

25

© 1948 SCIENTIFIC AMERICAN, INC