Supplement to “ Nature,” May 12, 1923 ix 
ATOMS AND ELECTRONS. 
We have hitherto pictured the inside of a star as 
a hurly-burly of atoms and ether-waves. We must 
now introduce a third population to join in the dance. 
There are vast numbers of free electrons—unattached 
units of negative electricity. More numerous than 
the atoms, the electrons dash about with a hundred- 
fold higher velocity—corresponding to their small 
mass, which is only 1/1850 of a hydrogen atom. 
These electrons have come out of the atoms, having 
broken loose at the high temperature here involved. 
An atom has been compared to a miniature solar 
system ; a composite central nucleus carrying positive 
charge corresponds to the sun, and round it revolve 
in circular and elliptic orbits a number of negative 
electrons at comparatively large distances corre- 
sponding to the planets. We know the number of 
satellite electrons for each element ; sodium has 11, 
iron 26, tin 50, uranium 92. Our own solar system 
with 8 revolving planets represents an atom of oxygen. 
The thermodynamical theory due mainly to Nernst 
permits us to calculate roughly how many of these 
break loose under given conditions of temperature 
and density ; and in a typical star a large proportion 
of them must have become free. 
This condition solves for us our chief difficulty as 
to the molecular weight of stellar material. We need 
to know it in order to perform our calculations as to 
the state of the star; and at first sight it might seem 
hopeless to arrive at the molecular weight without 
knowing the elements which constitute the bulk of 
the material. But suppose first that the temperature 
is so high that all the satellite electrons have broken 
away. Anatom of sodium will have separated into 12 
particles, namely, 11 electrons and rx mutilated atom ; 
its atomic weight 23 is divided between 12 independent 
particles, so that the average weight of each is 
23/12=1-92. Next take iron: the atomic weight 56 
is divided between 27 particles; average 2:07. For 
tin we have 119 divided by 51; average 2°34. For 
uranium, 238 divided by 93; average 2:56. It 
scarcely matters what element we take ; the average 
weight of the ultimate particles (which is what we 
mean by the molecular weight) is always somewhere 
about 2. If only the stars were a bit hotter than they 
actually are, it would make our task very easy. Un- 
fortunately, they are not hot enough to give complete 
separation, and the actual degree of separation will 
depend on the temperature of the star, thus intro- 
ducing a difficult complication. Generally at least 
half the electrons are detached and the molecular 
weight must be taken as between 3 and 4. I hope 
that the theory of this dissociation of electrons will be 
improved, because at present it is the chief bar to 
rapid progress with the theory of stellar constitution. 
It is a great help to know that the molecular weight 
is between 3 and 4; but we have reached a stage 
when it is becoming necessary for progress to know 
it for each star within much closer limits. 
BRIGHTNESS AND Mass. 
We pictured a physicist on a cloud-bound planet 
who was able from laboratory data to predict how 
large would be the masses into which the material 
of the universe must aggregate. Let us now set him 
a harder task. We inform him that we have observed 
these masses of gas, and, choosing one equal, say, to 
his 34th sphere, we ask him to predict how brightly 
it will shine. I have already mentioned that the star 
keeps practically the same brightness so long as it is 
a perfect gas ascending in temperature, so it should 
not be necessary to give the physicist any data except 
the precise mass. ‘To use the same plan as before, we 
imagine a series of lamps of ro candle-power, 100 
candle-power, tooo, and so on; and his task is to 
pick out which lamp in this series corresponds ap- 
proximately to the star. I believe that it is now 
possible for him to perform this task and to pick out 
(correctly) the 31st lamp. But for this purpose it is 
not enough that he should know all about the heat 
stored in the interior of the star; the brightness of 
the star depends on the rate at which the ether- 
waves are leaking out, and that introduces a new 
subject—the obstructive power of the material atoms 
which dam back the radiant flow. 
Another name for this obstructive power is opacity. 
A substance which strongly obstructs the passage of 
light and heat waves is said to be opaque. The rising 
temperature towards the centre of the star urges the 
heat to flow outwards to the lower temperature level ; 
the opacity of the material hinders this flow. The 
struggle between these two factors decides how much 
light and heat will flow out. We have calculated the 
internal temperature-distribution, so that we know all 
about the first factor; if then we can observe the 
outward flow which occurs, that should settle the value 
of the second factor—the opacity. The outward flow 
is capable of observation because it constitutes the 
heat and light sent to us by the star. 
One of the troubles of astronomy is that our in- 
formation about the stars is so scattered. We know 
the mass of one star very accurately, but we do not 
know its absolute brightness ; we know the brightness 
of another but not its mass ; for a third we may have 
an accurate knowledge of the density but nothing else. 
For Sirius, Procyon, and a Centauri our knowledge is 
fairly complete and accurate ; but not any of these are 
