Constitution and Evolution of the Stars 107
netic fields. The resulting forces deflect atoms of dif-
ferent weights in different directions, and bring each kind
to a separate focus upon a photographic plate, producing
images when the plate is developed. By measuring these
plates the atomic weights may be determined; and Aston
has found that, in every case but one, the true atomic
weights are exact integers, within the accuracy of meas-
urement, which is about one part in a thousand. When
the chemist finds an atomic weight which is not an in-
teger, such as 35.46 for chlorine, this is really the aver-
age for two different kinds of atoms of the same chemi-
cal properties, but different weights, both of which are
integers—35 and 37 in this case. The one exception is
hydrogen, for which the chemist’s determination 1.008
is exactly confirmed.
Now it is more than a century since Prout suggested
that, since the atomic weights are so nearly integers, the
atoms themselves are built up out of simple units. We
now transfer this idea to the atomic nuclei, which con-
tain practically all the mass, and Aston’s beautiful re-
searches practically compel belief. The hydrogen nucleus,
or “proton,” is the lightest of all, and we would naturally
look to it as the fundamental unit. Rutherford’s success
in knocking the nuclei of elements such as oxygen, nitro-
gen and sodium to bits, by collision with fast moving
alpha particles, has furnished a definite proof that pro-
tons, and alpha particles as well, are actual constituents
of these nuclei. Many nuclei must also contain electrons,
which prevent the net electric charge from getting too
high. It looks, for example, as if an alpha particle was built
of four protons and two electrons, held together by forces
of whose nature we are ignorant. This would give ex-
actly the right electric charge; but the mass of the four