286 
would therefore emit much less light than they would 
do later, and stand a poor chance of being seen.*’ 
We know, as yet, very little about the colour-index 
and temperature of stars of those varieties of Class M 
(Mb and Mc) which are evidently furthest along in 
the spectral series, and it may well be that a star 
usually reaches the temperature corresponding to 
these stages by the time that it begins to shine at all 
brightly. In any case, stars in these very early stages 
should be of small or moderate luminosity, and rare 
per unit of volume, and hence very few of them would 
be included in our catalogues. 
The great luminosity and extreme redness of the 
stars of Class N suggest that they belong at the 
beginning of the series of giant stars; but the rela- 
tions of this very distinct spectral type to the others 
are not yet quite clear, and it would be premature to 
give it a definitive place in the sequence. It seems 
clear, however, that these stars must be in a very 
primitive condition, rather than in a very late one, as 
believed by Lockyer. The stars of Class O (Wollf- 
Rayet stars) are of very great average luminosity, 
and probably lie beyond those of Class B at the apex 
of the temperature scale, as Lockyer supposes. But 
in the absence of data concerning their masses, densi- 
ties, and the like, we cannot place them definitively, 
except that Oe5 and Oe come almost certainly just 
above Be. 
One further application of the theory may be very 
briefly mentioned. If we have a large number of 
contracting masses of gas endowed with various 
moments of momentum, more and more of them will 
split up into pairs as they grow denser, and the pairs 
latest formed will have the shortest periods. A large 
percentage of spectroscopic binaries, especially of 
short period, is therefore direct evidence of a fairly 
advanced state of evolution, and the occurrence of 
this condition among the stars of Classes B and A 
supports—indeed, almost by itself compels—the view 
that they are far removed from a primitive condition. 
Most of the stars which have been investigated for 
radial velocity are giants, and the absence of spectro- 
scopic binaries of short period among the redder stars 
is in agreement with the view that they are in earlier 
stages of evolution. 
The distribution of the visual binaries and physical 
pairs among the various spectral classes depends 
mainly upon a quite different factor, namely, the 
resolving power of our telescopes, which allows us to 
separate the closer pairs of short period only among 
the nearer stars, so that the systems for which orbits 
have been determined are nearly all dwarf stars. 
I have endeavoured in the past hour to set before 
you the present state of knowledge concerning the 
real brightness, masses, densities, temperatures, and 
surface brightness of the stars, and to sketch the 
theory of stellar evolution to which the study of these 
things has led me. This theory is inconsistent with 
the generally accepted view. Its fundamental prin- 
ciple is identical with that of Lockyer’s classification, 
but it differs radically from the latter in the principles 
according to which it assigns individual stars, and 
even whole classes of stars, to the series of ascending 
or descending temperature. (For example, Lockyer 
puts such conspicuously giant stars as Canopus. 
Capella, Arcturus, and 8 Cygni, and all the stars of 
Class N, into the descending series, and places 
8 Hydri and 6 Pavonis (which are clearly dwarf stars) 
in the ascending series.) 
Two things have gone farthest to convince me that 
it may be a good approximation to the truth—the 
way in which it explains and coordinates character- 
istics of the different spectral types which previously 
appeared to be without connection or reason, and the 
27 Russe'l, Sczexce, N.S. 
NO, 2224. VOR. 498K 
Vol. xxxvii, p. 646, 1973. 
NATURE 
[May 14, 1914 
way in which a number of apparent exceptions to its 
indications have disappeared, one by one, as more 
accurate information concerning spectra, orbits of 
double stars, and the like, became available, until 
only one doubtful case remains. 
I have purposely made no attempt at this time to 
touch upon certain other interesting matters, such as 
the difference of the mean peculiar velocities of stars 
of the various spectral classes, although, with the aid 
of simple and very reasonable assumptions, they may 
be added to the list of things explainable by the new 
theory. My reason for this has been less for lack 
of time than because there is at present no definite 
reason, assignable from general considerations in 
advance, why we should expect an old star to be 
moving faster or slower than a younger one, while 
there is such a reason why we should suppose that a 
dense star is in a later stage of evolution than one of 
low density. It seems better to find out what we 
can about the order of evolution from data of the 
latter kind, and then apply our results to the study of 
problems of the former sort, than to attack them 
without such aid or by means of unproved assump- 
tions. The assumptions that are necessary on the new 
theory are simple and probable enough, but they do 
not form an integral part of the theory, and cannot 
be established directly from general considerations, 
and so I will not discuss them now. 
The new theory will not explain everything about 
the stars—I should be rather afraid of it if it did; for 
example, it leaves the phenomena of preferential 
motion, or ‘‘star-streaming,’’ as puzzling as ever. I 
have only tried to interpret some of the facts most 
obviously capable of evolutionary explanation, on the 
fundamental assumption that the properties of matter, 
and the forces in operation, among the stars, are the 
same as those with which we are familiar in the 
laboratory. He would be a bold man indeed who 
would assert that this assumption is entirely true, but 
it seems clear that it should be thoroughly tried out 
before the existence of new forces can legitimately 
be postulated. 
If the ideas to which you have so kindly listened 
to-day shall prove of any help toward removing the 
need for belief in unknown forces, and extending the 
domain of those already known, my labour will be 
far more than repaid; but it should not be forgotten 
that the real labourers have been those who, through 
long and weary nights, accumulated bit by bit, and, 
through monotonous days, prepared for the use of 
others the treasures of observational knowledge with 
which it has been my pleasurable lot to play in the 
comfort of my study. 
I need scarcely add that, if what I have said proves 
of interest to any of you, your frank and unsparing 
criticism will be the greatest service which vou can 
render me. 
UNIVERSITY AND EDUCATIONAL 
INTELLIGENCE. 
BIRMINGHAM.—At its last meeting the University 
council passed the following resolution :—‘* That the 
council desires to record its deep sorrow at the death 
of Prof. Poynting, who so faithfully served the Mason 
College and the University for thirty-four years. 
During his distinguished career as professor of physics 
he was not only an inspiring teacher and investigator, 
but bore a considerable part in the development of 
the college and of the University. His keen interest 
in all that concerned the University, its staff, and its 
students, his genial and attractive personality, will 
be long .and affectionately remembered; his death 
leaves a gap which it will be most difficult to fill.” 
i f 
