228 NATURE [Fuly 8, 1886 
Dr. Huggins of course only include the violet and the ultra- | its hottest portion, and there we have the spectra such as Dr, 
violet. 
This is the place to tell you that in the eclipse of 1882 most 
of these lines which you see in the spectrum of Sirius and a 
Lyrze—lines which are entirely cloaked in the ordinary spectrum 
of the one, and in most of the other stars—was actually photo- 
graphed in the hottest part of the sun’s atmosphere during that 
eclipse ; so that you see that there were two prophecies with 
rd to this set of lines, both of which were fulfilled. 
Now, when in science a working hypothesis suggests a certain 
result under certain conditions—enables us to prophecy in fact, 
and the prophecy comes true (that is essential)—we have a right 
to believe that the hypothesis may be well founded. 
Those, therefore, who hold that these differences are due to 
temperature, have considered their opinions to be considerably 
fortified by those two fulfilments of prophecy to which I have 
referred. 
You see in the diagram that although in the upper 
representing Sirius and. a Lyre the lines are very thick, as those 
erientae ines thin out other lines come in; so that in passing 
down the diagram from the upper horizon to the lower one, we 
get two conditions of things—one which we leave when we get 
ew lines thick, and another which we reach where we get a very 
considerable number of lines thin, as many people believe be- 
these substances gradually, by reduction of temperature in 
the atmosphere of the star, have gone into combination with 
themselves or something else, and formed other more complex 
nodies, which give us of course new lines practically at the ex- 
sense of the old ones. There then we get on that diagram 
representing a part of Dr. Huggins’s magnificent work, the pos- 
sible explanation of the passage from Class I. to Class II. 
A reference to the two classes of absorption-spectra we owe 
chiefly to the work of Dunér and Vogel may very fitly follow 
e records of Dr. Huggins’s work. 
re 
spectra 
cause 
CARBON 
Fic. 23.—Duneér's spectra of Class 111. compared with carbon 
I did not tell you (there was no necessity for it at the time) 
that one of the rarest substances apparently in the atmosphere of 
the sun at the present time is carbon. There is a possible trace 
of carbon in the sun’s atmosphere I think, but the assertion 
depends upon the existence of a single fluting, so far 
tions go, in the ultra-violet part of the spectrum. 
Now in these two classes of stars, Class IIIa. and Class III. 
of Vogel, which I have called Class III. and Class IV. to 
make things easier, in the former we get a spectrum which at 
present has not been investigated ; and we cannot say to what 
substances the absorptions in stars (there are hundreds of them 
mind you) which give you that spectrum is due. 
But with regard to the stars of the latter class (and there are 
hundreds of them) they give us absorptions about which there is 
no question whatever. There the light of the star, instead of 
being absorbed by iron vapour, by hydrogen vapour, by calcium 
vapour, and by nickel, and by cobalt vapours and the like, as in 
the atmosphere of our sun, is absorbed by carbon vapour, and 
carbon vapour almost pure and simple, for when you have taken 
these big flutings out of the spectrum, there is little left—which 
means that when you get rid of the absorption of carbon in the 
atmosphere of those stars there is very little absorption left. I 
mean that the remaining spectrum is very simple. 
I have already pointe d out that it is fair to say that if our sun 
were hotter its absorption would more resemble the spectrum of 
as observa- 
Huggins photographed for us in Class I. and Class II. I have 
just told you that there we really do get the same lines as in our 
sun during an eclipse, when we can best get at the spectrum of 
the hottest portion. 
Are we to suppose then that if our sun was very much cooler 
than it is we should get the spectrum of carbon developed in this 
enormous way? That is a question which at present it is not 
possible to answer. It is quite probable, but then if that be so, 
you will see two things : first, that the carbon, 7/7 exists as such 
in such enormous quantities in the sun at the present moment, 
must be so far outside the region of high temperature that it can- 
not absorb in that manner. The second point is'that it cannot 
be that particular substance which gives us the continuous spec- 
trum in the photosphere, because if it were we should be certain, 
I think, to get more indications of bright carbon vapour, both in 
the spectrum of sunspots and in the spectra of prominences than 
we do, so that although these stellar spectra may set us thinking 
about the sun, they are rather more important to us at the present 
moment for telling us what possibly is not rather than for what 
probably is. 
With regard to the stars with bright lines, the only point that 
I need make about them now is that it is most important that 
every endeavour should be made to determine the origin of those 
bright lines which, as I have already pointed out to you, are not 
coincident with the lines of hydrogen. But I would rather say 
what I have to say on that subject in connection with the next 
part of the comparison which we are making. 
In the spectra which have already been indicated to you nothing 
has been said about change of star-light except at long periods. 
It has been hinted that possibly a star which at one stage gives 
you a spectrum of very thick lines, may at other stages undergo 
changes which will make it a star of the second class, in which we 
have a greater number of thinner lines and so on. 
Here the question of stellar evolution is suggested. On this 
subject I cannot enter, but a few general remarks may be made. 
We may say that we now know that comets are clouds of stones, 
and experiments, to which I will refer again later on, have been 
made which suggest that if nebulz are of like nature the differ- 
ences between cometary and nebular spectra may be explained 
by differences of temperature, that of the nebula being higher 
than that of the comet. Now comets ordinarily, 7.e. when 
coolest, give us the spectrum of carbon, but when the tempera- 
ture is increased, as it was in the case of the comet of 1882, 
sodium and iron are added. Imagine a comet with a nucleus 
the light of which is absorbed by ordinary cometary vapours, 
and we shall have the spectrum of a star of the fourth class. 
On the nebular hypothesis, supposing, as seen above, that we 
started with ordinary cometary materials, then, on the beginning 
of a central condensation which in time is to become a star, as 
Kant and Laplace suggested, such central condensation should 
then give us a star of the fourth class. As the energy of condensa- 
tion increased and the temperature got higher, the spectra would 
change through the third and second classes, till ultimately, when 
the temperature was highest, the first class spectrum would be 
reached. On the slackening down of the temperature of the now 
formed star, the spectra of the second, third, and fourth classes 
would then be reproduced, but, of course, now in. the direct 
order. 
Nothing so far has been said-about changes which instead of 
taking millions and perhaps billions of years can be undergone 
in a few days, or weeks, or months. 
Careful observations in the heavens have shown a great many 
years ago that a large number of stars are subject to a consider- 
able change in their brilliancy. ‘The most important work which 
has been “recently done in what we may call the philosophy of 
variable stars we owe to the diligence of Prof. Pickering of 
Harvard Observatory in the United States. He has proposed 
a classification of variable stars, and for the purpose of this lecture 
I have just to make a slight alteration in his classification, as I 
did in the other one. First I will tell you broadly what the 
classification is, and then I will describe as briefly as may be 
some of the more important details which are of the greatest 
moment and interest. There is a whole mine of interest here 
which of course I cannot touch in the time at my disposal. 
The first class of these variable stars then may fitly form 
what are called temporary or new stars—stars which as we 
have had evidence during the last few years—in 1866 and 1876, 
and at the end of last year, suddenly burst into visibility in the 
heavens as if they were new creations, last for a certain time, 
