031

INORGANIC EVOLUTION AS STUDIED BY
SPECTRUM ANALYSIS.

LIST OF WORKS BY SIR NORMAN
LOCKYER.

CONTRIBUTIONS TO SOLAR PHYSICS.
CHEMISTRY OF THE SUN.
THE METEORITIC HYPOTHESIS.
THE SUN'S PLACE IX NATURE.
RECENT AND COMING ECLIPSES.
INORGANIC EVOLUTION.
STARGAZING, PAST AND PRESENT.

(In conjunction with G. M. Scabroke.}

ELEMENTARY LESSONS IN ASTRONOMY
PRIMER OF ASTRONOMY.
THE DAWN OP ASTRONOMY.
MOVEMENTS OF THE EARTH.

STUDIES IN SPECTRUM ANALYSIS.

THE SPECTROSCOPE AND ITS APPLICATIONS.

THE RULES OF GOLF.

(In conjunction with jr. Rutherford.)

INORGANIC EVOLUTION

AS STUDIED BY

SPECTRUM ANALYSIS.

BY

SIR NORMAN LOCKYER, K.C.B., F.R.S.

Correspondent or THE INSTITUTE OF FRANCE ; THE SOCIETY FOR THE PROMOTION

OF NATIONAL INDUSTRY OF FRANCE ; THE ROYAL ACADEMY , OF SCIENCE,

GOTTINGEN ; THE FRANKLIN INSTITUTE, PHILADELPHIA ; THE ROYAL

MEDICAL SOCIETY OF BRUSSELS ; SOCIETY OF ITALIAN SPECTROSCOPISTS ;

THE ROYAL ACADEMY OF PALERMO ; THE NATURAL HISTORY SOCIETY OF

GENEVA ; Member OF THE ROYAL ACADEMY OF LYNCEI, ROME ; AND

THE AMERICAN PHILOSOPHICAL SOCIETY, PHILADELPHIA; Honorary

Member OF THE ACADEMY OF NATURAL SCIENCE OF CANTANIA;

PHILOSOPHICAL SOCIETY OF YORK ; LITERARY AND PHILO-
SOPHICAL SOCIETY OF MANCHESTER ; AND LEHIGH UNIVERSITY ;
Director OF THE SOLAR PHYSICS OBSERVATORY ; AND

Professor OF ASTRONOMICAL PHYSICS IN THE ROYAL

COLLEGE OF SCIENCE, LONDON.

llon&on

MACMILLAN AND CO., LIMITED
1900

[All Rights Reserved.]

LOAN STACK

- 1 1 z.r

LONDON :

HARBISON AND SONS, PRINTERS IN ORDINARY TO HER MAJESTY,
ST. "M PUTIN'S LANE.

LSI

PREFACE.

THIS present volume contains an account of my most recent
inquiries into the chemistry of the stars, and of some questions
which have grown out of these inquiries. It has taken its
present form because several friends, upon whose judgment I
can rely, suggested that I should preface the account of the
work, and the conclusions I have derived from it, by a statement,
as clear and simple as I could make it, of the principles of
Spectrum Analysis and of the earlier steps in the various investi-
gations the convergence of which has led to the present stand-
point.

In my " Chemistry of the Sun," published in 1887, I dealt
chiefly with the then state of the problem, so far as the Sun was
concerned. In two later volumes, " The Meteoritic Hypothesis "
and the "Sun's Place in Nature," I included the stars in the
survey. The short story which I give in the earlier portion of
the present book consists of a resume of the three volumes, so
far as the question of dissociation is concerned ; this is followed
by evidence recently accumulated by other inquirers, all of which
tends to strengthen my original thesis. In the latter part of the
volume I endeavour to show how, in the studies concerning dis-
sociation, we have really been collecting facts concerning the
evolution of the chemical elements ; and I point out especially that
the first steps in this evolution may possibly be best studied by,
and most clearly represented in, the long chain of facts now at our
disposal touching the spectral changes observed in the hottest
stars.

My thanks are due (1) to Messrs. Lockyer, Fowler and Baxan-

205 2

vi PEEFACE.

dall, and other assistants at South Kensington, who have helped
me to do the work ; (2) to my colleagues, Professors Perry,
Howes and Farmer, and Professor Poulton and Dr. Woodward,
who have so freely given me information 011 several of the points
touched upon in the later chapters ; (3) to Professor Kayser, Sir
William Crookes, .Professor A. Schuster, and Dr. Preston, who
have been good enough to look over the portions referring to their
researches ; and (4) to the officers and council of the Royal Society
and to Messrs. Macmillan, for placing at my disposal some of- the
illustrations.

NOBMAN LOCKYER

Solar Physics Observatory,
South Kensington,
January 9, 1900.

CONTENTS.

BOOK I.— THE BASIS OF THE INQUIRY.

CHAP. PAGE

I. Principles and Methods ... ... ... ... ... 1

II. Some Pioneering Difficulties ... ... ... ... 18

III. The Present Position 29

BOOK IL— APPLICATION OF THE INQUIEY TO
THE SUN AND STARS.

IV. The Sun's Chromosphere 38

V. Stellar Atmospheres ... ... ... ... ... 45

VI. The Chemistry of the Stars ... 55

VII. A Chemical Classification of Stars ... 66

BOOK III.— THE DISSOCIATION HYPOTHESIS.

VIII. Recent Opinion 73

IX. Stellar Evidence 78

X. " Series " Evidence 83

XI. "Shifting "of Lines Evidence 101

XII. Magnetic Perturbation Evidence ... ... ... 109

XIII. Fractionation Evidence . 116

BOOK IV.— OBJECTIONS TO THE DISSOCIATION
HYPOTHESIS.

XIV. The Chemistry of Space 120

XV. The General Distribution of Stars 124

XVI. The Distribution of Chemical Groups of Stars ... 129

XVII. The Result of the Inquiry ... ... ... ... 141

XVIII. Replies to Special Objections ... ... ... ... 144

Vlll CONTEXTS.

BOOK V.— INORGANIC EVOLUTION.

CHAP. PAGE

XIX. What Evolution means : Organic Evolution ... ... 152

XX. The Stellar Evidence regarding Inorganic Evolution ... 157

XXI. The Simplest Elements appear first ... ... ... 162

XXII. The Relations of the Organic and Inorganic Evolutions 168

XXIII. Inorganic Evolution from a Chemical Standpoint ... 175

XXIV. Inorganic Evolution from a Physical Standpoint ... 184

ILLUSTRATIONS.

FIG. PAGE

1. Arrangement of candle, prism and eye 2

3. Use of the simple spectroscope ... ... ... ... 3

3. A continuous and a discontinuous spectrum ... ... ... G

4. Use of a circular and line slit ... ... ... ... ... 7

5. Observation of a flame with ordinary spectroscope ... ... 7

6. Steinheil spectroscope ... ... ... ... ... ... 8

7. Angstrom's grating spectrometer ... ... ... ... 9

8. Spectra of barium and iron contrasted ... ... ... 9

9. Fluting of carbon 10

1 0. Fluting of magnesium ... ... ... ... ... ... 10

1 1 . The " series " of cleveite gases ... ... ... ... 11

12. Fraunhofer's solar spectrum ... ... ... ... ... 15

1 3. A stellar spectroscope ... ... ... ... ... ... 16

14. Spectroscope attached to a telescope for sun observations ... 21

15. First method of spectroscopic work 21

16. New method of using a lens 22

17. Long and short lines of strontium and calcium ... ... 23

18. Long and short lines of sodium ... ... ... ... 24

19. Spectrum of a sun-spot... .... ... ... ... 25

20. Spot and prominence lines contrasted... ... ... ... 26

21. Different rates of motion registered by different iron lines ... 27

22. Comparison of chromospheric and test spectra 40

23. Spectrum of chromosphere contrasted with Fraunhofer lines 40

24. Temperature curve ... ... ... ... ... ... 47

25. Comparison of the spectrum of a Cygni with enhanced

metallic lines 49

26. Comparison of the spectrum of a Cygni and chromosphere ... 53

27. Map of chemical substances present in stars of different tem-

peratures ... ... ... ... ... ... ... 62

28. Comparison of the spectra of Sirius and a Cygni ... ... 69

29. Comparison of the spectra of Procyon and y Cygni ... .. 69

30. Simple flutings of nitrogen 84

X ILLUSTRATIONS.

TIG. PAGE

.31. Spectrum of the cleveite gases ... ... ... .. 84

32. Same spectrum sorted out into " series " ... ... ... 85

33. Triplets in spectra 92

34. Map showing residual lines in the spectra of calcium and

magnesium 96

35. Changes of wave-length due to pressure 105

36. Results of magnetic perturbations on spectrallines .. ... Ill

37. A f ractionation diagram ... ... ... ... ... 118

38. Comparison of relative numbers of stars 130

39. Distribution of bright-line stars in the Milky Way 133

40. Photograph of glass globe showing Milky AVay 134

41. Photograph of the same showing distribution of bright-line

stars 135

42. Spectrum of Nova Aurigse ... ... ... ... ... 138

43. Diagram showing time and temperature relations of organic

and inorganic evolution ... ... ... ... ... 173

INORGANIC EVOLUTION AS STUDIED BY
SPECTRUM ANALYSIS.

BOOK L— THE BASIS OF THE INQUIRY.

CHAP. I. — PRINCIPLES AND METHODS.

THE thirty years' work to which I have to refer in this book has had
to do with various points raised by the investigation of the radiation
and absorption of light ; the science of spectrum analysis is involved.

Spectrum analysis, indeed, is now becoming so far-reaching, espe-
cially in inquiries having to do with the conditions of the various
celestial bodies, that there are many who are anxious to know some-
thing of its teachings. To some of these, however, the terms used by
men of science, a very necessary shorthand, are unfamiliar, and appear
hard to understand, because the opportunity of seeing the things they
are intended to define, and which they generally do define in most
admirable fashion, has never presented itself. I propose, therefore, to
attempt to show that there is nothing recondite about these terms ; that
it is possible without any expensive apparatus for every one who will
take a little trouble, to observe the phenomena for himself, after which
the meanings of the terms employed will present no difficulty whatever.

One key to the hieroglyphics, the light story, which is hidden in
every ray of light, is supplied to us by the rainbow. It teaches us that
the white light with which nature bountifully supplies us in the sun's
rays is composed of rays of different kinds or of different colours ; and
it is common knowledge that there is an almost perfect analogy
between these coloured lights and sounds of different pitches.

The blue of the rainbow may be likened to the higher notes of the
key-board of a piano, and the red of the rainbow, on the. other hand,
may be likened to the longer sound waves which produce the lower
notes ; and as we are able in the language of music to define each
particular note, such as B flat and G sharp, and so on, so light- waves
are defined by their colours or wave-lengths.

2 INORGANIC EVOLUTION. [CHAP.

What nature accomplishes by a rain-drop we can do with a prism
or a grating. A prism is a piece of glass or other transparent mate-
rial through which the light is bent out of its course or refracted in the
process. A grating is a collection of wires, or scratches on glass or
metal, equidistant, very near together, and all parallel. When light
passes through, or is reflected by such a system, it is said to be diffracted,
and one result that we are concerned in is very similar to that of
passing light through a prism.

It is rapidly becoming a familiar fact to many that when a ray of
white light is refracted by a prism or diffracted by a grating a band of
colour similar to a rainbow is produced, and that this effect follows
because white light is built up of light of every colour, each colour
having its own special length of wave and degree of refrangibility.
Our rainbow band is called a spectrum.

Such a glass prism or grating is the fundamental part of the instru-
ment called the spectroscope, and the most complicated spectroscope
which we can imagine simply utilises the part which the prism or
grating plays in breaking up a beam of white light into its constituent
parts from the red to the violet. Between these colours we get that
string of orange, yellow, green, and blue which we are familiar with in
the rainbow.

A Simple Spectroscope.

For sixpence any of us may make for ourselves an instrument which
will serve many of the purposes of demonstrating some of the mar-
vellously fertile fields of knowledge which have been recently opened
up to us. From an optician we can buy a small prism for sixpence ; get
a piece of wood from 20 to 10 inches long (the distance of distinct
vision), 1 inch broad, and J an inch thick. On one end glue a cork

Prism
Candle

FiG. 1.-- Arrangement of candle, prisrn and eye.

2 inches high, at the other end fasten, by melting the bottom, a stump
of a wax candle of such a height that the dark cone above the

I.] PRINCIPLES AND METHODS. 3

wick is level with the top of the cork. Then glue the prism on the
cork, so that by looking sideways through the prism the coloured
image or spectrum of the flame of the candle placed at the other end
of the piece of wood can be seen.

We get a band of colour, a spectrum of the candle flame built up
of an infinite number of images of the flame produced by the light
rays of every colour. But, so far, the spectrum is impure, because the
images overlap. We can get rid of this defect by replacing the candle
by a needle.

If we now allow the needle to reflect the light of the candle flame,
taking care that the direct light from the candle does not fall upon the
face of the prism, we then get a much purer band of colour, because
now we have an innumerable multitude of images of the thin needle
instead of the broad flame close together. The needle is the equi-
valent of the slit of the more complicated spectroscopes used in
laboratories.

FIG. 2. — Use of the simple spectroscope. •

We can vary this experiment by gumming two pieces of tin oil
with two perfectly straight edges on a piece of glass so that the straight
edges are parallel and very near together. In this way we have a
slit ; this should be fixed close to the candle and between it and the
prism.

Now the light of the candle is white, and the preceding experiment
tells us that such light gives us a band containing all the colours

B 2

4 INORGANIC EVOLUTION. [CHAP.

without any breaks or gaps. We have what is called a continuous

The Continuous Spectrum.

If we burn a piece of paper, or a match, or ordinary coal-gas, we get
a white light identical to that given us by the candle ; solids which do
not liquefy when made white-hot, and liquids which do not volatilise under
the same condition, and some dense gases when heated, do the same.

This effect is produced because there is light of every wave-length
to produce an image of the needle (or the slit) ; these images blend
together continuously from one end of the spectrum to the other.

Let us then consider this fact established, namely, that solid or
liquid bodies and dense gases, when heated to a vivid incandescence,
give a continuous spectrum. Under these circumstances the light to
the eye, without the spectroscope, will be white, like that of a candle or
white-hot poker.

The Length of the Continuous Spectrum -varies with Temperature.

If we put a poker in a fire, it becomes red-hot ; if we heat a platinum
wire by passing a feeble current of electricity along it, it becomes red-
hot like the poker.

In both cases examination by means of the prism shows that the
red end only of the spectrum is visible. But if the poker or wire be
gradually heated more strongly, the yellow, green, and blue rays will
successively appear. Finally, when a brilliant white heat has been
attained, the whole of the colours of the spectrum will be present.

Hence we learn that *if the degree of incandescence be not high, the
light will only be red. But, so far as the spectrum goes — and it will
expand towards the violet !as the 'incandescence increases, as before
stated — it will be continuous. ' >%

The red condition comes from the absence of blue light; the white
condition comes from the gradual addition of blue as the temperature
increases.

One of the laws formulated by Kirchhoff in the infancy of spectro-
scopic inquiry has to do with the kind of radiation given out by bodies
at different temperatures. The law affirms that the hotter a< mass of
matter is the further its spectrum extends into the ultra-violet.

. Gaslight is redder than the light of an incandescent lamp because
the latter is hotter. The carbons in a so-called arc-lamp give out a
bluish- white light because they are hotter still.

By similar reasoning from experiment we are bound to consider

I.] PRINCIPLES AND METHODS. 5

the bluish- white stars, the white stars, the yellow, red and blood-red
stars to indicate a decreasing order of temperature.*

We shall not go far wrong in supposing that the star with the most
intense continuous radiation in the ultra-violet is the hottest, inde-
pendently of absorbing conditions, which, in the absence of evidence to
the contrary, we must assume to follow the same law in all.

An inquiry into the facts placed at our disposal by stellar photo-
graphs, shows that there is a considerable variation in the distance to
which the radiation extends in the ultra-violet, and that the stars can
be arranged in order of temperature on this basis.

Judged by this criterion alone, some of the hottest stars so tar ob-
served are 7 Orionis, fOrionis, aVirginis, y Pegasi, ?? Ursse Majoris,
and A. Tauri. Of stars of lower, but riot much lower, temperature than
the above, may be named Kigel, f Tauri, aAndromedae, /3Persei,
a Pegasi, and j3 Tauri.

In this way spectrum analysis helps us with regard to temperatures,
both on the earth and in the heavens.

Discontinuous Spectra with Bright Lines.

Let us next pass from a solid which retains its incandescence like
platinum wire without melting, or a liquid which retains its incan-
descence, like molten iron, without volatilising and see what happens.
We have found that when the light entering the slit consists of every
colour and every tone, we have a continuous band of colour. If there is
any defect in the light we must have a discontinuous one, for the reason
that an image of the slit cannot be produced in any particular part of
the spectrum if there be no light of that particular colour to produce it.

There are many artificial flames which are coloured, and if their
light is analysed in the same way as the light of the candle, a perfectly
new set of phenomena present themselves.

Let us again make use of our improvised spectroscope, and allow
the needle to be illuminated by the flame of a spirit lamp into which
salt is gradually allowed to fall • we see at once why the flame is
yellow. It contains no red, green, blue, or violet rays, so that we
should not represent the spectrum by

* On this point 1 wrote as follows in 1892 : " An erroneous idea with regard to
the indications of the temperature of the stars has been held by those who have
not considered the matter specially. It has been imagined that the presence of the
series of hydrogen lines in the ultra-violet was of itself sufficient evidence of a very
high temperature. The experiments of Cornu, however, have shown that the
complete series of lines can be seen with an ordinary spark without jar. Hence
the high temperature of such a star as Sirius is not indicated by the fact that its
spectrum shows the whole series of hydrogen lines, but by the fact that there is
bright continuous radiation far in the ultra-violet.

6 INORGANIC EVOLUTION.

as in the case of the candle, but simply by

[CHAP.

We see one image of the needle coloured in yellow.

We have passed from the spectrum of polychromatic to that of
monochromatic light — from white light to coloured light — from light of
all wave-lengths to light of one wave-length ; from an infinite number
of slit images giving a continuous band of every colour, to one image
of the slit produced by light of one refrangibility, the colour of the
image depending upon the refrangibility. What we shall see in pass-
ing from the spectrum of the candle to that of sodium vapour in the
spirit lamp is shown in the accompanying woodcut.

Candle-flame spectrum.

Straight slit.^j g . it

>lamp flame
King slit. J wif

FIG. 3. — A continuous and a discontinuous spectrum.

That we are truly dealing with an image of the needle (or a slit)
can be proved by using a slit of any shape. This can be shown by
slightly altering our needle experiment. Take a piece of glass and a
piece of tin-foil 1J inches square, cut out of the centre of the tin-foil a
disc slightly larger than a threepenny-piece, and gum the remainder on
the glass. In the centre, where the disc has been cut away, gum a
threepenny-piece. The interval between the threepenny-piece and the
tin-foil constitutes a circular slit. Let it replace the needle, and ex-
amine the flame of the spirit lamp charged with salt through it with
the prism as before.

It will readily be grasped, from what has been stated, that in the
case of coloured flames, the light passing through the spectroscope
being only red, or yellow, or green, as the case may be, will go to build
up an image of the slit in the appropriate part of the spectrum, and

I.]

PRINCIPLES AND METHODS.

that the image thus built up will take the form of a line or circle,
according to the slit we use.

Many chemical substances, salts of various metals, become lumin-
ous by inserting them into flames, as we have treated common salt
(chloride of sodium). With each metal the colour imparted to the
flame is different. The resulting spectrum is called a discontinuous
spectrum, because it is only here and there that images of the slit are
produced ; because some coloured rays, and not all, are present.

FIG. 4. — The spectrum of a complicated light-source as seen with a circular
and a line slit.

The usual laboratory arrangement for^ observing the spectra of
flames, is shown in the woodcut (Fig. 5).

Further, the system of images of the needle (or slit) varies for each
substance, and it is on this ground that the term spectrum analysis is
used, because we can in this way recognise the various substances in
the flame.

Fia. 5. — Observation of a flame spectrum -with ordinary spectroscope with com-
parison prism, a, prism ; b, collimator ; d, slit ; e e, flames to be compared ;
/", observing telescope ; g, scale illuminated by h and reflected by the second
surface of the prism into the telescope.

S INORGANIC EVOLUTION. [CHAP.

But we are not limited to flame temperatures ; substances in a state
of gas or vapour may be made to glow by electricity. At these higher
temperatures very complicated spectra are produced, and again the
spectrum is special to each chemical substance experimented on ; the
images of the needle (or slit), occupying different positions along the
spectrum according to the nature of the source of light.

Fig. 5 gives us a laboratory prism spectroscope of small disper-
sion ; with the more complicated spectra the phenomena are often
better seen if more than one prism is employed. Fig. 6 shows an
instrument in which four prisms are used.

FIG. 6.— Steinheil spectroscope with four prisms.

It is in the case of the more complicated spectra that the wave-
length has to be specially considered from the point of view of denn-
ing the position of a line. It is not enough to say, as was said in the
case of the sodium line, that it is located in the orange.

The lengths of the various light-waves are very small. The wave-
length of the sound-wave of the middle C of a piano is about 4 feet,
while the wave-length of yellow light as defined by that of a line very
accurately measured is -0005895 of a millimetre, that is 5895 ten-
millionths of a millimetre ; so that there are 43,088 waves in a British
inch. The unit of wave-length usually employed is the ten-millionth
of a millimetre. These wave-lengths get shorter as we pass from the
red to the violet.

For accurate measures of the wave-lengths of the lines a grating
is employed as shown in Fig. 7.

I.J PRINCIPLES AND METHODS 9

So much then in general for the radiations given out by light
sources, and the manner in which the spectroscope shows them and the
student records their positions.

Spectrum analysis was established when experiment proved that no
two substances which give a line spectrum give the same order of lines
from one end of the spectrum to the other ; in other words, the line
spectrum of each chemical substance differs from that given by any
other.

Here then is one of the secrets of the new power of investigation
of which the spectroscope has put us in possession : we can recognise

FIG. 7. — Angstrom's grating spectrometer.

each element by its spectrum, whether that spectrum is produced in
the laboratory or is given by light travelling earthwards from the most
distant star, provided the element exists both here and there.

It is in this way that spectrum analysis helps us with regard to

FIG. 8.— Parts of the spectra of (A) barium and (B) iron (from a
photograph).

10

INORGANIC EVOLUTION.

[CHAP.

chemistry ; the spectrum varies according to the chemical substance
which produces it in a manner that will be gathered from an inspec-
tion of the photograph (Fig. 8) which shows the difference between the
spectrum of barium (A), and that of iron (B).

Flutings.

The earliest spectroscope observations revealed the fact that in some
spectra the lines, instead of being irregularly distributed along the
spectrum, were arranged in an easily seen rhythmic fashion. Such
allocations of lines are called flutings, as a succession of them gives
rise to an appearance strongly recalling the flutings of a Corinthian
column seen under a strong side light.

FlG. 9. — Fluting of carbon.

Our improvised spectroscope helps us here too ; use the candle and
straight slit in front of it as before, but shorten the slit, and only
allow the blue light from the base of the candle flame to pass through
it to the prism. We see two or three sets of flutings. These are the
flutings of carbon; they are amongst the most beautiful examples
known and are thoroughly typical.

FIG. 10.— Flutins: Of magnesium.

Series.

One of the most important discoveries made in recent years, teaches
us that in the case of many chemical elements, the apparently irregular
distribution of the lines is really dominated by a most beautiful law,
and that the most exquisite orderly rhythm can be obtained by sorting

I.] PRINCIPLES AND METHODS. 11

out the lines into what are termed " series," that is lines numerically
related to each other.

Messrs. Runge and Paschen* showed, in 1890, that the spectra of
lithium, sodium, and potassium were the summation of the spectra of
various " series." Later they have shown that the same is true in the
case of the cleveite gases.

Violet. Red

l"iG."ll.— The series in the cleveite gases.

A " series " of spectral lines may be defined as a sequence of lines,
the intensity of which decreases with the wave-length, and the wave-
number or wave-frequency of which may be determined by the
formula

A + BM2 + G//i4,

where n represents the integers from three upwards, and the constants
A, B, and C are determined for each element separately. The shorter
the wave-lengths the greater number of waves there will be in a given
length ; hence the wave-frequency varies inversely as the wave-length.

The fact that lines must close up to one another, as the violet end
of the spectrum is reached, indicates that the character of a " series "
is best brought under notice in the ultra-violet end of the spectrum.
In the visible part of the spectrum the lines forming " series " are too
far apart to be recognised as belonging to a series.

The accompanying diagram (Fig. 11) shows how the apparently
irregular lines observed in the spectra of the cleveite gases can be
arranged into the most exquisite order when the six series of lines which
build up the spectra are shown separately.

Some of these series are composed of triplets and some of doublets
instead of single lines.

* AWi. k. AJcad. Wist., Berlin, 1890.

12 . INORGANIC EVOLUTION. [CHAP.

I wrote thus on this subject in 1879 : —

" I am at present engaged in investigating this question of rhythm,
and I have already found that many of the first order lines of iron
may probably arise from the superposition or integration of a number
of rhythmical triplets. All this goes to show how long the series of
simplifications is that we bring about in the case of the so-called ele-
mentary bodies by the application of a temperature that we cannot as
yet define.

" Indeed, the more one studies spectra in detail, and especially
under varying conditions of temperature which enable us to observe
the reversal now of this set of lines, now of that, the more complex
becomes the possible origin. Some spectra are full of doublets ; others
again are full of triplets, the wider member being sometimes on the
more, sometimes on the less, refrangible side."*

Mascartf had noted this recurrence of similar features in spectra
ten years earlier.

Discontinuous Spectra with Dark Lines.

It is time now to make still another experiment with our needle
and prism.

If we study sunlight (taking care again to shield the prism), by-
allowing a sunbeam to illuminate the needle, we get a spectrum of a
kind differing from those we have seen before, inasmuch as the con-
tinuous band of colour is broken, it is full of dark lines ; that is, some
of the coloured rays are lacking ; and hence images of the needle are
not forthcoming in places. The positions of some of the chief dark
lines lettered by Fraunhofer are shown in Fig. 12.

We now know that this result is produced by what is termed the
absorption of light. To understand it we have only to look at a candle
through glasses of different colours : a blue glass absorbs or stops the
red light, and only the blue end of the spectrum remains; a red
glass absorbs or stops the blue, and only the red end remains.

In these cases large regions of the spectrum are alternately
blotted out as differently coloured glasses are used, but the absorption
with which we have to do mostly is of a more restricted character :
lines, that is, single images of the slit, are in question.

One of the most important results that has been gathered from the

* Proc. Roy. Soc., vol. xxviii, March, 1879.

t In 1869, he wrote as follows : " Jl semble difficile quo la reproduction d'un
pareil phenomene soit tin effet du hasard : ii'est-il pas plus naturel d'admettre que
ces groupes de raies semblables sont des harrnoniques qui tiennent a la constitution
moleculaire du gaz luniineux ? II faudra sans doule un grand nombre d' observa-
tions analogues pour decouvrir la loi qui regit ces harnioniques."

[.] PRINCIPLES AND METHODS. 13

study of these absorption effects is that if we look at a light source
competent to give us a continuous spectrum through any of the
vapours or gases we have so far considered as producing bright lines,
provided the .light source is hotter than the gases or vapours, the par-
ticular rays constituting the bright line or discontinuous spectrum of
each of the vapours as gases will be cut out from the light of the con-
tinuous spectrum.

Explanation of Absorption.

While in the giving out of light we are dealing with molecular
vibration taking place so energetically as to give rise to luminous
radiation ; absorption phenomena afford us evidence of this motion of
the molecules when their vibrations are far less violent. The molecules
can only vibrate each in its own period, and they will even take up
vibrations from light which is passing among them, provided always
that the light thus passing among them contains the proper vibrations.

An illustration from what happens in the case of sound will help to
make this clear. If we go into a quiet room where there is a piano,
and sing a note and stop suddenly, we find that note echoed back from
the piano. If we sing another note, we find that it is also re-echoed
from the piano. How is this 1 When we have sung a particular note,
we have thrown the air into a particular state of vibration. One wire
in the piano was competent to vibrate in harmony with it. It did so,
and, vibrating after we had finished, kept on the note.

This principle may be illustrated in another and very striking
mariner by means of two large tuning-forks mounted on sounding-boxes
and tuned in exact unison. One of the forks is set in active vibration
by means of a fiddle-bow, and then brought near to the other one, the
open mouths of the two sounding-boxes being presented to each other
to make the effect as great as possible. After a few moments, if the
fork originally sounded is damped to stop its sound, it will be found
that the other fork has taken up the vibration and is sounding, not so
loudly as the original fork was, but still distinctly. If the two forks
are not in perfect unison, no amount of bowing of the one will have
the slightest effect in producing sound from the other. Again, suppose
we have a long room, and a fiddle at one end of it, and that between it
and an observer at the other end of the room there is a screen of fiddles,
all tuned like the solitary one, we can imagine that in that case the
observer would scarcely hear the note produced upon any one of the
open strings of the solitary fiddle. Why 1 The reason is that the air-
pulses set up by the open string of this fiddle, in tune with all the
others, would set all the other similar strings in vibration ; the air pulses
set in motion by the vibration of the fiddle cannot set all those strings

14 INORGANIC EVOLUTION. [CHAP.

vibrating and still pass on to one's ear at the other end of the room as
if nothing had happened to them.

Now apply this to light. Suppose we have at one end of a room a
vivid light source giving us all possible waves of light from red to
violet. This we may represent as before by

. Sly w a m © Y © L^

Also suppose that we have in the middle of the room a screen of
molecules, say a sodium flame, capable of emitting yellow light,

What will happen 1 Will the light come to our eyes exactly as if
the molecules were not there ? No ; it will not. What then will be
the difference 1 The molecules which vibrate at such a rate that they
give out yellow light, keep for their own purpose — filch, so to speak,
from the light passing through them — the particular vibrations which
they want to carry on their own motions, and we shall have

\V7 n PET) (f\i /TS\ nET)

\J 'J LED vLfl (^) Lr\i

as a result ; the light comes to us minus the vibrations which have thus
been utilised, as we may put it, by the screen of vapour. We have, in
fact, an apparently dark space which may be represented thus :

- w a [§ © Y © \&

In the spectroscope we see what would otherwise be a continuous
spectrum, with a dark band across the yellow absolutely identical in
position with the bright band observed when the molecules of the
vapour of which the screen is composed radiated light in the first in-
stance. It is not, however, a case of absolute blackness, or absence of
that particular ray, for the molecules are set in vibration by the rays
which they absorb, and therefore give out some light, but it is so feeble
as to appear black by contrast with the very much brighter rays coming
direct from the original source.

This great law may be summed up as follows : Gases and vapours,
wJien relatively cool, absorb those rays which the)/ themselves emit wlien
incandescent ; the absorption is continuous or discontinuous (or selective)
as the radiation is continuous or discontinuous (or selective).

I have referred to this matter at. some length because in our light
sources, in the sun, an.d in most of the stars we have light from a more
highly heated centre passing through an envelope of cooler vapours,
and on this account absorption phenomena are produced.

PRINCIPLES AND METHODS.

15

It was Fraunhofer, at the beginning of the century which is now
so rapidly passing away, who was the discoverer of the fact that the
spectrum Jof the sun was discontinuous with dark lines.

When we wish to go further afield than the sun, that is, to the stars,

16

INORGANIC EVOLUTION.

[CHAP.

we must first use a telescope to collect the light, and then employ a
spectroscope.

- Fig. 13 shows a spectroscope thus attached at the eye-piece end of
the great Lick refractor. In astronomical inquiries the same methods
of work are employed, and although it will be seen that we are now
far beyond the improvised spectroscope with which we began, both in
construction and use, no new' principle is involved.

FIG. 13. — A stellar spectroscope attached 'to the Lick equatorial.

Now if my reader has not hesitated to invest his or her sixpence
in a prism, and has had the patience (no other quality is needed) to do
what I have suggested, the way is open to read without difficulty most

I.] PRINCIPLES AND METHODS. 17

books involving spectrum analysis which he or she- is likely to come
across ; terms such as

Spectrum Fluted spectra

Continuous spectrum Discontinuous (or selective) spectrum

Grating Fraunliofer lines

Prism Wave-length, wave-frequency

Spectroscope Kadiation

Slit Absorption

Line spectra Series,

should now have acquired a definite meaning, and I trust the expressive
ness of the terms will be acknowledged while they are accepted as part
of the future mental stock-in-trade.

IS

CHAP. II. — SOME PIONEERING DIFFICULTIES.

I began to endeavour to apply the principles of spectrum
analysis to the investigation of the nature of the heavenly bodies in
1865, the then idea, based upon Kirchhoff and Bunsen's work of
1859, was that the spectrum of a chemical element was one and in-
divisible— that it could not be changed by temperature or by anything
else.

Looking back it is easy to see now that this idea largely depended
upon the fact that in the early days low j flame temperatures were
generally employed, and that it so happens that the substances best
visible in the flame and which were therefore chosen to experiment
upon, such as sodium, calcium, potassium and the like, give us line
spectra at low stages of heat.

Hence the first spectroscopic ideas entirely agreed with those of
the chemist, that the chemical " atom," defined by a certain " atomic ""
weight was a manufactured article, indivisible, indestructible. Chemi-
cal elementary substances were either composed of these atoms, these
indivisible units ; or of " molecules " consisting of two or more of
them, hence the terms " diatomic " and " polyatomic " molecule.

The difference between the spectra of the same element in the solid
and gaseous, states, in which we have first a continuous and secondly a line
spectrum, was ascribed to the restricted motion of the atom in the solid
and its freedom in the gaseous state — it was a question of " free path."
The difference between the states which gave us the continuous and dis-
continuous spectra was a physical difference having nothing to do with
chemistry. According to the kinetic theory of gases, the particles of all
bodies are in a state of continual agitation, and the difference between
the solid, liquid, and gaseous states of matter is that in a solid body the
molecule never gets beyond a certain distance from its initial position.
The path it describes is often within a very small region of space. Prof.
Clifford, in a lecture upon atoms, many years ago illustrated this very
clearly. He supposed a body in the middle of a room held by elastic
bands to the ceiling and the floor, and in the same manner to each side
of the room. Now pull the body from its place ; it will vibrate, but
always about a mean position ; it will not travel bodily out of its
place ; it will always go back again.

We next come to liquids. Concerning these we read : "In fluids,
on the other hand, there is no such restriction to the excursions of a

CHAP. II.] PIONEERING DIFFICULTIES. 19'

molecule. It is true that the molecule generally can travel but a very
small distance before its path is disturbed by an encounter with some
other molecule ; but after this encounter, there is nothing which
determines the molecule rather to return towards the place from
whence it came than to push its way into new regions. Hence in
liquids the path of a molecule is not confined within a limited region, as
in the case of solids, but may penetrate to any part of the space occu-
pied by the liquid.

Now we have the motion of the molecule in the solid and the liquid..
How about the movement in a gas 1 "A gaseous body is supposed to-
consist of a large number of molecules moving very rapidly." For in-
stance, the molecules of air travel about 20 miles in a minute. " During,
the greater part of their course these molecules are not acted upon by
any sensible force, and therefore move in straight lines with uniform
velocity. When two molecules come within a certain distance of each
other, a mutual action takes place between them which may be com-
pared to the collision of two billiard balls. Each molecule has its
course changed, and starts in a new path."

The collision between two molecules is denned as an " encounter" ;.
the course of a molecule between encounters a " free path." " In
ordinary gases the free motion of a molecule takes up much more time
than is occupied by an encounter. As the density of the gas increases
the free path diminishes."

It will be seen at once that on the view first held that the differ-
ence between continuous and discontinuous spectra depended simply
upon the solid and gaseous states, no solid could give us 'a line spec-
trum ; and the well-known absorption spectra of didymium glass and
other solid bodies would be impossible.

Another important series of facts was soon brought to the front.
Pliicker and Hittorf in the year 1865, announced that "there is a
certain number of elementary substances which when differently
treated furnish two kinds of spectra of quite a different character, not
having any line or band in common." The difference in character to-
which reference is here made consists in the spectrum produced at the
lower temperature being composed of flutings, which are replaced by
lines when the higher temperature is reached.

This was the first blow aimed at the general view — one element
one spectrum — to which I have referred above. It was met in two-
way s.

Taking the line spectrum as representing the true vibration of the
chemical unit, I have already shown that the continuous spectrum was
explained as due to its physical environment, the solid or liquid state.
This, then, had not to be considered from the chemical point of view.

C 2

20 INORGANIC EVOLUTION. [CHAP.

The fluted spectra were boldly ascribed to " impurities," but not
always wisely, for, to get rid of the difficulty presented by the two
spectra of hydrogen, two perfectly distinct spectra were ascribed to
acetylene. Again the " bell-hypothesis " was suggested, according to
which the spectrum did not depend so much upon the substance as
upon the way it was made to vibrate. According to this view the
same chemical atom might have a dozen spectra if struck in a
dozen differant ways.

But it was answered that this argument proved too much ; and for
this reason. Mitscherlich showed in 1864 that some bodies known
to be chemical compounds when raised to incandescence, give us a
spectrum special to the compound ; that is, they have a spectrum of
their own ; no lines of either of the constituents are seen.

I showed later that when the temperature was sufficient to produce
decomposition, the lines of the elementary bodies of which the com-
pound was composed made their appearances according to the tempera-
ture employed. And I also showed that precisely the same thing
happens with regard to the fluted and line spectra of the same chemical
element. We may get the first alone at a low temperature ; we may
increase the temperature and dim it slightly, some lines making their
appearance ; and next, by employing a very high temperature, we can
abolish the fluted spectrum altogether and obtain one with lines only.

Since then the difference between the two spectra of the same ele-
ment was no more marked than the difference between the spectrum of
a known compound and its constituents after the compound had been
broken up by heat, it was as logical to deny the existence of compound
bodies as to deny that more molecular complexities than one were in-
volved in spectral phenomena.

Attacks like these finally caused the chemists to reconsider their
position, and some time later, being under the impression, which has
turned out to have no justification, that " monatomic " elements like
mercury have not fluted spectra, they conceded that the fluted spectra
might represent the vibration of the " diatomic " molecule in the
" diatomic " elements. This, of course, was to give up the " bell-
hypothesis."

At the time when the differences of opinion arising from the ex-
istence of fluted as well as line spectra in the case of many elements
were being discussed, solar observations were beginning to bring before
us a perfect flood of facts apparently devoid of any law or order. In
1866 I threw an image of the sun on the slit of a spectroscope (Fig. 14),
in order to observe the spectra of its different parts, and in this way
the spectra of sun-spots (Fig. 19) and eventually of prominences
were observed.

II.]

PIONEERING DIFFICULTIES.

'21

FIG. 14. — Spectroscope attached to a large refractor which throws an image of'
the sun on the slit plate.

In the first method of work adopted in the laboratory the spectro-
scope was directed to the light source, so that the spectrum was built
up of the light coming from all parts of it without distinction.

FIG. 15. — The first method of work with the slit of the spectroscope close to the light
source. In the experiment illustrated the light source is an electric spark produced
by an induction coil with Leyden jar in circuit. The slit end of the spectroscope
is shown to the right.

"22 INORGANIC EVOLUTION. [CHAP.

In 1869 I introduced into laboratory work the method adopted in
the case of the sun in the observatory ; that is, an image of each light
source experimented on was thrown on to the slit by a lens (Fig. 16),
so that the spectrum of each part of it could be observed, and some of
the results obtained by the new method were the following : —

FIG. 16. — The method of throwing an image of the light source (in this case a
candle flame) on the slit plate of a laboratory spectroscope.

The spectral lines obtained by using such a light source as the
electric arc or spark were of different lengths ; some appeared only in
the spectrum of the centre of the light source, others extended far
into the outer envelopes. This effect was best studied by throwing the
image of a horizontal arc or spark on a vertical slit. The lengths of
the lines photographed in the electric arc of many metallic elements,
were tabulated and published in 1873 and 1874.

In Figs. 17 and 18 these so-called "long arid short lines" are
illustrated. In one case we deal with a mixture of the salts of calcium
and strontium, in the other with the metal sodium. The richness of the
lines in the spectrum of the core of the arc will be best gathered from
Fig. 17, the variations in the lengths of the lines from Fig. 18.

Here then was the first glimpse of the idea that the complete
spectrum of a chemical element obtained at the highest temperature
might arise from the summation of two or more different line spectra,
produced at different degrees of temperature, and therefore bringing us
in presence of two or more molecular complexities ; that is, different
molecules broken up at different temperatures. So soon as experi-

„.]

PIONEERING DIFFICULTIES

23

merits in the laboratory had given a definite result with regard to the
spectrum of a metal in this way, I proceeded to study the sun with
a view of determining how that metal behaved in the sun.

This involved, first, photographs of the solar spectrum with its dark
lines, photographic comparisons of these dark lines with the bright

a o

* PH

1 J

-3 *

bC
O

I

lines constituting the spectra of the metallic elements. This enabled
us to compare the total light given by each light source with the light
received from all parts of the sun indiscriminately.

24

JNORtiANIC EVOLUTION.

[CHAP

. 18. — The longs and shorts of sodium taken under the same conditions, showing
that the orange line extends furthest from the poles.

Next the spectra of different parts of the sun — chromosphere and
prominences and spots — were compared with different parts of the
light source, the core of , the arc, and the centre of the spark, and the
outer regions of both.

It will be seen that the inquiry now had a very broad base, and it
could be immediately tested in many ways at every stage.

Wonderful anomalies were at once detected ; lines known to belong
to the same chemical element behaved differently in several ways.
Some were limited to prominences, others to spots (Fig. 20), and in
solar storms different iron lines indicated different velocities (Fig. 21).
In the spectrum of the hottest part of the sun open to our inquiries,
the region namely immediately overlying the photosphere, which I
named the chromosphere, the anomalies became legion ; suffice to say
that in the hottest part of the sun we could get at, the spectrum of
iron then represented in Kirchhoff's map of the ordinary solar spec-
trum by 460 lines was reduced to three lines.

It was no longer a question merely of settling the difficulties raised
by the observations of Pliicker and Hittorf.

Many observations and cross references of this kind during the
next few years convinced me that the view that each chemical element
had only one line spectrum was erroneous, and that the results ob-

II.] PIONEERING DIFFICULTIES. 25

tained suggested that the various terrestrial and solar phenomena were
produced by a series of simplifications brought about by each higher
temperature employed. That is, that- the new instrument, the spectro-
scope, showed that higher temperatures than those previously em-
ployed were doing for chemistry what previous similar inquiries had
done, namely, indicating the existence of finer constituents in matter
supposed at each point of time to be elementary.

This was the first glimpse of dissociation in relation to the produc-
tion of changes in the line spectrum.

By the year 1872 the work of Rutherfurd and Secchi on stellar
spectra enabled the base of the inquiry to include the stars as well as
the sun. In some of the stars the existence of hydrogen, magnesium,
and carbon were beyond question. The point that first struck me
was that in white stars like a Lyrse and Sirius, with continuous spectra
extending far into the violet — stars therefore hotter than their fellows
of a yellow or red colour — we had to do with hydrogen almost alone.

It was in 1873 that I first called the attention of the Royal Society
to the very remarkable facts which had even then been brought to-
gether regarding the possible action of heat in the sun and stars.
Referring more especially to the classification of stars by Rutherfurd,
I wrote as follows : — *

"I have asked myself whether all the above facts cannot be grouped
together in a working hypothesis which assumes that in the reversing

FIG. 19. — Spectrum of a sun-spot as compared with the general spectrum,
showing that certain metallic lines (sodium and calcium in this instance)
are widened. The darker portion represents the spectrum of the spot.

layers of the sun and stars various degrees of ' celestial dissociation *
are at work, which dissociation prevents the coming together of the
atoms which, at the temperature of the earth and at all artificial tem-
peratures yet attained here, compose the metals, the metalloids and
compounds."

Subsequently in a private letter to M. Dumas, who took the

* Phil Trans., vol. clxiv, Part IF, p. 491.

26

INORGANIC EVOLUTION.

[CHAP.

keenest interest in my solar work, I wrote, " II semble que plus une
etoile est chaude plus son spectre est simple."

I also pointed out the close relation of hydrogen to calcium, mag
nesium and other metals (it was on this ground that I had named the

I

substance which gave D3, which always varied with hydrogen, helium),
and the absence of all other terrestrial gases from the solar spectrum.
An interesting discussion at the Paris Academy of Sciences was thus
•concluded by M. Dumas :

PIONEERING DIFFICULTIES.

27

II.]

"En resume, quandje soutenais devant 1'Academie que les elements
de Lavoisier devaient etre considered, ainsi qu'il avait e*tabli lui-meme,
non comme les elements absolus de 1'univers, mais comme les elements
rehtifs de 1'experience humaine ; quand je professais, il y a longtemps,

FIG. 21. — Different rates of motion registered by different iron lines.

que lliydrogbie etait plus pres des me'taux que de toute autre classe de
corps ; j'emettais des opinions que les decouvertes actuelles viennent
•confirmer et que je n'ai point a modifier aujourd'hui."*

One of the replies to my working hypothesis was that the various
chemical elements probably existed in different proportions in the
different stars, and that it so happened that in Vega and Sirius one of
•them, hydrogen, existed practically alone.

In 1878 I went further, and showed that thousands of solar pheno-
mena which had been carefully recorded during the previous years
•could only be explained by assuming that the changes in the various
intensities of lines in the line spectrum itself indicated successive dis-
sociations. I pictured the effect of furnaces of different temperatures,
and I wrote as follows :f

* Chemistry of the Sun, p. 205.

f Proc. Soy. Soc., vol. xxviii, p. 169.

See also Chemistry of the Sun, chap.

atvin.

28 INORGANIC EVOLUTION. [CHAP. II.

" It is abundantly clear that if the so-called elements, or, more
properly speaking, their finest atoms — those that give us line spectra
— are really compounds, the compounds must have been formed at a
very high temperature. It is easy to imagine that there may be no
superior limit to temperature, and therefore no superior limit beyond
which such combinations are possible, because the atoms which have
the power of combining together at these transcendental stages of heat
do not exist as such, or rather they exist combined with other atoms,
like or unlike, at all lower temperatures. Hence association will be a
combination of more complex molecules as temperature is reduced, and
of dissociation, therefore, with increased temperature, there may be
no end."

In 1878 I went back to the study of the changes in the line spectra
in relation to the changes observed when known compounds were dis-
sociated, and after discussing certain objections, I submitted the con-
clusion that the known facts with regard to the changes in line
spectra " are easily grouped together, and a perfect continuity of
phenomena established on the hypothesis of successive dissociations
analogous to those observed in the cases of undoubted compounds."*

It is thus seen that the conclusions to which my spectroscopic work
up to the year 1880 had led me, tended in exactly the same direction
as that indicated by more purely chemical inquiries thus referred to by
Berthelot in that year : —

"L'etude approfondie des proprie'tes physiques et chimiques des
masses e'le'mentaires, qui , constituent nos corps simples actuels, tend
chaque jour d'a vantage a les assimiler, non a des atonies indivisibles,
homogenes et susceptibles d'eprouver seulement des mouvements
d'ensemble, . . . il est difficile d'imaginer un mot et une notion
plus contraires a 1'observation ; mais a des Edifices fort complexes,,
dou^s d'une architecture specifique et anime's des mouvements intestins
ires varies."!

* Proc. Roy. Soc., vol. xxviii, p. 179.
f Comptes rendu*, 18SO, vol. xc, p. 1512.

CHAP. III. — THE PRESENT POSITION.

IN the last chapter I referred to some of the difficulties encountered
by the earlier researchers in spectrum analysis. In the present one I
propose to pass over the history of nearly twenty years' work, with all
its attendant doubts and difficulties, and deal with what that work has
brought us, a perfect harmony between laboratory, solar and stellar
phenomena.

It has been proved beyond all question that not only are both
fluted (or channelled-space) spectra and line spectra visible in the case
of most of the elements, but that many of the metallic elements with
which I shall have to deal in the sequel have at least two sets of lines
accompanying, if not resulting from, the action of widely differing
temperatures.

It is important to mention that the different chemical elements
behave very differently in regard to the action of heat and electricity
upon them as we pass from the solid to the liquid and vaporous forms ;
tha£ is, the two different forms of energy are apt to behave very differ-
ently ; the permanent gases as opposed to the elements which generally
exist in the solid form is the first differentiation ; the elements of low
atomic weights and low melting point as opposed to the rest, is the
second.

In the cases in which heat-energy can go so far, we first get an
increase in the free path of the molecules, and ultimately the latter are
made to vibrate.

In the case of high-tension electricity, on the other hand, increase
of free path is scarcely involved, and hence we may have effects similar
to those produced by high temperature, with scarcely perceptible effects
of heat in the ordinary sense.

Conversing on this subject with my friend Clifford, many years
ago, we came to the conclusion that the energy imparted to a molecule
might cause (1) an extension of free path; (2) a rotation; and (3) a
vibration. To get concrete images of these effects we spoke of patli-
heat, spins-heat, and wobble-heat. The facts seemed to show that heat
energy had no effect in producing line-spectra until the two first results
had been obtained, and, further, that in all gases and many metals it
had no effect in producing vibrations ; while, on the other hand, elec-
trical energy generally acted as if it began at the third stage and is
effective in the case of every chemical substance without exception.

30 INORGANIC EVOLUTION. [CHAP,

However this may be, we now know that many elements present
changes at several widely differing stages of heat. The line spectra of
elements like sodium, lithium, and others may be obtained by the heat
of the flame of a spirit lamp, or an ordinary Bunsen's burner, the sub-
stance being introduced into the flame by a clean platinum wire twisted
into a loop at the end.

This temperature has no effect upon iron and similar metals. To
get any special spectral indication from them a higher temperature
than that of the Bunsen is required ; the blowpipe flame may be resorted
to ; in this a stream of air is blown through the centre of a flame of
coal gas burning at the end of a cylindrical tube.

We get in this way what is called a " flame-spectrum," in which
flutings and some lines are seen. In order to obtain the complete line-
spectra of some of the less volatile metals, like iron and copper, we
are driven to use electrical energy and employ the voltaic current, and
(for choice) metallic poles, which are so strongly heated by the passage
of the current that the vapour of the metal thus experimented on is
produced and rendered incandescent.

We may say generally that no amount of heat-energy will render
visible the spectra of gases. These are obtained by enclosing the gases
in glass tubes, and illuminating them by means of an electric current.
We may go further and say that the ordinary voltaic current used in
laboratories is equally inoperative. We must have the induced
current, and with different tensions different spectra are produced.

We have then arrived so far. Heat-energy, which does give us
line-spectra in some cases when metals are concerned, fails us in the case
of the permanent gases and many metals. A voltaic current gives us
spectra when metals are in question, but, like heat-energy, it will not
set the particles of the permanent gases vibrating.

But when both metals and the permanent gases are subjected to the
action of a strong induced current, that is, a current of high tension,
when an induction coil with Leyden jars and an air break are employed,
we get this vibration ; gases now become luminous, a distinct change in
the spectra of the metals is observed, a change as well marked, or
perhaps better marked, than any of the previous lower temperature
changes to which I have already drawn attention.

When the tension is still further increased, the differences in the
spectra are most marked in the case of gases, for the reason that, being
enclosed in tubes, they cannot escape from the action of the current ;
all the molecules are equally affected. The spectrum is sometimes NOT a
mixed one. In the case of the metals the spark is made to pass between
two small pointed poles, and the region of most intense action is a very
limited one ; we get from the particles outside this region the spectrum

HI.] THE PRESENT POSITION. 31

obtained with a lower degree of electrical energy. The spectrum is a
mixed one. Even when we take the precaution of throwing an image
of the spark on the slit of the spectroscope, the outer cooler layers
pierced by the line of sight add their lines to the spectrum of the
centre.

Not only so, but the individuality of the various chemical elements-
conies out in a remarkable manner.

To take one or two instances. I will begin with the gases with a
weak and strong induced current. Hydrogen gives us what is termed
a structure spectrum, a spectrum full of lines ; this changes to a series..
Oxygen gives us series which change into a complicated line spectrum in
which no series has been traced. Nitrogen gives us a fluted spectrum,
which changes into a complicated line-spectrum.

I next pass to the metals, and again, for brevity's sake, I will deal
with three substances only. In the case of magnesium, iron, and
calcium, the changes observed on passing from the temperature of the
arc to that of the spark have been minutely observed. In each new
lines are added, or old ones are intensified at the higher temperature.
Such lines have been termed enhanced lines.

These enhanced lines are not seen alone ; as in the case of the spark,,
so in the arc outside the region of high temperature in which they are
produced, the cooling vapours give us the lines visible at a lower tem-
perature.

Bearing in mind what happens in the case of the gases, we can con-
ceive the enhanced lines to be seen alone at the highest temperature in
a space sufficiently shielded from the action of all lower temperatures,,
but such a shielding is beyond our laboratory expedients ; still, as I
shall show, in the atmospheres of the stars we have probably the closest
approximation open to our observation of that equally heated space
condition to which I have referred.

The enhanced lines are very few in number as compared with those
seen at the temperature of the arc. In the case of iron thousands are
reduced to tens.

The above statements are only general; if we include the non-
metals, more stages of temperature are required, and it then becomes-
evident that different kinds of spectra are produced at the same tem-
perature in the case of different elements ; in other words, at many
different heat-levels changes occur, always in one direction, but differing
widely for different substances at the lower temperatures. At the
highest temperatures — at the limit — there is much greater constancy
in the phenomena observed if we disregard the question of series. If
considered from the series point of view, there is no constancy at all.

It is obvious that with all these temperature effects observed in a-

32 INORGANIC EVOLUTION. [CHAP.

large number of elements, very many comparisons are rendered possible.
All these suggest that if dissociation is really in question, in some
cases at least more than two simplifications in the line stage are
necessary to explain the facts. It is possible that the effects at first
ascribed to quantity may be due to the presence of a series of molecules
of different complexities, and that this is the true reason why " the more
there is to dissociate, the more time is required to run through the
series, and the better the first stages are seen."*

After this general statement of the changes in spectra observed to
accompany change in the quantity and kind of energy used in the
experiments, I propose to refer briefly to the most recent work on this
subject, touching the changes observed on passing from the arc to the
spark in the case of many of the metallic elements. By the kindness
of Mr. Hugh Spottiswoode, the photographs of the enhanced lines have
been obtained by the use of the large induction coil, giving a 40-inch
spark, formerly belonging to Dr. Spottiswoode, P.K.S. I am anxious
to express here my deep obligation to Mr. Hugh Spottiswoode for
the loan of such a magnificent addition to my instrumental stock-in-
trade.

The spark obtained by means of the Spottiswoode coil is so luminous
that higher dispersions than those formerly employed can be effectively
used, and in consequence of this, the detection of the enhanced lines
becomes more easy; their number therefore has been considerably
increased.

At the higher temperature enhanced lines have been found Lo maKe
their appearance in the spectra of nearly all the metals already ex-
amined. Lithium is one exception.

Neglecting then all changes at the lowest temperatures, but
including the flame spectrum, four distinct temperature stages are
indicated by the varying spectra of the metals ; for simplicity I limit
myself to iron as an example. These are : —

1. The flame spectrum, consisting of a few lines and flutings only,
including several well-marked lines, some of them arranged in triplets.

2. The arc spectrum, consisting, according to Rowland, of 2,000
lines or more.

3. The spark spectrum, differing from the arc spectrum in the
enhancement of some of the short lines arid the reduced relative
brightness of others.

4. A spectrum consisting of a relatively very small number of lines
which are intensified in the spark. This, as stated above, we can
conceive to be visible alone at the highest temperature in a space
efficiently shielded from the action of all lower ones, since the enhanced

* Pr.jc. Roy. Soc., 1879, No. 200.

III.] THE PRESENT POSITION. 33

lines behave like those of a metal when a compound of a metal is
broken up by the action of heat.

Each line of each element, at whatever temperature it is produced,
can at once be compared in relation to position in the spectrum with
the lines visible in celestial bodies with a view of determining whether
the element exists in them.

At the time at which the earlier inquiries of this kind were made
it was only possible for the most part to deal with eye observations of
the heavenly bodies. The results were, therefore, limited to the visible
spectrum.

During the last few years photographs of the spectra of the brighter
stars and of the sun's chromosphere during eclipses have been obtained ;
it became of importance, therefore, to extend the observations of terres-
trial spectra into the photographic regions for the purpose of making
the comparisons which were necessary for continuing the inquiry.

The recent work has been done with this object in view.

The way in which the enhanced lines have been used is as follows.
Those belonging to some of the chief metallic elements have been
brought together, and thus form what I have termed a " test-spectrum."
This has been treated as if it were the spectrum of an unknown element,
and it has been compared with the various spectra presented by the sun
and stars.

How marvellous, how even magnificent, the results of this inquiry
have been, I shall show later in detail ; but I may here say by way of
anticipation that the test-spectrum turns out to be practically the
spectrum of the chromosphere, that is, the spectrum of the hottest
part of the sun that we can get at ; and that a star has been found in
which it exists almost alone, nearly all the lines of which had previously
been regarded as " unknown."

This last result is of the highest order of importance, because it
should carry conviction home to many who were not satisfied with
the change of spectrum as seen in a laboratory, where always the
enhanced lines seen in the spectrum of the centre of the spark
have alongside them the lines in the spectrum of the outer envelope,
which of course is cooling, and in which the finer molecules should
reunite. For twenty years I have longed for an incandescent bottle in
which to store what the centre of the spark produces. The stars have
now provided it, as I shall show.

Although I have promised to pass over the history of the work
generally, I must still point out that the enhanced lines in the test-
spectrum actually include all those first studied years ago when every-
thing was dim, and we were seeing through a glass darkly ; not as we
are now, face to face. To show the rigid connection of the new with

D

34 INORGANIC EVOLUTION. [CHAP.

the old, it is desirable to refer briefly to some of the work undertaken
in relation to some of the first anomalies noted.

One advantage of this method of treatment is that it shows that
the immense mass of facts now available supports all the conclusions
drawn from the meagre evidence available a quarter of a century ago.

Some of the anomalies were as follows : they are given as specimens
of many.

1. Inversion of intensity of lines seen under different circum-
stances.

I showed in 1879 that there was no connection whatever between
the spectra of calcium, barium, iron and manganese and the chromo-
sphere spectrum beyond certain coincidences of wave-length. The
long lines seen in laboratory experiments are suppressed, and the
feeble lines exalted in the spectrum of the chromosphere. In the
Fraunhofer spectrum, the relative intensities of the lines are quite
different from those of coincident lines in the chromosphere.

2. The simplification of the spectrum of a substance at the tem-
perature of the chromosphere. To take an example, in the visible
region of the spectrum, iron is represented by nearly a thousand
Fraunhofer lines ; in the chromosphere it has only two representatives.

3. In sun spots we deal with one set of iron lines, in the chromo-
sphere with another.

4. At the maximum sun-spot period the lines widened in spot
spectra are nearly all unknown ; at the minimum they are chiefly due
to iron and other familiar substances.

5. The up-rush or down-rush of the so-called iron vapour in the
sun is not registered equally by all the iron lines, as it should be on
the non-dissociation hypothesis. Thus, as I first observed in 1880,
while motion is sometimes shown by the change of refrangibility of
some lines attributed to iron, other adjacent iron lines indicate a state
of absolute rest.

Laboratory work without stint has been brought to bear, with a
view of attempting to explain the anomalies to which attention has
been directed.

I only refer here to the work done on iron, magnesium and calcium,
to show that in those metals the anomalies were to a large extent due to
the lines now termed enhanced — that is, the lines seem to considerably
change their intensities when the highest temperatures are employed.

Iron.

In the course of my early observations of the spectrum of the
chromosphere, I discovered on June 6, 1869, a bright line at 1474 on

III.] THE PRESENT POSITION. 35

Kirehhoff' s scale, which I stated to be coincident with a line of iron.
On June 26 I discovered another at 2003'4 of the same scale.

The later researches on the spectrum of iron have shown that the
iron line which I observed in 1869 to be coincident with the bright
chromospheric line at 1474 on Kirehhoff' s scale, having a wave-length of
5316-79, is an enhanced line, agreeing absolutely with Young's latest
determination of the wave-length of the 1474 chromospheric line.

Similarly the line at 2003'4 of Kirchhoff s scale, with a wave-length
of 4924, is also an enhanced line of iron.

The first experiments were made to explain my own and the
Italian observations of the chromosphere which proved the presence of
only these two lines of iron in the part of the spectrum ordinarily
observed ; the ordinary spectrum of iron, in which 460 lines had been
mapped at that time, was entirely invisible.

The anomalies were investigated in the experimental work with
sparks produced by quantity and intensity coils, with and without jars
in the circuit. The outcome of these experiments was to show that the
chromospheric representatives of iron were precisely the lines which
were brightened on passing from the arc to the spark, while the lines
widened in spots corresponded to a lower temperature.

The next anomaly observed was that in a sun spot the iron line at
4924 often indicated no movement of the iron vapour, while the other
iron lines showed that it was moving with considerable velocity.

It seemed perfectly clear then that in the sun " we were not
dealing with iron itself, but with primitive forms of matter contained
in iron which are capable of withstanding the high temperature of the
sun, after the iron observed as such has been broken up, as suggested
by Brodie."*

On this view, the high temperature iron lines of the chromosphere
represent the vibrations of one set of molecules, while the lines which
are widened in spots correspond to other molecular vibrations.
Similarly, the idea of different molecular groupings provides a satis-
factory explanation of the varying rates of movement of iron vapour
indicated by adjacent lines, the lines being produced by absorption of
different molecules at different levels and at different temperatures.

.-'•' i
Magnesium.

In 1879 I passed a spark through a flame charged with vapours of
different substances. In the case of magnesium the effect of the higher
temperature of the spark was very marked ; some of the flame lines
being abolished, while two new ones made their appearance, one of

•* Proc. Rot/. Soc., vol. xxxii, p. 234.

D 2

36 INORGANIC EVOLUTION. [CHAP.

them at 448 1 . The important fact was that the lines special to the
flame did not appear among the Fraunhofer lines, while some of those
of the spark did appear.

This line at 4481 now takes its place among the enhanced lines like
those of iron previously mentioned ; special cases now form pait of the
more general one.

Here again the experiments pointed to varying degrees of dissocia-
tion at different temperatures as the cause of the non-appearance of
some of the magnesium lines in the Fraunhofer spectrum.

From these experiments, the results of which were subsequently
mapped in relation to the various heat-levels indicated by solar pheno-
mena, I drew the following conclusions in 1879 : —

"I think it is not too much to hope that a careful study of such
maps, showing the results already obtained, or to be obtained, at
varying temperatures, controlled by observations of the conditions
under which changes are brought about, will, if we accept the idea
that various dissociations of the molecules present in the solid are
brought about by different stages of heat, and then reverse the process,
enable us to determine the mode of evolution by which the molecules
vibrating in the atmospheres of the hottest stars associate into those
of which the solid metal is composed. I put this suggestion forward
with the greater confidence, because I see that help can be got from
various converging lines of work."*

Calcium.

In 1876 I produced evidence that the working hypothesis that the
molecular grouping of calcium which gives a spectrum having its prin-
cipal line at 4226'9 is nearly broken up in the sun, and quite broken
up in the spark, explained the facts which are that the low temperature
line loses its importance in the spectrum of the sun, in which H and K
are by far the strongest lines,

I summed up the facts regarding calcium as follows : — " We have
the blue line differentiated from H and K by its thinness in the solar
spectrum while they are thick, and by its thickness in the arc while
they are thin. We have it again differentiated from them by its ab-
sence in solar storms in which they are almost universally seen, and,
finally, by its absence during eclipses, while the H and K lines have
been the brightest seen or photographed."

I afterwards attempted to carry the matter further by photograph-
ing the spectra of sun spots. In all cases H and K lines were seen
reversed over the spots, just as Young saw them at Sherman, while

* Proc. Hoy. Soc., 1879, vol. xxx, p. 30.

III.] THE PRESENT POSITION. 37

the blue calcium line was not reversed. The oldest of these photo-
graphs which has been preserved bears the date April 1, 1881.

The experimental results in the case of calcium, therefore, followed
suit with those obtained from iron and magnesium, and indicated that
the cause of the inversion of intensities in the lines of a substance
under different circumstances is due to the varying degrees of dissocia-
tion brought about by different temperatures.

Both in the case of iron, magnesium and calcium, the high tempera-
ture lines involved are not seen at all at lower temperatures, and even
in the case of calcium, when photographic exposure of 100 hours' dura-
tion have been employed. It should be sufficiently obvious to every-
body from this that temperature alone is in question.

Finally, then. The similar changes in the spectra of certain ele-
ments, changes observed in laboratory, sun and stars are simply and
sufficiently explained on the hypothesis of dissociation. If we reject
this, so far no other explanation is forthcoming which co-ordinates and
harmonises the results obtained along the different lines of work. Nor
is this all : as I shall show later on, there are other branches of
physical inquiry which suggest the same hypothesis.

38

BOOK IL— APPLICATION OF THE INQUIRY TO THE SUN

AND STARS.

CHAP. IV. — THE SUN'S CHROMOSPHERE.

I STATED in the previous chapter (p. 33), that in order to utilize the
information placed at our disposal by the discovery of the new lines
seen in the spectra of metals exposed to high temperatures, I had
brought the enhanced lines of the chief metallic elements together, and
thus formed a " test-spectrum " to use as a new engine of research in
regions of work where help might be expected from it.

In this chapter I shall deal with the application of this test-spectrum
to a study of the sun.

It is obvious that the general spectrum of the sun, like that of -stars
generally, is built up of all the absorptions which can make themselvc*
felt in every layer of its atmosphere from bottom to top, that is from
the photosphere to the outermost part of the corona. It is important
to note that this spectrum is cJiangeless from year to year.

Now sun spots are disturbances produced in the photosphere ; and
the chromosphere, with its disturbances, called prominences, lie directly
above it. Here, then, we are dealing with the lowest part of the sun's
atmosphere. We find first of all that in opposition to the changeless
general spectrum, great changes occur with the sun-spot period, both
in the spots and chromosphere.

The spot spectrum is indicated, as was found in 1866, by the widen-
ing of certain lines; the chromospheric spectrum, as was found in 1868,
by the appearance at the sun's limb of certain bright lines. In both
cases the lines affected seen at any one time are almost always rela-
tively few in number.

Since 1868 we have been enabled to observe not only the spectrum
of the sun's spots, but that of the chromosphere as well, every day
when the sun shines. The chromosphere is full of marvels. At first,
when our knowledge of spectra was very much more restricted than
now, almost all the lines observed were unknown. In 1868 I saw a
line in the yellow, which I found behaved very much like hydrogen,
though I could prove that it was not due to hydrogen ; for laboratory
use the substance which gave rise to it I named helium. Next year,
as I stated in the last chapter, I saw a line in the green at 1474 of
Kirchhoff's scale. That was an unknown line, but in some subsequent

CHAP. IV.] THE SUN'S CHROMOSPHERE. 39

researches I traced it to iron. From that day to this we have observed
a large number of lines.

But useful as the method of observing the chromosphere without
an eclipse, which enables us

" . . . to feel from world to world,"

as Tennyson has put it, has proved, we want an eclipse to see it face to
face.

During the eclipses of 1893, 1896, and 1898, a tremendous flood of
light has been thrown upon it by the use of large instruments con-
structed on a plan devised by Kespighi and myself in 1871. These
give us images of the chromosphere painted by each one of its radia-
tions, so that the exact locus of each chemical layer is revealed. One
of the instruments employed during the Indian eclipse lias also been used
in photographing metallic spectra and the spectra of stars, so that it is
now easy to place photographs of the spectra of the chromosphere ob-
tained during a total eclipse, and of the various metals and stars side
by side.

As in the case of the photographs taken with the prismatic cameras
in 1893 and 1896, the spectrum of the chromosphere in 1898 is very
different from the Fraunhofer spectrum, so that we have not to deal
with a mere reversal of the dark lines of ordinary sunlight into bright
ones.

Many very strong chromospheric lines, the helium lines for example,
are not represented among the Fraunhofer lines, while many Fraunhofer
lines are absent from the chromospheric spectrum (Fig. 23).

But the most remarkable result is that in the eclipse photograph of
the chromosphere spectrum, the most important of the metallic lines
are precisely those included in the " test-spectrum " (Fig. 22). This
photograph in fact deals chiefly with the enhanced metallic lines.

I recognise in this result a veritable Rosetta stone, which will
enable us to read the terrestrial and celestial hieroglyphics presented
to us in spectra, and help us to study them and get at results much
more distinctly and certainly than ever before. The result proves
conclusively that the absorption in the sun's atmosphere which pro-
duces the Fraunhofer lines is not produced by the hottest lowest
stratum, the chromosphere.

It is imperative in order to clear the ground for the future study
of stellar spectra, to inquire fully into the true locus of absorption.
One of the most important conclusions we draw from the Indian
eclipse is that, for some reason or other, the lowest hottest part of the
sun's atmosphere does not write its record among the lines which build
up the general spectrum so effectively as does another.

INORGANIC EVOLUTION.

[CHAP.

IV.] THE SUNS CHROMOSPHERE. 41

This conclusion differs considerably from the opinion generally held.
In my paper on the eclipse of 1893,* I referred at length to this
point. The matter is so important that I do not hesitate to quote
what I then said.

" As a result of solar spectroscopic observations, combined with
laboratory work, Dr. Frankland and myself came to the conclusion,
in 1869, that at least in one particular, Kirchhoff's theory of the
solar constitution required modification. In that year we wrote as
follows : — f

" ' May not these facts indicate that the absorption to which the
reversal of the spectrum and the Fraunhofer lines are due takes
place in the photosphere itself, or extremely near to it, instead of in
an extensive outer absorbing atmosphere ? '

"In an early observation of a prominence on April 17th, 1870, I
found hundreds of the Fraunhofer lines bright at the base, and
remarked that ' a more convincing proof of the theory of the solar
constitution put forward by Dr. Frankland and myself could scarcely
have been furnished.' J

" During the eclipse of 1870, at the moment of disappearance of
the sun, a similar reversal of lines was noticed; we had, to quote
Professor Young, ' a sudden reversal into brightness and colour of
the countless dark lines of the spectrum at the commencement of
totality.' On these observations was based the view that there was a
region some 2"' high above the photosphere, which reversed for us all
the lines visible in the solar spectrum • and on this ground the name
' reversing layer ' was given to it.

" Continued observations, however, led me, in 1873, to abandon
the view that the absorption phenomena of the solar spectrum are
produced by any such thin stratum, and convinced me that the absorp-
tion took place at various levels above the photosphere. I need not
give the evidence here ; it is set forth in my Chemistry of the Sun.§
On the latter hypothesis the different vapours exist normally at
different distances above the photosphere according to their powers of
resisting the dissociating effects of heat.|j

" My observations during the eclipse of 1882, in the seven minutes
preceding totality, to my mind set the matter at rest. * We begin with
one short and brilliant line constantly seen in prominences, never
seen in spots. Next another line appears, also constantly seen in

* Phil. Trans., 1890, vol. clxxxvii, A, p. 603.

f Proc. Soy. Sot., vol. xvii, p. 8S.

J Ibid., vol. xviii, p. 353.

§ Chapter XXII, pp. 303—309.

|j Proc. Hoy. Soc., vol. xxxiv, p. 292.

42 INORGANIC EVOLUTION. [CHAP.

prominences; and now, for the first time, a longer and thinner line
appears, occasionally noted as widened in spots ; while, last of all,
we get, very long, very delicate relatively, two lines constantly seen
widened in spots, and another line, not seen in the spark, and never
yet recorded as widened in spots.'*

" This is one of the most mportant points in solar physics, but
there is not yet a concensus of opinion upon it. Professor Young
and others, apparently, still hold to the view first announced by Dr.
Frankland and myself in the infancy of the observations, that the
Fraunhofer absorption takes place in a thin stratum, lying close to the
photosphere."

I next proceeded to discuss the numerous photographs obtained
during the eclipse, and I gave a map showing that there was only the
slightest relation between the intensities of the lines common to the
Fraunhofer and the eclipse spectrum, and further, that only a few of
the Fraunhofer lines are represented at all. Not only this, but in the
eclipse photographs there are many bright lines not represented at all
among the Fraunhofer lines.

The chromosphere, which represents that part of the sun's atmo-
sphere underlying the true reversing layer, is admirably pourtrayed in
the photographs of the eclipse of 1898. So complete is the record
that it is quite sufficient for our present purpose, and is the more to be
relied on since it represents it at the same instant of time ; I have
elsewhere pointed out that Young's list of chromospheric lines may be
misleading because it is a summation of results obtained at different
times and of different conditions: prominences even may be, and
doubtless are, involved. The lengths and intensities of the lines are
faithfully recorded in the photographs.

An examination of the eclipse photographs shows that the temperu:
ture of the most luminous vapours at the sun's limb is not far from
that produced by an electric spark of very high tension, the lines,
which we have seen to be enhanced on passing from the arc to such a
spark, being present.

The chromosphere, then, is certainly not the origin of the Fraun-
hofer lines, either as regards intensity or number. From the eye
observations made since 1868, there is ample evidence that the quiescent
chromosphere spectrum indicates a higher temperature than that at
which much of the most valid absorption takes place ; in other words,
the majority of the lines associated with lower temperature are pro-
duced above the level of the chromosphere, and hence the true reversing
layer, instead of being at the bottom of the chromosphere, as held by
some, is really above it.

* Proc. Soy. Soc., vol. xxxiv; p. 297.

IV.] THE SUN'S CHROMOSPHERE. 43

The eclipse photographs, however, at the same time afford evidence
by the relative lengths of some of the lower temperature lines that we
need not locate the region which produces the absorption indicated
by the Fraunhofer lines at any great height above the chromosphere.

I may say that for some time I was of opinion that in the sun many
of the darkest lines indicated absorptions high up in the atmosphere,
for the reason that the bright continuous spectrum of the lower levels
might have an important effect upon line absorption phenomena by
superposing radiation, and so diminishing the initial absorption. The
observations of the eclipses of 1893, 1896 and 1898, however, indicate
that this opinion is probably only strictly true when the strata of the
sun's atmosphere close above the photosphere are considered.

Let us next turn to the highest regions of the solar surroundings to
see if we can get any effective help from them.

In this matter we are dependent absolutely upon eclipses, and
certainly the phenomena observable when the so-called corona is
visible, full of awe and grandeur to all, are also full of precious teach-
ing to the .student of science. The corona varies like the spots and
prominences with the sun-spot period.

It happened that I was the only person that saw both the eclipse of
1871 at the maximum of the sun-spot period and that of 1878 at mini-
mum ; the corona of 1871 was as distinct from the corona of 1878 as any-
thing could be. In 1871 we got nothing but bright lines indicating the
presence of gases, namely hydrogen and another since provisionally
called coronium. In 1878 we got no bright lines at all ; so I then
stated that probably the changes in the chemistry and appearance of
the corona would be found to be dependent upon the sun-spot period,
and recent work has borne out that suggestion.

I have now specially to refer to the corona as observed and photo-
graphed in 1898 in India by means of the prismatic camera, remark-
ing that an important point in the use of the prismatic camera is that
it enables us to separate the spectrum of the corona from that of the
prominences.

One of the chief results obtained is the determination of the posi-
tion of several lines of probably more than one new gas, which, so far,
have not been recognised as existing on the earth.

Like the lowest hottest layer, for some reason or other, this upper
layer does not write ivs record among the lines which build up the
general spectrum.

Up to the employment of the prismatic camera insufficient atten-
tion had been directed to the fact that in observations made by an ordi-
nary spectroscope no true measure of the height to which the vapours

44 ' INORGANIC EVOLUTION. [CHAP. IV.

or gases extended above the sun could be obtained ; early observations,
in fact, showed the existence of glare between the observer and the
dark moon ; hence it must exist between us and the sun's surround-
ings.

The prismatic camera gets rid of the effects of this glare, and its
.results indicate that the effective absorbing layer — that namely, which
gives rise to the Fraunhofer lines — is much more restricted in thickness
than was to be gathered from the early observations.

We learn from the sun, then, that the absorption which defines its
ordinary spectrum is the absorption of a middle region, one shielded
both from the highest temperature of the lowest reaches of the atmo-
sphere where most tremendous changes are continually going on, and
from the external region where the temperature must be low, and
where the metallic vapours must condense.

This is the first great teaching of the test-spectrum. The next
chapter will deal with the second.

45

CHAP. V. — STELLAR ATMOSPHERES.

AFTER the laboratory work undertaken with the view of att3mpting to
find explanations of the various phenomena presented by the sun had
reached a certain stage, it became necessary to endeavour to get an
idea of the sun's place among the stars by a discussion of all I/he
existing spectroscopic observations which might throw light upon the
subject.

At that time a very large number of the most important lines,
both bright and dark, recorded in stellar spectra were of unknown
origin. The inquiry, therefore, in the case of all the hotter stars had
to do with the spectral lines as hieroglyphics, not as special chemical
representatives.

When I began the inquiry, the prevailing ideas were that the first
period of a star's life was one of the highest temperature, and that all
the differences observed were due to different stages of cooling having
been reached. With regard to the nebulae, they, it was imagined,
formed a different order of created things from the stars.

Passing over the old views, among them one that the nebulae were
holes in something dark, which enabled us to see something bright
beyond ; and another that they were composed of a fiery fluid, I may
say that not long ago they were supposed to be masses of gases only,
existing at a very high temperature ; and it was also suggested that
they, perchance, represented the residua in space left after all the stars
had been formed.

The upshot of this inquiry forms the subject matter of two com-
panion volumes,* so I need not dwell upon it in any detail here. But
it is necessary that I should state, as briefly as may ba, the results to
which the discussion of all the then available spectroscopic observations
led me.

All the observations were satisfied by the working hypothesis of
the evolution of all cosmical bodies from meteorites, the various stages
recorded by the spectra being brought about by the various conditions
which follow from the hypothesis.

The nebulas present us with the first stage. They are taken to be
sparse swarms of meteorites colliding together, and thus producing
their luminosity, which spectroscopically is found to be due to permanent

* The Meteoritic Hypothesis and The Sun's Place in Natiire. Maemillan.

46 INORGANIC EVOLUTION. [CHAP.

gases, hydrogen and the cleveita gases and carbon compounds driven
out of the meteorites as a result of the heat produced by the collisions ;
and to a less extent to the low temperature lines of some of the chemical
metallic elements known to exist in meteorites.

We have then to deal with the colliding particles of the swarm
and the permanent gases given off and filling the interspaces. The tem-
perature is relatively low ; since gases may glow at a low temperature
as well as at a high one, the temperature evidence depends upon the
presence of cool metallic lines and the absence of the enhanced ones.

The nebulae, then, are relatively cool collections of some of the per-
manent gases and of some cool metallic vapours, and both gases and
metals are precisely those I have referred to as writing their records
most visibly in stellar atmospheres.

If the nebulae are thus composed, they are bound to condense to
centres, however vast their initial proportions, however irregular ths
first distribution of the cosmic clouds which compose them. Each
meteorite, the motion of which is stopped by collisions, must at once
fall to the centre of gravity of the swarm.

Each pair of meteorites in collision puts us in mental possession of
what the final stage must be. We begin with a feeble absorption of
metallic vapours round each meteorite in collision ; the space between
the meteorites is filled with the permanent gases driven out further
afield, and having no power to condense. Hence dark metallic and
bright gas lines. As time goes on, the former must predominate, for
the whole swarm of meteorites will then form a gaseous sphere, with a
strongly heated centre, the light of which will be absorbed by the
exterior vapour.

As condensation goes on, the temperature at the centre of condensa-
tion always increasing, all the meteorites of the parent swarm in time
are driven into a state of gas. The meteoritic bombardment practically
now ceases for lack of material, and the future history of the mass of
gas is, speaking generally, that of a cooling body, the violent motions
in the atmosphere while condensation was going on now being replaced
by a relative calm, producing a quiescent reversing layer the observa-
tion of which alone enables us to define the temperature of the star.

The temperature- order of the group of stars with bright lines as
well as dark ones in their spectra, has been traced, and typical stars
indicating the spectral changes have been as carefully studied as those
in which absorption phenomena are visible alone, so that now there are
very few breaks in the line connecting the nebulas with the stars on the
verge of extinction.

We find ourselves here in the presence of minute details exhibiting
the workings of a law associated distinctly with temperature ; and

STELLAR ATMOSPHERES.

47

more than this, we are also in the presence of high temperature fur-
naces, entirely shielded by their vastness from the presence of those
distracting phenomena which we are" never free from in the most
perfect conditions of experiment we can get here.

Thanks to the spectroscope, the old guesses have now been replaced
by the result of a general inquiry, in which hundreds of thousands of

eu " '*'"•} '••''.''''<'•'.•'>. ••• -f

observations have been used, and for my part I do not think it prob-
able that the scheme of celestial evolution which I have sketched above

48 INORGANIC EVOLUTION. [CHAP,

and which is indicated in the accompanying temperature curve, will be
greatly changed in its essential points ; it rests upon so wide a basis of
induction.

When this view of celestial evolution was first formulated as the
result of the wide spectroscopic inquiry to which I have referred, most
of the lines in the nebulae, and in the stellar groups III, IV, and Vr
were of unknown origin; the groups were established by accepting
their presence as criteria, without any reference to chemistry. In the
lower groups I, II, and VI, the chemistry was obvious, and the identi-
fication of many metallic flutings made it clearer still.

When engaged later on, in 1893, in the classification of stars, accord-
ing to their photographic spectra* I came across two very important sets
of lines of unknown origin, one in the hottest stars, the other in stars
of intermediate temperature.

After the discovery of a terrestrial source of helium by Professor
Ramsay, I showed in a series of seven notes communicated to the Royal
Society,! May — September, 1895, that the cleveite gases, which I
obtained by the process of distillation, accounted to a very great
extent for the first set.

This result proved to be the key to the chemistry of groups III and
IV, which contains the hottest stars.

In 1897, in a series of three communications to the Royal
Society,! I pointed out that some of the other set of unknown lines in
the stars of intermediate temperature, taking a Cygni as an example,
were due to the enhanced spark lines of iron and other metals, the arc
lines being almost entirely absent.

The recent developments of this research, and the ultimate forma-
tion of a " test-spectrum," have been referred to in Chapter III. The
result of this has been to greatly strengthen the argument based upon
the first observations.

In the accompanying photograph, a comparison is shown between
the lines of a Cygni and the enhanced lines of the substances thrown
together to form the " test-spectrum." The extraordinary number of
coincidences is seen at a glance. The facts are as follows : —

The number of lines measured in the spectrum of o Cygni at
Kensington between \ 3798'! and A 4861 '6 is 307

Of these the number which approximately coincides with the
enhanced metallic lines so far observed is . .. .. ..120

* Phil. Trans., A, vol. clxxxiv, p. 675.

f 1st note, Proc. Roy. Soc., vol. Iviii, t>. 67; 2nd, ibid., vol. Iviii, p. 113 ; 3rd,
ibid., vol. Iviii, p. 116 ; 4th, ibid., vol. Iviii, p. 192 ; 5th, ibid., vol. Iviii, p. 193 ;.
6th, ibid., vol. lix, p. 4 ; 7th, ibid., vol. lix, p. 342.

J Proc. Roy. Soc., vol. Ix, p. 475; ibid., vol. hi, p. 148 ; ibid., vol. Ixi, p. 441.

STELLAR ATMOSPHERES.

The number of lines (excluding the hydrogen serie*) in a Cygiii of
intensity over 4 (the maximum being represented by 10) is . . 40

Of this number, the coincidences with enhanced metallic lines with
the dispersion employed amount to ..- .. .. .. .. 38

9

< Ci

<•< CO

cc

Ci

CO

c

< <M

pq

The lines of the stars of intermediate temperature, like a Cygni,
have long been recognised by the Harvard observers as well as by myself
as presenting great difficultie3.

E

50 INORGANIC EVOLUTION. [CHAP.

In 1893 I wrote as follows:* "With the exception of the K line,
the lines of hydrogen and the high temperature line of magnesium at
A4481, all the lines ma}^ be said to be at present of unknown origin.
Some of the lines fall near lines of iron, but the absence of the strongest
lines indicates that the close coincidences are probably accidental."
In the Harvard Spectra of Bright Stars, 1897, p. 5, the following words
occur, relating to the same stars : " This system of lines should perhaps
lie regarded as forming a separate class, as in the case of the Orion
lines, and should not be described as ' metallic,' as has just been done
in the absence of any more distinctive name."

It will be seen then that the second set of "unknown lines "has now
been as effectively disposed of by the determination of the enhanced
lines of the metallic elements as the first set was by the discovery of the
cleveite gases. The secrets of the " unknown lines " in the hottest
stars now stand revealed.

Now that the chemical story is so nearly complete, or at all events
so much more complete than it was, we are in a position to inquire
what the stars teach us concerning their chemistry ; but in the first
instance we must examine the origin of the information they afford us,
that is, amongst other things, we must study their absorbing conditions,
.and next their chemistry in relation to temperature.

With regard to the origin of the absorption phenomena, to which,
for the most part, our inquiries will be directed ; in the case of the
sun, we have a star so near us that we can examine the different parts of
its atmosphere, which we cannot do in the case of the more distant
stars.

We have seen in Chapter IV the facts with regard to the sun—
that the most valid absorbing layer occupies a certain region in the
atmosphere not high up, not at the bottom, but slightly above the
'bottom — that is the chromospheric — layer.

Now the spectrum of Arcturus resembles the spectrum of the sun
.almost line for line ; what is true for the sun therefore must be equally
true for Arcturus, which exactly resembles it. The next point we
have to consider is whether the absorption in stars generally, which the
spectrum indicates for us, takes place from top to bottom of the atmo-
sphere, or only in certain levels.

In many of these stars the atmosphere may be millions of miles
high. In each the chemical substances in the hottest and coldest
portions may be vastly different ; the region, therefore, in which this
absorption takes place, which spectroscopically enables us to dis-
criminate star from star, must be accurately known before we can
obtain the greatest amount of information from our inquiries.

* Phil. Trans ^ A, vol. clxxxir, p. 694.

V.] STELLAR ATMOSPHERES. 51

Assuming that the most valid absorbing vapours in any particular
stafr are all near one temperature, we can proceed to investigate the
origins of the spectrum lines by first getting a clue as to the probable
temperature from the extent of continuous spectrum, and then inquir-
ing into the presence or absence of the lines which are longest in the
spectra of various substances at that temperature. If, however, the
absorptions take place at different levels in the atmosphere of a star,
the proper spectrum of each substance to be thus investigated can only
be determined by a comparison of the stellar with the terrestrial lines
of the substance under varying temperature conditions.

TMs method of looking for the longest lines will fail in the case of
stars which are hotter than our hottest spark. In such case, therefore,
we must necessarily rely on a comparison with lines which, from our
study of the spectra at different temperatures, would most probably be
longest in the spectrum at a temperature higher than any at which
experiments can be carried on.

It is in connection with such an inquiry as this that the study of
the conditions of the sun's atmosphere is of supreme importance, that
is why I have devoted the previous chapter to it. It is obvious that a
knowledge of the solar conditions must be of the utmost value in
enabling us to apply a well-established series of facts, gathered in the
case of the star nearest to us, to the phenomena presented by the more
distant bodies.

By doing this we have obtained facts which suggest in what parts
of the atmosphere the absorption takes place which produces the
various phenomena on which the chemical classification can be based ;
these facts we are bound to accept in a discussion of the origin of
stellar absorption in the absence of evidence to the contrary. And we
are justified in extending these general conclusions to all the stars that
shine in the heavens. I go further than this, and say that in the
presence of such definite results, it is not philosophical to assume that
the absorption may take place at the bottom of the atmosphere of one
star, or at the top of the atmosphere of another. The onus probandi
rests upon those who hold such views.

So much then, in brief, for solar teachings in relation to the record
of the absorption of the lower parts of stellar atmospheres.

If we are justified in arguing from a star with a photosphere as
well developed as that of the sun to one in which it is in all probability
much less marked in consequence of a much higher temperature, then
we must consider that the absorptions which mark our the various star
groups are more conditioned by the temperatures of the absorbing
regions merely than by the thickness of the absorbing atmospheres, or
by the densities of the various vapours. Another consideration to be

E 2

52 INORGANIC EVOLUTION. [CHAP.

borne in mind is that if the atmospheres are in part composed of
condensable vapours, and not entirely of gases permanent at all stejlar
temperatures, condensation must always be going on outside at the
region of lowest temperature.

The absorption phenomena in stellar spectra are not identical at the
same mean temperature on the ascending and descending sides of the
curve, on account of the tremendous difference in the physical conditions.

In a condensing swarm, the centre of which is undergoing
meteoritic bombardment from all sides, there cannot be the equivalent
of the solar chromosphere ; the whole mass is made up of heterogeneous
vapours at different temperatures, and moving with different velocities
in different regions.

In a condensed swarm, of which we can take the sun as a type, all
action produced from without has practically ceased ; we get relatively
a quiet atmosphere and an orderly assortment of the vapours from top
to bottom, disturbed only by the fall of condensed metallic vapours.
But still, on the view that the differences in the spectra of the heavenly
bodies chiefly represent differences in degree of condensation and tem-
perature, there can be, cm fond, no great chemical difference between
bodies of increasing and bodies of decreasing temperature. Hence we
find at equal mean temperatures on opposite sides of the temperature
curve, this chemical similarity of the absorbing vapours proved by
many points of resemblance in the spectra, especially the identical
behaviour of the enhanced metallic and cleveite lines.

Now that the test-spectum has led us to such a very definite con-
clusion with regard to a Cygni and other stars resembling it, it is
necessary to tarn back to Chapter IV, in which the solar atmosphere
was discussed. It was pointed out what a marvellous resemblance there
was between the test-spectrum and the sun's chromosphere, photo-
graphed during the eclipse of 1898. If the spectra of the valid absorb-
ing atmosphere of a Cygni and of the sun's chromosphere resemble the
test-spectrum as they do, the atmospheres must resemble each other,,
both in chemistry and temperature.

Here, then, we have an almost undreamt-of opportunity of noting
the close connection between solar and stellar phenomena, not merely
in noting the identity of the action of the absorbing layers as we do
when we find the spectra of the sun, Arcturus and Capella, almost
identical, line for line, but in studying the relation of the absorbing
layer of one star to the underlying layer in another.

While we find, on the one hand, that the absorbing layer of the sun
is similar to those of Arcturus and Capella, we find, on the other, that
the spectrum of the sun's chromosphere resembles that of the reversing
layer of a Cygni. The " test-spectrum " fits them both.

V-]

STELLAR ATMOSPHERES.

Now the chromosphere by a well-known physical law must be
hotter than anything outside it, but we know that the reversing layer

o

T5

O

I
I

I

a

i §
2.2

lies outside it, therefore the reversing layer of a Cygni must be hotter
than the reversing layer of the sun.

In the chromosphere of 1898, the enhanced lines are all of greater
intensity than the corresponding Fraunhofer lines, and they are also

54 INORGANIC EVOLUTION. [CHAP. V.

relatively stronger, as referred to the arc lines, than they are in
the experimental spark. Hence, the incandescent iron vapour in the
chromosphere must be at a temperature at least as high as that of the
spark, and certainly higher than that of the iron vapour which is most
effective in the production of Fraunhofer lines.

The evidence is complete that the temperature in the reversing
layer of a Cygni is higher than that of the reversing layer of the sun.
What do we find ? Of lines disappearing we have the arc lines of iron,
calcium, magnesium, strontium, and so on, some thousands in number.
Of lines increasing in importance we have the small number represent-
ing the enhanced lines of iron, the lines of hydrogen, and some others
which we cannot at present associate with the name of any known
substance. Here, then, we get a series of phenomena which is simply
and sufficiently explained by the statement that on passing from the
temperature of the sun to a Cygni, among other changes brought
about, the complicated line spectrum of iron is giving way to a more
simple one consisting of the enhanced lines. Further inquiries show
that the other metallic spectra are behaving in the same way.

In passing from the absorbing layer of the sun to that of a Cygni,
then, we pass from the arc lines of the metallic elements to the enhanced
lines. Truly a most tremendous change which the test-spectrum puts
beyond all question. The significance of this will come later.

In the case of the sun, the enhanced test-spectrum was the only
one we could employ with advantage. But in the pase of the hottest
stars, stars that is, with the longest spectrum, we can go still further.
These are so much hotter than the sun, that they give us the oppor-
tunity of noting another break; really of employing another test-
spectrum, that afforded by the summation of the lines of hydrogen and
the cleveite gases.

As we have seen, the arc metallic lines give way to the enhanced
metallic lines in stars of intermediate temperature, like our sun and
a Cygni, so, in the hottest stars the enhanced metallic lines vanish
almost entirely, and give place to a spectrum almost purely gaseous.

To take iron as an example, for the sake of simplicity, it will be
seen then that the actual stellar phenomena might have been pre-
dicted up to a certain point, from a consideration of laboratory and
solar phenomena. But the stars carry us further than our predictions ;
we see the gradual increase of hydrogen and the cleveite gases. The
facts demonstrate that as temperature increases hydrogen increases,
and, together with the cleveite gases not obvious before, finally replaces
iron which has disappeared.

55

CHAP. VI. — THE CHEMISTRY OF THE STARS.

THE recent advances in our knowledge which have come from the
combination and interaction of solar, stellar and laboratory research,,
carried on by the aid of instruments of much greater power than those
formerly used, have given us a firm chemical hold on all the groups of
stars in my classification of them. These groups were established by
discussing sequences of lines before the origin of the lines had been
made out ; as I have already said, a series of hieroglyphics is now re-
placed by chemical facts ; and we can now study the chemistry of the
stars, as well as their order in a system of classification.

The first question which naturally arises is this : Do the chemical
elements make themselves visible indiscriminately in all the celestial
bodies, so that practically, from a chemical point of view, the bodies
appear to us of similar chemical constitution 1 This is not so.

From the spectra of those stars which resemble the sun, in that
they consist of an interior nucleus surrounded by an atmosphere which
absorbs the light of the nucleus, and which therefore we study by
means of this absorption ; it is to be gathered that the atmospheres of
some stars are chiefly gaseous, i.e., consisting of elements we recognise
as gases here, of others chiefly metallic, of others again mainly com-
posed of carbon or compounds of carbon.

Here then we have spectroscopically revealed the fact that there is
considerable variation in the chemical constituents which visibly build
up the stellar atmospheres.

This, though a general, is still an isolated statement. Can we con-
nect it with another ?

By means of one of the first principles of spectrum analysis referred
to in Chapter I, we know that the hotter a thing is the light of which
produces a continuous spectrum, the further does the- spectrum stretch
into the violet and ultra-violet.

Hence the hotter a star is, the further does its complete or con-
tinmus spectrum lengthen out towards the ultra-violet, and, cwteris
paribus, the less is it absorbed by cooler vapours in its atmosphere.

Now to deal with three of the main groups of stars, we find the
following very general result : —

Gaseous stars .. .. Longest spectrum.
Metallic stars . . . . Medium spectrum.
Carbon stars . . . . Shortest spectrum.

56 INORGANIC EVOLUTION. [CHAP.

We have now associated two different series of phenomena, and we
-are entitled to make the following general statement : —

G-aseous stars . . . . Highest temperature.
Metallic stars . . . . Medium temperature.
Carbon stars . . . . Lowest temperature.

Hence the differences in apparent chemical constitutions are asso-
ciated with differences of temperature.

This, then, is the result of our first inquiry into the existence of
the various chemical elements in the atmospheres of stars generally.
We get a great diversity, and we know that this diversity accompanies
changes of temperature. We also find that the sun, which we inde-
pendently know to be a cooling star, and Arcturus, are identical
chemically.

Can we associate with the two to which I have already called atten-
tion still a third class of facts 1

Laboratory work enables us to do this.

The cleveite gas spectrum and the spectrum of enhanced metallic
lines come to our help and enable us to get a step forwarder. In
studying the appearance of these lines in stellar spectra, we have a
third series of phenomena available, and we find that the results are
absolutely in harmony with what has gone before. Thus

Gaseous stars .. Highest temperature .. Strong cleveite gas and

faint enhanced lines.
rFeeble cleveite gas and

,.- , ,,. ,, ,. strung enhanced lines.

Metallic stars , , Medium temperature . . •< „ . .

No cleveite gas, and

«- strong arc lines.
Carbon stars . . Lowest temperature . . Faint arc lines.

It is clear now, not only that the spectral changes in stars are asso-
ciated with, or produced by, changes of temperature, but that the study
of the enhanced spark and the arc lines lands us in the possibility of a
rigorous stellar thermometry, such lines being more easy to observe
than the relative lengths of spectrum.

What then, is the chemical law 1 It is this. In the very hottest
stars we deal, speaking generally, with the gases hydrogen, helium,
asterium, and doubtless others still unknown, almost exclusively. At
the next lowest temperatures we find these gases being replaced by
metals in the state in which they are observed in our laboratories
when the most powerful jar-spark is employed. At a lower tempera-
ture still the gases almost disappear entirely, and the metals exist in
the state produced by the electric arc.

I said " speaking gencrall ," but we really can go further than this

VI.]

CHEMISTRY OF THE STARS.

57

general statement, and I next pass from the general to the particular,
and give the detailed results recently obtained in the case of stars as
hot or hotter than Arcturus — taking -Arcturus to represent the solar
temperature — in the light of the most recent work, some of which has
already been referred to in the preceding chapters.

Proto-metals.

With regard to the metals, the recent work on the enhanced lines
in the spectrum of metals, a Cygni* and the sun's chromosphere enables
us to deal with the lines observed at the highest temperature in the
spectra of the following substances : magnesium, calcium, iron, man-
ganese, nickel, chromium, titanium, copper, vanadium, strontium, sili-

cmm.

The untouched reproductions of photographs of the spectra of the
chromosphere and a Cygni, given on page 53, have already shown the
wonderful similarity which exists between these three spectra.

As we have to deal both with the arc and spark lines of these sub-
stances, for the sake of clearness I call the latter " proto-metallic " lines,
and consider the substances which produce them, obtained at the
highest available laboratory temperatures, " proto-metals," that is, a
finer form of the metal than that which produces the arc lines, corre-
sponding to the " meta-elements " imagined by Crookes.

The temperature ranges of the enhanced lines of these metals have
been investigated in various stars with the following results : —

Metal.

Eange of temperature
(upward series).

.Range of temperature
(downward series).

Magnesium
Calcium. .
Iron
Titanium
Manganese
Nickel ..
Chromium
Vanadium
Copper ..
Strontium

a Ursee Min. to y Argus
a Tauri to 7 Argus
a Tauri to £ Taurif
a Tauri to £ Tauri
a Ursse Min. to a Cygni
a TJrsoe Min. to a Cygni
a Ursa3 Min. to a Cygni
a Ursee Min. to a Cygni
I a Ursse Min. to a Cygni
a Tauri to a Cygni

a Eridani to Procyon.
o Eridani to Arcturus.
ft Persei to Arcturus.
ft Persei to Arcturus.
/3 Persei to Procyon.
ft Persei to Procyon.
7 Lyrae to Procyon.
Sirius to Procyon.
ft Persei to Procyon.
Sirius to Arcturus.

* Nature, vol. Ixix, p. 342.

•f This is one of the most extraordinary spectra which has been met with in the
Kensington series of photographs, as I have already pointed (Proc. Roy. Soc., vol.
Ixi, p. 184). While the lines of hydrogen are fairly sharp and not very broad,
many of the lines, especially those of the cleveite gases, are broadened almost into

58 INORGANIC EVOLUTION. [CHAP.

The enhanced lines of the above substances seem to account for
almost all of the more marked lines in a Cygni. It is on this ground
that I have investigated their behaviour in other stars before waiting
for the results of the complete inquiry. Another reason has been that,
although in addition to the enhanced lines of the metals shown in the
foregoing table, those of barium, cadmium, molybdenum, lanthanum,
antimony, lead, palladium, tantalum, erbium and yttrium, tungsten,
cerium, uranium, cobalt, and bismuth have already been investigated
with lower dispersion, and a spark obtained with the use of a much
less jar capacity, so far as I have no certainty that any of these sub-
stances exist in the reversing layers of stars of intermediate tem-
perature.

The temperature ranges of the arc lines of some of the metals have
also been investigated, and the results are shown in the following
table :

Metal.

Kange of temperature
(upward series).

Range of temperature
(downward series).

Iron
Calcium . . . .
Manganese

a Tauri to a Cygni.
a Tauri to a Ursae Min.
a Tauri to a Urssc Min.

•
a Canis Majoris to Arcturus.
a Canis Majoris to Arcturus.
a Canis Majoris to Arcturus.

So much, then, for the metals. I now turn to the gases.

Proto-liydrogen.

Some little time ago Professor Pickering, of the Harvard Observa-
tory, found on examining the spectra of the southern stars, that one of
them on the poop (Lat. Puppis), hence called f Puppis, of the ship which
forms the constellation Argo, contained a system of lines not hitherto
recognised, and he naturally concluded that it indicated a new
element.* On further inquiry he found reason to suppose that this
new series was in some way connected with hydrogen, since the lines
occupied the same positions as those computed from the same formula
and constants from which the ordinary series of hydrogen was calcu-

invisibility. On the meteoritic hypothesis this is explained by the great differences
of velocity and direction of the meteoritic streams, the special broadening of the
lines of the cleveite gases indicating that these gases are chiefly concerned in dis-
turbances at high temperatures.

On account of the indistinctness of many of its lines, £ Tauri is omitted from
the present discussion.

* See Astrophysical Journal, vol. iv. p. 369, and vol. 5, p. 95.

V..]

CHEMISTRY OF THE STARS.

kited, the only difference in the employment of the formula being that
even values of n were used instead of odd values.*

Professors Pickering and Kayser both concede that this new form
of hydrogen is due most probably to a high temperature, and Professor
Kayser expressly states "that this series has never been observed
before can perhaps be explained by insufficient temperature in our
Geissler tubes and most of the stars."

If, as suggested both by Professor Kayser arid myself, this new
series and the one previously known are probably of the subordinate
type, the principal series of hydrogen is still beyond our ken, unless
indeed one of the still " unknown " lines represents it, as suggested
by Professor Eydberg. Another possibility is that, even in the hottest
stars so far considered, the temperature is not high enough to allow
its molecule to exist uncombined.

On the view that the new series of probable hydrogen lines in
£ Puppis represents the effect of a transcendental temperature, an
attempt has been made to produce this spectrum in the laboratory.
In the high-tension spark in hydrogen at atmospheric pressure the
ordinary series is represented by broad lines. The use of the spark
with large jars in vacuum tubes results in the partial fusion of the
glass and the appearance of lines which have been traced to silicium,
but the new series has not yet been observed.

In his first communication Professor Pickering mentions lines at
4698, 4652, 4620, and 4505, but he does not refer to them in his-
second paper, which has special reference to the new series. The line
4505 was at first taken to be one of the components of the new series,,
but this seems to have been subsequently superseded by the employ-
ment of the line about 4544, which agrees better both as regards-
intensity and the calculated position 4543'6.

* The two series are as follows : —
Old Series.

New Series.

».

6

8

Computed.
6563 -0
4861 -5

Obserred. n.
6563 -0 5
4861-5 7

Computed.
10128 -1
5413 -9

Observed
(means).

10

4340 -6

4340 7

9

4543 -6

—

12

4101 -9

4101 -8

11

4201 -7

4200-4

14

3970 -2

3970 -2

13

4027 -4

4026 '8

16

3889 -2

3889 -1 15

3925 -2

3924 -7

18

3835 -5

3835 -5 17

3859 -8

3858 -7

20

3798 -0

3798-1 19

3815 -2

3815 -9

21

3783 -4

3783 -4

These figures are taken from Professor Pickering's article in Astrophysical
Journal, vol. v, p. 93. See also Kayser's article on p. 95 of the same journal.

-60 INORGANIC EVOLUTION. [CHAP.

As this new hydrogen series seems to bear the same relation to the
well-known one as the pro to-metallic lines bear to the metallic, I
call the gas which produces it proto-hydrogen for the sake of clearness.

The new series of lines has been found in the spectra of £, e, S, and
K Orionis photographed at Kensington in 1892.

Professor Pickering himself has since found this system of lines in
other stars than £ Puppis, 29 Canis Majoris among them, and Mr.
McClean, in his admirable work on the brightest stars of the southern
hemisphere, has obtained photographs of the spectrum of 7 Argus, in
which the new series appears.

From a discussion of these stars in relation to the others photo-
graphed, there can be little doubt that we are here face to face with
the very hottest stars so far known : and that the new series of hydro-
gen lines represents one among the last stages of chemical simplifica-
tion so far within our ken.

We are, therefore, now in a better position to determine the rela-
tion of this new gas to other gases, both known and unknown, appear-
ing in stars of nearly equal temperature.

Other New Lines.

But even with our present knowledge of stellar spectra we find
that in relation to the hottest stars there are still some gaps in our
chemical knowledge ; not only is this so, but have we any right to
assume, taking into account the limitations of our means of observa-
tion and of the strict limitation of our observations to the relatively
small part of space nearest us, enormous though it is, that we are
as yet really in touch with the highest stellar temperatures 2

Again, we cannot be certain that the small number of stars as yet
studied puts us in presence of the highest stellar temperatures. Those
stars which apparently are at the very apex of the temperature curve
are involved in unknown lines, and require a special study.

Two typical unknown lines have wave-lengths at 4089*2 and
4649'2,* and besides these three other unknown lines occur in 7 Argus.

As these most probably reveal still undiscovered gases, I include
them in the following table showing the limits of stellar temperature
to which the various known and unknown lines, probably of gaseous
origin, extend.

Mr. McClean has stated that certain of the oxygen lines (amongst
which is the strong triplet at XX 407(H, 4072*4, and 4076-3) appear in
the spectrum of j3 Crucis and other stars of nearly equal temperature.
My own observations, so far as they have gone, tend to confirm this

* Proc. Roy, Soc., vol. Ixii, p. 52.

VI.]

CHEMISTRY OF THE STARS.

61

Origin.

X of chief lines.

Range in ascending
series of stars.

Range in descending
series of stars.

f 44571

Unknown

4451 L

[3876)

Seen only in

7 Argus.

Hydrogen
'(new)

f 4544 -0 1
[4200 -4 J

£ Orionis to 7 Argus

No stars available.

Unknown

4649 -2

a Crucis to £ Orionis.

o Eridani.

Helium ..

f 4471 '6 \
1 4026 -3 J

Rigel to 7 Argus

a Eridani to 7 Lyrse.

Asterium

f 43881
14009 f

Rigel to 7 Argus

a Eridani to 7 Lyne.

Hydrogen

Complete series

Aldebaran to 7 Argus

o Eridani to Arcturus.

view ; but other photographs and more laboratory work are needed
to explain certain changes of intensity which have been observed^
The lines attributed by Mr. McClean to oxygen have been noted
between a Crucis and £ Orionis in the upward series, and in stars at
about the a Eridani stage of temperature in the downward series.

There is evidence that the strongest lines of nitrogen at X 3995*2
and X 4630-9 make their appearance in stars at about the temperature
of a Crucis. These lines appear from Rigel to f Orionis in the upward
series, and are present in the stars at the a Eridani stage in the down-
ward.

I pointed out many years ago* that at high temperatures the
fiutings of carbon in the violet are replaced by a line at X 4267-5..
There is a line at this wave-length in the spectra of stars ranging in
temperature from that of Rigel to £ Orionis on the up side, and from
a Eridani to ft Persei on the downside of the temperature curve.

There is no known line of gases or metals to which this line can be
assigned. It is probable, therefore, that carbon exists in stars of
the same temperature as that at which oxygen and nitrogen have been
traced.

Two lines in the spectrum of silicium (X 4128-5 and X 4131-5) have
been traced in stars between the temperatures of a Ursse Min. and
a Crucis in the upward series, and between those of a Eridani and
Procyon on the downward.

The accompanying map shows the facts relating to stars as hot as,,
or hotter than, the sun, as we know them at present.

Description of Map.

The map is arranged on the following plan. The temperature of
the sun and Arcturus forms the lowest stage. The upper limit is,

* Proc. Hoy. Soc., vol. xxx, p. 461.

62

INORGANIC EVOLUTIOX.

[CHAP.

ftSB
40094
4481-3

OCTAURI.

I

•

o
7

Cft

^

acRucis. |

gORIONIS.

i

' Unknown.
Proto- hydrogen
Asterium.
Helium
Proto-magnesium.
Hydrogen.
;Prdtb-calcium.
Unknown.
Unknown.
Oxygen.
Nitrogen.
Carbon.
Silicium.
Proto-iron.
Proto-titomium.
Proto- copper
Proto- manganese
Proto- nickel.
Proto -chromium
Proto-vanadium.
Proto-stronllum.
Iron.
Calcium.
Manganese .

39338
40892

39952
42675

441285
441315

40330

=

=

—

^—

aERIDAN

AAAA.[\
*l'f 'TTV

4556-1
43444
40674
45888
40539

•

• —

— «MMM

m
m

r-

(404.59
lAOR^-R

(4071 £

42269

•§•••

•M— —

(4030 9
j 4033 2

j 4034 6

§

0:

aCANISMIN

aCANIS MAJ,

FIG. 27.- — Map of chemical substances present in stars of different temperatures.

defined by 7 Argus, the hottest star so far known. On the left the
stars named are those of increasing temperature, on the right those of
decreasing temperature. Those on the same horizon represent equal
mean temperatures so far as the cleveite gas and enhanced lines help

VI.] CHEMISTRY OF THE STARS. 63

us to determine them. The blank spaces indicate that so far no star
has been photographed in the spectrum of which the enhanced lines
exactly match those on the opposite side.

The names of the various chemical substances included in the dis-
cussion are given at the top. I have retained the prefix " proto " to
that condition of each metallic vapour which gives us the enhanced
lines alone, and I have added it to that form of hydrogen seen only in
the hottest stars.

The behaviour of the most typical line of each chemical substance
is indicated by a double line looped at the top at its highest range.
The length and varying thickness of the lines in stars on both sides of
the temperature curve are derived from the observed appearance and
intensity of the lines, noted in the different stars.

The wave-lengths of the lines discussed are shown at the bottom of
the map.

Details of Changes observed.

The facts embodied in the map present to us the spectral changes
noted in stars of Groups III, IV, and V of my classification,* and are
a result of a more general inquiry than those referred to in my pre-
vious papers f the origins of a very considerable number of stellar
lines having since then been traced to enhanced lines of metals and to
known gases.

It will be seen that this more general inquiry entirely justifies the
prior statement J that the metallic lines are thickest in stars increasing
their temperature, and the hydrogen lines thickest in stars decreasing
their temperature, in other words, on the opposite arms of the tempera-
ture curve. I have already stated a possible explanation^

It will be observed that, so far, I have not been able to find stellar
spectra on the downward side corresponding to those of 7 Argus and
£ Orionis ; but it is more than probable that near the apex of the curve
only a small change will be observed ; their default, therefore, is of less
consequence than it might have been.

The same remark applies to a Cygni and Sirius : but here it is cer-
tain that the differences in the relative intensities of the gaseous and
enhanced lines will be considerable, judging from what happens above
and below the heat stages represented by them.

The stars used in the discussion give us very definite results, show-
ing that the various chemical forms are introduced at six very distinct
heat levels.

* Proc. Eoy. Soc., vol. xliii, p. 117 (1887).

f Proc. Roy. Soc., vol. xliv, p. 1 (18S£) ; ibid., vol. xlv, p. 380 (1889) ; PHI.
Trans., A., 184, (1893), p. 725.

J Proc. Eoy. Soc., vol. Ixi, p. 182.
§ Proc. Boy. Soc., vol. Ixi, p. 183.

64 INORGANIC EVOLUTION. [CHAP.

The Temperature Ranges.

I next proceed to make some remarks upon the series of facts now
for the first time brought together; it must, however, be borne in
mind that all the chemical elements and all parts of the spectrum have
not yet been included in the survey.

The facts indicate individual peculiarities ; some chemical forms
appear to be longer lived than others, and, further, the important
spectral changes in the case of different substances do not occur at
the same temperature.

(1) Hydrogen appears throughout both series of stars from top to
bottom. Proto-magnesium and proto-calcium follow suit very nearly ;
but the highest intensity of the former is reached at the stage repre-
sented by a Cygni, and of the latter at the solar temperature represented
by a Tauri and Arc turns.

(2) With the above exceptions all the chemical forms so far traced
are relatively short-lived.

This is the first important differentiation. In the light of (1) we
are justified in assuming that tho substances in (2) would be visible in
the stellar reversing layers if they were there.

(3) In the stars of higher temperatures we deal generally with
gases. Below the stages represented by /3 Orionis and y Lyra? we
deal with proto-metals and metals, hydrogen being the only exception.

(4) The proto-metals make their appearance at about the same heat-
level at which the gases (with carbon), always excepting hydrogen,
begin to die out.

This is the second important differentiation. It is interesting to
notice the distinct difference of behaviour of carbon and silicium in the
descending series ; the former goes through the same stages as oxygen
and nitrogen, the latter behaves like the proto-metals.

(5) With the exception of iron the metals, as contra-distinguished
from the proto-metals, only make their appearance in stars at and
below the heat level of Sirius.

This is the third important differentiation. It is accompanied with
a notable diminution of hydrogen and proto-magnesium, and with an
increase of proto-calcium ; indeed, the latter seems generally to vary
inversely with the hydrogen.

The question arises whether the order of visibility at reduced tem-
peratures now indicated does not explain the absence of proto-hydro-
gen, oxygen, and nitrogen from the spectra of the sun and nebulae ; the
metals present in, and the absence of quartz from, meteorites, and the
similarity of the gaseous products obtained from meteorites and metals,
native and other, in vacuo at high temperatures.

VI.] THE CHEMISTRY OF THE STARS. 65

The Chemistry of the Cooler Stars. .

I have shown, on page 57, how the discovery of new lines in the
spectra of the metallic elements by using the most powerful induction
coil in existence has put us in possession of the chemistry of stars of
intermediate temperature ; and, further, how the discovery of the
cleveite gases has helped us in tracing the origins of very many lines
of the hotter stars.

Our knowledge of the chemistry of the cooler stars is little short of
marvellous ; we have two distinct groups of coolest ones, the evidence
of their much lower temperature being the shortness of their spectra.
In one of these groups we deal with absorption alone, as in those
already considered; we find an important break in the phenomena
observed ; helium, hydrogen, and the enhanced lines of metals have
practically disappeared, and we deal with metallic arc lines and carbon
absorption chiefly.

But the other group of coolest stars presents us with quite new
phenomena. We no longer deal with absorption alone, but accompany-
ing it we have radiation, so that the spectra contain both dark lines
and flutings and bright ones. Now such spectra are visible in the case
of new stars, as they are called, the ephemera of the skies, which may
be said to exist only for an instant relatively. In the case of these
bodies, when the disturbance which gives rise to their sudden appear-
ance has ceased, we find their places occupied by nebulae ; we cannot,
therefore, be dealing here with stars like the sun, which has already
taken some millions of years to slowly cool, and requires more millions
to complete the process into invisibility.

Hence in this class of coolest " stars " we are obviously dealing with
swarms of meteorites, the condensation of which has scarcely com-
menced, and hence it is that this class provides us with more " variable
stars " than any other.

66

CHAPTER VII. — A CHEMICAL CLASSIFICATION OF STARS.

IN the attempts made to classify the stars by means of their spectra,
from Eutherfurd's time to quite recently, the various criteria selected
were necessarily for the most part of unknown origin ; with the excep-
tion of hydrogen, calcium, iron and carbon, in the main, chemical
origins could not be assigned with certainty to the spectral lines.
Hence the various groups defined by the behaviour of unknown lines
were referred to by numbers, and as the views of those employed in
the work of classifying differed widely as to the sequence of the
phenomena observed, the numerical sequences vary very considerably, so
that any co-ordination becomes difficult and confusing.

The recent work referred to in the last chapter has thrown such a
flood of light on the chemistry of the stars that most definite chemical
groupings can now be established, and the object of the present
chapter is to give an account of the general scheme of classification in
which they are employed, which I have recently proposed.

The fact that most of the important lines in the photographic region
of the stellar spectra have now been traced to their origins renders
this step desirable, although many of the chemical elements still remain
to be completely investigated from the stellar point of view.

The scheme is based upon a minute inquiry into the varying inten-
sities, in the different stars, of the lines and flutings of the under-
mentioned substances : —

Certain unknown elements (probably gaseous, unless their lines
represent " principal series ") in the hottest stars, and the new form of
hydrogen discovered by Professor Pickering (which I term " proto-
hydrogen " for the sake of clearness), hydrogen, helium, asterium,
calcium, magnesium, oxygen, nitrogen, carbon, silicium, iron, titanium,
copper, manganese, nickel, chromium, vanadium, strontium ; the
spectra being observed at the highest available spark temperatures.
The lines thus observed I term "enhanced" lines, and I distinguished the
kind of vapour which produces them by the affix " proto," e.g., proto-
magnesium, for the sake of clearness.*

# Boy. Soc. Proc., vol. Ixir, p. 398.

CHAP. VII.] A CHEMICAL CLASSIFICATION OF STARS. 07

Iron, calcium, and manganese at arc temperatures.

Carbon (flutings) at arc temperatures.

Manganese and iron (flutings) at a still lower temperature.

In the last chapter I stated the results arrived at recently with
regard to the appearances of the lines of the above substances in stars
of different temperatures, and the definitions of the different groups or
genera to be subsequently given are based upon the map given
on page 62, together with more minute inquiries on certain additional
points the examination into which was suggested as the work went
on.

So far as the inquiry has .ii present gone, the various most salient
differences to be taken advantage of for grouping purposes are repre-
sented in the following stars, the information being derived from the
researches of Professor Pickering* and Mr. McClean,f as well as from
the Kensington series of photographs.

Hottest Stars.

Two stars in the constellation Argo (( Puppis and y Argus }).

Alnitam (e Orionis). This is a star in the belt of Orion shown on
maps as Alnilam. Dr. Budge has been good enough to make inquiries
for me, which show the change of word to have been brought about by
a transcriber's error, and that the meaning of the Arabic word is "a.
belt of spheres or pearls."

Stars of intermediate Temperature.

Ascending Series.

/3 Crucis.
£ Tauri.
Rigel.
a Cygni.

[ 1

Polaris.

Aldebaran.

Descending Series*

Achernar.

Algol.

Markab.

[. ]

Sirius.

Proeyon.

Arcturus.

* Astro-pk^s.Journ., vol. v. p. 92, 1897.

t Spectra of Southern Stars.

£ The spectrum of this star contains bright line?, but -when these occur with
dark lines, the latter alone have to te considered for purposes of chemical classi-
fication.

F 2

68

INORGANIC EVOLUTION.

[CHAP. vii.

Stars of lowest Temperature.
Ascending Series. Descending

Antares, one of the brightest
stars in Duner's Catalogue of
Class IILr.*

[Nebulse.]

19 Piscium, one of the brightest
stars in Duner's Catalogue of
Class lllb.

[Dark Stars.]

In order to make quite clear that both an ascending and a
descending series must be taken into account, I give herewith two
photographs showing the phenomena observed on both sides of the
temperature curve in reversing layers of stars of nearly equal mean
temperatures, as determined by the enhanced lines. The stars in
question are : —

Sirius (descending),
a Cygni (ascending).

Procyon (descending). "1 -p.^ Og
y Cygni (ascending). ]

The main differences to which I wish to draw attention are the very
different intensities of the hydrogen lines in Sirius and a Cygni, and the
difference in the width and intensities of the proto-metallic and metallic
lines in Procyon and y Cygni. These differences, so significant from a
classification point of view, were first indicated in a communication to
the Eoyal Society in 1887f, and the progress of the work on these lines
has shown how important they are. I have based the group — or generic
—words upon the following considerations.

As we now know beyond all question that a series of geological
strata from the most ancient to the most recent brings us in presence
of different organic forms, of which the most recent are the most com-
plex ; is it possible that the many sharp changes of spectra observed
in a series of stars from the highest temperature to the lowest, bring us
in presence of a series of chemical forms which become more complex
as the temperature is reduced 1 If so, we are in the stars studying
the actual facts relating to the workings of inorganic evolution on
parallel lines to those which have already been made available in the
case of organic evolution. I shall discuss this matter later.

In the meantime, regarding the typical stars as the equivalents of
ihe typical strata, such as the Cambrian, Silurian, &c., it is convenient

* Sur les etoiles a spectres de la troisnjine elasse.
f Proc Roy. Soe., vol. xliii, p. 145.

FIG. 28.

. 2H.

of-

rd

12 -s^

5 O
O

2

O
>-

u

5K,

70 INORGANIC EVOLUTION. [CHAP

that the form of the words used to define them should be common to
both ; hence I suggest an adjectival form ending in ian. If the typical
star is the brightest in a constellation, I use its Arabic name as root ; if
the typical star is not the brightest, I use the name of the constellation.
The desideratum referred has to a certain extent determined the
choice of stars where many were available. I have to express my
great obligations to Dr. Murray for help generously afforded in the
consideration of some of the questions thus raised. The table runs as
follows : —

CLASSIFICATION OF STARS INTO GENERA DEPENDING UPON THEIR
CHEMISTRY AND TEMPERATURE.

Highest temperature, simplest chemistry.

Argonian.
Alnitamian.

Crucian.
Ǥ Taurian.
i Bigelian.
g, Cygnian.

1 Polarian.
^ Aldebarian.
Antarian.

Achernian.

Algolian.

Markabian.

Sirian.
Procyonian.
Arcturian.
Piscian.

The cheirical definitions of the various groups or genera are as
follows : —

DEFINITIONS OF STELLAR GENERA.

Argonian.

Predominant. — Hydrogen and proto-hydrogen.
Fainter. — Helium, unknown (\4451, 4457), proto-niagnesium,
proto-calcium, asteriuni.

Alnitamian.

Predominant. — Hydrogen, helium, proto-silicium, unknown
(A 4649-2).

Fainter. — Asterium, proto-hydrogen, proto-magiiesiuni, proto-
caleium, oxygen, nitrogen, carbon.

C Crucian. \ Achernian.

Predominant. — Hydrogen, he- \ Same as Crucian.
Hum, asterium, oxygen, nitrogen,
carbon.

Fainter. — Proto -magnesium,
proto-calcium, proto-silicium, un-
known (A 4649'2), silicium.

VII.]

A CHEMICAL CLASSIFICATION OF STARS.

71

Taurian.

Predominant. — Hydrogen, he-
lium, proto-magnesium. asterium.

Fainter. — Proto-calcium, sili-
ciuiu, nitrogen, carbon, oxygen,
proto-iron, proto-titanium.

Rigelian.

Predominant. — Hydrogen, proto-
calcium, proto-magnesium, helium,
silicium.

Fainter. — Asterium, proto-iron,
nitrogen, carbon, proto-titanium.

Cygnian.

Predominant. — Hydrogen, proto-
calcium, proto magnesium, proto-
iron, silicium, proto-titanium,
proto-copper, proto-chromium.

Fainter. — Proto-nickel, proto-
vanadium, proto-manganese, proto-
strontium, iron (arc).

Polarian.

Predominant. — Proto-calcium,
proto-titanium. hydrogen, proto-
magnesium, proto-iron, and arc
lines of calcium, iron, and manga-

Fainter. — The other proto-metals
and metals occurring in the Sirian
genus.

Aldebarian.

Predominant. — Proto-calcium, arc
lines of iron, calcium, and manganese,
proto -strontium, hydrogen.

Fainter. — Proto iron and proto-tita-
nium.

Antarian.

Predominant. — Flutings of manga-
nese.

Fainter. — Arc lines of metallic ele-
ments.

Algolian.

Predominant.— Hydrogen, proto-
magnesium, proto- calcium, helium,
silicium.

Fainter. — Proto-iron, asterium,
carbon, proto-titanium, proto-cop-
per, proto-manganese, proto-nickel.

MarJcabian.

Predominant. — Hydrogen, proto-
calcium, proto-magnesium, sili-
cium.

Fainter. — Proto-iron, helium,
asterium, proto-titanium, proto-
copper, proto-manganese, proto-
nickel, proto-chromium.

Sirian.

Predominant. — Hydrogen, proto-
calcium, proto-magnesium, proto-
iron, silicium.

Fainter. — The lines of the other
proto-metals and the arc lines of
iron, calcium; and manganese.

Procyonian.
Same as Polarian.

Arcturian.
Same as Aldebarian.

Pis dan.

Predominant. — Flutings of carbon.
Fainter. — Arc lines of metallic ele-
ments.

72 INORGANIC EVOLUTION. [CHAP VI I.

We may take for granted that as time goes on new intermediate
genera will have to be established; the proposed classification lends
itself conveniently to this, as there are no numerical relations to be
disturbed.

A still more general chemical classification is the following, it being.
understood that in it only the most predominant chemical features are
considered, and that there is no sharp line of separation between these
larger groups. The peculiar position of calcium and magnesium renders-
this caveat the more necessary.

CLASSIFICATION OF STARS.
Highest temperature.

Gaseous stars

r Proto-hydrogen stars . . . ( ^rgoman.
J I Almtamian.

f Crucian.

*» Cleveite-gas stars < „

I Taunan.

. TEigelian.

Proto-metallic stars 1 Cygnian

Metallic stars fPolarian.

I Aldebarian.
Stars with fluted spectra Antarian.

Lowest temperature.

Achernian.

Algolian.

Markabian.

Sirian.
Procyonian.
Arcturian.
Piscian.

The detailed chemical facts to be gathered from the definitions of
the several genera indicate many important differences between the
order of appearance of the chemical substances in the atmospheres of the
stars and that suggested by the hypothetical " periodic law." I shall
refer to this point later on.

BOOK III.— THE DISSOCIATION HYPOTHESIS.
CHAPTER VIII. — RECENT OPINION.

WHEN stating in Chapter II some of the difficulties encountered by
the early workers in spectrum analysis who found it impossible to
reconcile the facts which the new method of work was accumulating
with the then received chemical view, I pointed out that as early as
1873 I had suggested that many of our difficulties would vanish if it
were conceded that the " atoms " of the chemist were broken up, or
dissociated, into finer forms by the high temperatures necessarily
employed in the new method of investigation.

The year 1873 was 27 years ago ; I propose, therefore, to briefly
refer, as judicially as I can, to the recent state of opinion on this
subject, or rather on some of the main points of it.

Only some of the views I had brought forward from time to time
have received general acceptance, those include the breaking up of the
solid metal giving (from whatever cause) a continuous spectrum into
smaller molecular groupings giving fluted and line spectra.

My view as to the subsequent dissociation of molecules, when once
the line spectrum stage has been reached, was still rejected by many.
For myself, I am not surprised at this. In a question of such tran-
scendental importance, caution must be redoubled; an absence of
work and expression of opinion in such a line of inquiry with questions
of pure science only involved, is almost inherent to the nature of the
investigations. The chemist has little interest in an appeal to celestial
phenomena, and astronomers do not generally concern themselves with
chemistry. The region investigated by the chemist is a low tem-
perature region dominated by monatomic and polyatomic molecules.
The region I have chiefly investigated is a high temperature region, in
which mercury gives us the same phenomena as manganese. In
short, the changes with which spectrum analysis has to do take place
at a far higher temperature level than that employed in ordinary
chemical work, and hence probably it is that I can only refer to one
chemical experiment bearing on the subject.

It is important, however, to point out that in cases where the two
regions overlap, vapour density determinations and other work have
been in harmony with the spectroscopic results, e.g., the changed

74 INORGANIC EVOLUTION. [CHAP.

density of iodine at changed temperatures and with a change in
spectrum.

The specific gravity of iodine vapour was found by Deville and
Troost to be 8'72 (air = 1), which corresponds to the density 125-9,
proving that the molecule or two volumes of iodine gas weighs
126-53 x 2 = 253-06. When iodine vapour is heated to 700° its specific-
gravity begins to diminish until at higher temperatures it becomes
constant, and is half that at 700°, the vapour consisting of free
atoms.*

Another, but less direct, argument in favour of dissociation,
independently of the changes in the intensities of the lines, was based
upon some observations I had made in an attempt to work out a
spectroscopic method for the detection of impurities. I noted the
presence of what I termed "basic lines," that is, short lines which
remained common to two or more spectra, after " long lines " had
been eliminated as being due to impurities.

I now refer to these different points seriatim.

Flutings represent Vibrations of Complex Molecules.

I take the change of the continuous spectrum successively into
flutings and lines first, and in justification of the statement that in
this matter my view is now generally accepted, I give the following
quotations from Schuster and Eder and Yalenta : —

" That the discontinuous spectra of different orders (line and band
spectra) are due to different molecular combination I consider to be
pretty well established, and analogy has led me (and Mr. Lockyer
before me) to explain the continuous spectra by the same cause ; for
the change of the continuous spectrum to the line or band spectrum
takes place in exactly the same way as the change of spectra of
different orders into each other."!

"Spater fiihrte Lockyer weiter aus, dass die Gase, solange ihre
Molekiile aus mehreren Atomen zerfallen, Linienspectren geberi
miissen. Diese Anschauung wurde seither ziemlich allgemein accep-
tirt/'|

The question of flutings was early conceded generally, but special
exceptions were made, carbon furnishes one instance.

Messrs. Liveing and Dewar in 187 9§ objected to my hypothesis,

V Viet. Meyer, Ser. DeutscTi. CJiem. Ges., vol. xiii, pp. 394, 1010, 1103; Meier
and grafts, Compt. Rend., rol. xc, p. 690; rol. xcii, p. 39.

f \chuster, Phil. Trans., 1879, Part I, vol. clxix, p. 39.

J E£er and Valenta, DenJcschriften der Jcaiserlichen AJcademie der Wissen-
schaften, Wien, vol. Ixi, p. 426, 1894.

§ Froc. Hoi,'. Soc., vol. xxx, p. 508.

VIII.] RECENT OPINION. 75

that the sets of carbon fiutings in the green represent molecular
groupings of that substance other than that (or those) which gives us the
lino spectrum, as gratuitous. I showed that the flutings, which Messrs.
Liveing and Dewar ascribed to a hydrocarbon, were present in the
spectrum of tetrachloride of carbon which gave no trace of hydrogen,
This experiment at first gave them no reason to modify their con-
clusion, but later they repeated and endorsed it, and finally admitted
that "the spectrum of the flame of hydrocarbons is not necessarily
connected with the presence of hydrogen,"* and so far as I can under-
stand their paper they seem to accept the idea of different molecular
groupings, which they began by characterising as " gratuitous."

The Complexity of the Line Spectrum.

With regard to the view that the line spectrum integrates for us
the vibrations of several sets of molecules, as I have already stated
this was not accepted. The number of objections is legion, and it is
impossible to refer* to all of them here. But, at the same time, the
opinion of some of those workers who have approached the subject
from both points of view was, I think, coming round to my side, and I
shall briefly refer to one or two instances.

Attention has recently been drawn to the variations in the appear-
ance of the magnesium lines in the celestial bodies by Dr. Scheiner, of
the Potsdam Observatory, who is not apparently acquainted with my
work of 1879 ; he, however, accepts the idea that the variations furnish
us with a precise indication of stellar temperature,! and he is now
employing it in the work of the observatory .J

* Proc. Roy. Soc., vol. xxxiv, p. 423.

t Astronomical Spectroscope, Frost's Translation, p. viii.

£ Dr. Scheiner points out that in the spectra of nearly all stars of Class Ta
(Group IV) the line at 4481 " generally appears as a broad line — in some spectra
as strong as the hydrogen lines — but its intensity decreases just in proportion as
the number of lines in the stellar spectrum increases, so that it is hardly of the
average intensity in the solar spectrum, or other spectra of type Ha, and the
author is unable to detect it in the spectrum of a Orionis." My prior work, dating
from 1879, being probably unknown to Dr. Scheiner, Messrs. Liveing and Dewar
are credited with the discovery of the peculiar behaviour of this line in laboratory
experiments, and it is added that " the dependence of the line upon the temperature
thus readily suggests that the temperature of the absorbing vapours upon the stars
of Class Ilia (Group II) is something like that of the electric arc, while that of
the stars of Class Ha is higher, and that of stars of Class la is at least as high as
the temperature of the high-tension spark from a Leyden jar. This view receives
striking confirmation in the precisely opposite behaviour of the magnesium line at
A 4352-18. First becoming visible in the spectra of type la (Group IV), which
have numerous lines, it is strong in the spectra of type Ha (Groups III and V),
and increases jo as to be one of the strongest lines as we pass towards type Ilia

76 INORGANIC EVOLUTION. [CHAP.

Professors Eder and Valenta thus state the conclusions they have
recently arrived at in their study of the changes in the spectrum of
mercury : —

" Ferner ist die Erscheinung der ziemlich unvermittelten Auf-
blitzens des linien-reichsten Spectrums (siehe die Abbildung, Fig. 8,
der heliographirten Tafel) bei hochgradig gesteigerLer Starke des
Flaschenfunkens und gleichzeitigem Erhitzen der Capillare, beson-
ders das Auftauchen zahlreicher neuer Hauptlinien, welche friiher
nicht oder kaum sichtbar waren, und mancher Doppellinien an Stelle
von einfachen Linien, eine derartige, dass sie zu Lockyer's Theorie
der Dissociation der Elemente passen wiirde, wenn man iiberhaupt die
Zerlegbarkeit lingerer Elemente in die Discussion ziehen will."*

[Translation : —

" Moreover the appearance of the great brilliancy of the richly
lined spectrum with a high tension jar spark, the capillary being
heated, and especially the interchange of a great number of new
lines which were dim before, and also the change of single lines into
double ones ; these are such that would harmonise well with
Lockyer's theory of dissociation of the elements, if one is prepared
to bring into the discussion the possibility of the dissociation of the
chemical elements."]

I am glad to be able to quote the following opinion of Sir William
Crookes,t to which I attach great weight : —

" Until some fact is shown to be unreconcilable with Mr. Lockyer's
views, we consider ourselves perfectly justified in giving them our
provisional adhesion, as a working hypothesis to be constantly tested
by reference to observed phenomena."

I am anxious to refer here also to the opinion expressed by my
colleague, Professor Sir William Roberts-Austen, whose researches
have mostly been carried on at high temperatures : —

" Mr. Lockyer has, however, since done far more : he has shown

(3roup II). Now, as was found by Liveing and Dewar, this line exhibits just the
same peculiarities in the laboratory ; in the spark spectrum it is hardly recognis-
able, in the arc spectrum it is very strong."

My most recent work suggests that Dr. Scheiner is wrong in identifying the
magnesium line 4352'IS in the cooler stars with the line nearly in the same position
in the hotter stars. In the hot stars the line behaves almost exactly like the enhanced
line of magnesium 4481'3, and I have previously pointed out that the stellar line
was therefore possibly not due to cool magnesium. This is now justified by the
discovery of an important enhanced line of iron at 4351*93, which accounts for the
line in the hot slars, and really strengthens Dr. Schemer's argument.

* DenJcschrijten der Jcaiserlichen Akademie der Wissenschaften, Wien, vol. Ixi,
p. 429, 1894.

f Chetn. News, 1879, vol. xxxix, p. G3.

VIII.

RECENT OPINION. 77

that the intense he;it of the sun carries the process of molecular
simplification much farther; and, if we compare the complicated
spectra of the vapours of metals produced by the highest tempera-
tures available here with the very simple spectra of the same metals
as they exist in the hottest part of the sun's atmosphere, it is diffi-
cult to resist the conclusion that the atom of the chemiit has itself
been changed. My own belief is that these ' atoms ' are changed,
and that iron, as it exists in the sun, is not the vapour of iron as we
know it upon earth."*

.The Basic Lines.

With regard to the basic line part of the inquiry, I think I shall
tiot be going too far in saying that it has been universally rejected,
.and chiefly on the ground that some lines which appeared coincident
at the dispersion I employed appeared double with higher disper-
sions. I have pointed out in the Chemistry of the Sun (p. 377)
that this is not a sufficient answer, but I have left aside this branch
of the inquiry for some years in the hope that some chemist would
take up the question of spectroscopic impurities out of which it grew.
But it is evident that this basic line point of view, even though it
be considered a less direct attack on the problem than others since
begun, assumes a much more important and definite position in the light
of the new work. I will not go into this question at length now, but
will content myself here by asking whether one actual demonstration
•of dissociation will not take a form very like that which the chemist
has taken to be a proof of the existence of impurities.

I shall return to this later on.

Other Physical Researches now in progress. \

So much for opinion a year or two ago. In subsequent chapters I
shall refer to other attacks upon the problem of dissociation, which to
my mind and to many of the objectors sets the matter on a much
firmer basis by accumulating facts, not only with regard to the stars,
but in other fields of inquiry in which the idea of dissociation has to
be appealed to in order to explain the phenomena.

* Proc. Roy. In<f., vol. xiii, p. 509, 1892.

78

CHAPTER IX.— THE STELLAR EVIDENCE.

I NEXT proceed to consider whether the views which I found
necessary to enable me to group together harmoniously and con-
tinuously solar phenomena years ago when nothing was known of
stellar chemistry, are weakened or strengthened by the study of the
enormous new field of investigation opened out by the recent stellar
work, by which we have finally the sun taking its place as one term in a
long series, the complete study of which enables us to watch the work-
ings of the celestial evolution which has built up the heavens as we
know them.

The great increase of our knowledge we have gained from the
study of stars arises from the fact that they have revealed to us a
continuous series of spectral changes at temperatures much higher
than the sun affords us.

One of the minor advantages of this is, that we can, taking the sun
as our base, see what would happen if the sun were to become hotter.
Let us consider this point first.

In approaching this part of the subject, it is necessary to proceed
with great caution, since the things observed are different. The solar
work has consisted in observing different parts of the sun, the star
work puts us in presence of the total effects both of radiation and
absorption in the case of each body observed.

The facts with regard to the lower portions of the solar atmosphere
have already been detailed. They have been gathered from the
photographs secured during the eclipse of 1898.

Having these unimpeachable series of facts to go upon, we have
found that the absorption indicated by the Fraunhofer lines is not
caused by the chromosphere, and that the most valid absorbing layer
lies above the chromosphere. We have also seen that in the chromo-
sphere we find enhanced lines among the Fraunhofer lines, which are
chiefly arc lines. What must happen then if the sun is supposed to get
hotter 1

It is only possible to consider the results produced by a higher
temperature on two hypotheses. The first, the usual one, that the
chemical elements are indestructible ; the second, that they are not.

On the first hypothesis it is difficult to say what change could take
place which would alter the characteristics of the Fraunhofer spectrum

CHAP. IX.] THE STELLAR EVIDENCE. 79

very widely. We have a complex mixture of the vapours of metallic
substances and gases with paramount calcium, hydrogen, and the
cleveite gases. Temperature cannot -therefore vary the relative inten-
sities of the lines. H and K, the chief lines of calcium, must always
remain predominant, iron must remain because it cannot be destroyed,
and since the quantity of hydrogen and the cleveite gases present
cannot be increased, their lines cannot therefore become more impor-
tant in the spectrum.

It is also clear that any change of relative density on the usual
hypothesis cannot be brought about by an increase of temperature ;
this, then, cannot alter, it cannot change the relative proportions of
chemical substances present in any layer, and therefore the relative
intensities of the lines which indicate the existence of the various
substances in the different layers.

If now we turn to the other hypothesis, that, namely of dissociation,
we see at once, in the light of laboratory experiments, that with every
considerable increase of temperature in all such masses of vapour and
gas as those which now constitute the solar chromosphere and revers-
ing layer, a fundamental change in the appearance of the spectrum
must be brought about ; complex molecules would be broken up into
simpler ones, and the result of this action would bring new lines into
the spectrum, indicating the vibration of the molecules produced.
Now let us come to facts. Were the temperature of the reversing
layer to be increased, if dissociation takes place at this temperature, the
dissociation products must become visible, and we must look for them
among those lines which expand at the expense of those which contract
and disappear. Is any such experiment as this going on even at this
moment ? The answer is beyond question.

The lower, hotter chromosphere differs from the reversing layer
precisely because this change has taken place. As I have said before,
we pass on descending the sun's atmosphere from the arc lines in the
reversing layer to the enhanced lines in the chromosphere, from the arc
spectrum to the " test spectrum," from the metals to the proto-
metals.

What could only be pointed out with regard to only a line or two
20 years ago can now be proved for a whole set of lines, and the dis-
sociation argument is seen to be vastly strengthened the more it is
tested.

Next, let us see where the stellar evidence helps us ; here I shall
deal with the main outlines merely. If in the sun the chromosphere is
hotter than the reversing layer in a star slightly hotter than the sun,
the reversing layer which builds up the stars' absorption should
resemble the chromosphere.

80 INORGANIC EVOLUTION. [CHAP.

I have already stated the facts with regard to a Cygni. Now let
us look at them in the light of the dissociation hypothesis.

The evidence is complete that the temperature in the reversing
layer of a Cygni is higher than that of the reversing layer of the sun.
What do we find 1 Of lines disappearing we have arc lines of iron,
?ome thousands in number, calcium, magnesium, strontium, and so on.
Of lines increasing in importance we have the small number repre-
senting the enhanced lines of iron, the lines of hydrogen, and some
others which we cannot at present associate with the name of any
known substance. Here, then, we get a series of phenomena which,
on the hypothesis we are discussing, is simply and sufficiently ex-
plained by the statement that on passing from the temperature of the
sun to that of a Cygni, among changes brought about the complicated
line spectrum of iron is giving way to a more simple one consisting of the
enhanced lines. Further inquiries show that the other metallic spectra
are behaving in the same way. Looking for the lines which increase in
importance, while the others are reduced, we find the lines of hydrogen.

So far then up the scale of temperature the solar and stellar record
is the same ; the star at the next stage of heat above the sun has its
reversing layer as hot as the sun's chromosphere, and the same " test
spectrum " as we have seen fits both. I hold that dissociation simply
and sufficiently explains this all-important fact.

But this is as far as the sun can take us. The stars, however, con-
tinue the story.

If we consider another change higher up in the scale of temperature,
taking as the lower level a Cygni, at which we have arrived, we have
independent evidence that the so-called Orion stars are hotter than
such a star as a Cygni.

On proceeding to study the higher dissociating temperature at
work in the Orion stars, all the statements made with reference to
the changes likely to occur in the spectrum on the non-dissociation
hypothesis, strictly apply. We cannot expect any change in the rela-
tive intensity of the lines and the appearance of the spectrum cannot
be fundamentally altered.

On the dissociation hypothesis, on the other hand, if we find cer-
tain lines indicating certain substances disappearing, and other lines
indicating other substances making their appearance for the first time
(or if they were visible before, becoming much intensified), we shall
have an opportunity of studying the effects of the new dissociating
forces at work.

Now is there any change 1 The facts are that this increase of tem-
perature we are now considering is accompanied by the gradual extinc-
tion of the enhanced lines, an increase in the amount of hydrogen

THE STELLAR EVIDENCE. 81

present, and the lines of the cleveite gases, oxygen, nitrogen, and
carbon now appear for the first time.

Associating this with the former result, we get as distinct evidence
that an increase of the gas lines in the spectrum accompanies the disap-
pearance of the enhanced lines, as that an increased development of the
enhanced lines accompanies the decrease of the arc lines.

To take iron as an example, for the sake of simplicity ; it will be
seen that the actual stellar phenomena might have been predicted up
to a certain point, from a consideration of laboratory and solar phe-
nomena. But the stars carry us further than our predictions ; we see
the gradual increase of hydrogen and cleveite gases. The facts
demonstrate that as temperature increases hydrogen increases, .and,
together with the cleveite gases not obvious before, finally replaces
iron which has disappeared.

This is one of the great stellar revelations, and it must be re-
membered that we have now hundreds of photographs which we can-
bring together to study the gradual change. There are no " breaks in-
strata." One of the most wonderful things about this line of work to
my mind is the simplicity, coupled with continuity, of the phenomena.
It carries conviction with it.

We have then to face the fact that on the dissociation hypothesisr
as the metals which exist at the temperature of the arc are broken up
into finer forms, which I have termed proto-metals, at the fourth stage-
of heat (that of the high tension spark) which gives us the enhanced
spectrum ; so the proto-metals are themselves broken up at some tem-
perature which we cannot reach in our laboratories into other simpler
gaseous forms, the cleveite gases, oxygen, nitrogen and carbon being
among them.

Does the story end here 1 No, there is a still higher stage ; after
the cleveite gases have disappeared as the arc lines and enhanced lines
did at the lower stages ; the new form of hydrogen to which I have
before called attention and which we may think of as "proto
hydrogen," makes its appearance. But there are already evidences
that even this is not the end of the simplifications brought about by
the transcendental stellar temperatures we are now discussing.

It must always be remembered that the Spottiswoode coil (giving a-
40-inch spark) with a tremendous battery of condensers only carries
up to 7 Cygni, by which I mean that using this coil we obtain the
enhanced lines of the proto-metals of very nearly the same relative
intensities as those under which they appear in that star.

In the stars then we have a few distinct changes of spectra : these
changes we know independently by the increase in. the length of the
spectrum towards the ultra-violet accompany stages of increased tern-

G

82 INORGANIC EVOLUTION. [CHAP. IX.

perature. It is most natural to suppose that these increasing tem-
peratures produce increasing simplifications.

Dealing, then, with the changes which we can now study in stellar
bodies from the temperature of the sun upwards, we have the series of
spectral changes on which the new chemical classification (Chapter
VII) has been based.

Now if dissociation is not the cause of these changes where are we
to look for one equally simple and sufficient 1

It is quite clear that the phenomena to be observed with every
increase of temperature, that is in a series of stars with spectra
gradually extending more and more into the ultra-violet, must be
vastly different if the elements are dissociated from what they would
be if the elements remained unchanged.

The only change which we can imagine on the usual hypothesis, as
resulting from the increase of temperature, is that with the increase
in volume there will be a reduction in density, and all the lines will be
equally enfeebled. But this is exactly what does not happen.

It may be said that in consequence of a more exalted temperature
in the hottest stars the hydrogen and cleveite gases may, for some
reason or other, escape from among the metallic vapours, and form an
upper special atmosphere of their own, in which, in consequence of its
greater chemical simplicity, the lines of these substances will become
more important. But this argument is not philosophical, because we
have no right to assume such a change. These gases already exist in
the sun and give us no traces of their existence at any great height
above the chromosphere ; the gas that does exist in these elevated
regions is one about which we know nothing, so far, terrestrially, and
•of which no trace has yet been found in the spectrum of the hottest stars.

I hold, then, that the stars more than justify my appeal to the law
of continuity ; their verdict is that, as in all previous human expe-
rience, a higher temperature brings about simplifications, and it is not
strange that as our horizon is expanded by new work we find the
same laws in operation. We have, in fact, in these phenomena the
work of dissociation carried on before our eyes in the hottest stars, to
a point not reached anywhere else, and the stars also tell us that this
is possibly beyond our laboratory possibilities, for the highest tempera-
ture I have employed only carries us to the heat level of y Cygni, in
which star the cleveite gases, if visible, give only very faint traces. We
are thus brought finally face to face with the fact that iron is a com-
pound into the ultimate formation of which hydrogen, or the cleveite
gases, or both, may possibly enter.

S3

CHAPTER X. — THE "SERIES" EVIDENCE.
Introduction.

I DEFINED the meaning of the term "Series" on p. 10, and pointed
out how one of the important discoveries in recent years enables us
to study spectra from a new point of view. I propose in the present
chapter to deal with this subject in its most general aspect, and to
inquire whether this new method of inquiry helps us with any sug-
gestions or facts which may be utilized in the discussion of the disso-
ciation hypothesis : in other words, whether the new evidence afforded by
series, like the new evidence accumulated by the study of stellar spectra,
strengthens the view that the line spectra of the so-called chemical
elements are produced not by one but by more than one vibrating
particle.

To explain what is meant by "series," it is well to begin by
studying what are termed fluted spectra. I have already referred to
these and given photographs on p. 10 ; these flutings are perfectly
rhythmic from end to end. The whole of a fluting may be regarded
as a unit ; it is generally strongest towards the right or the red end of
the spectrum, its elements gradually becoming dimmer as we approach
the violet end. It is well seen in the accompanying untouched photo
graph of some of the flutings in the spectrum of nitrogen (Fig. 30).

But a fluting is generally more than this ; it is built up of sub-
sidiary flutings. Each of the subdivisions of it is in itself an almost
exact representatation in the small of what the whole thing is in the
great; so that we have the conceptions of a simple fluting and a
compound fluting. The compound flutings are well represented in
the flutings of carbon and magnesium (see Figs. 9 and 10). In all
cases we get exquisite rhythm, though in some cases it is apparently
overlaid by other lines, and generally the system is intensified towards
the red end of the spectrum.

Now when we leave these flutings and study an ordinary line
spectrum, in a great many cases all rhythm seems to have disappeared.
There is apparently no law and no order. I have already in Fig. 1 1
given the series observed in the spectra of the cleveite gases. Let
us go into this a little closer and compare these " series " with the
spectrum as ordinarily observed. Let us take the lines seen when we
expose the gas obtained from the mineral cleveite to the act/ion of a

G 2

84

INORGANIC EVOLUTION.

[CHAP.

strong electric current. We observe no rhythm, and there seems to
be a very irregular distribution (Fig. 31).

I may here state that it has always been customary with me in
reproducing spectra in the form of illustrations to show the red end

ir

,

02

of the spectrum on the right hand side and the violet end on the
left. As most of the workers on "series" do the opposite, seeing
that they have to deal with the numbers of waves instead of their
length, I propose in this chapter to depart from my usual custom

X.J

THE " SERIES EVIDENCE.

85

and place the red in series spectra on the left, so that all the series
illustrations may be comparable inter se>

Messrs. Eunge and Paschen have shown conclusively that when we
come to sort these lines out into series, there is just the same

K

HJ

exquisite order that we find in flutings. Fig. 32 shows how they
have all been resolved into two sets of three series which gradually
get nearer together towards the violet and stronger towards the
red; the irregular line spectrum when analyzed in this way, is
translated into a wonderful order. I suggested many years ago that

86 INORGANIC EVOLUTION , [CHAP.

the triplets in the ordinary line spectrum of a substance may really
be remnants of compound flutings, and such inquiries as these really
seem to justify that suggestion.

We arrive at the fact that the term " series " applies to related
lines. It is impossible to suppose that these wonderful rhythmic
series of lines are not related in some way to each other, and that
being so we have to study their wave-lengths, that is, their positions
in the case of any one element to find out and define the relationship ;
and not only so, but to see if any relation exists between the lines of
different elements.

A Shoi't History.

The history of this quite modern inquiry is not very long, but
short as it is I only propose to refer to it in the briefest possible
manner.

The first attempt to discover relationships among the lines of
spectra was made by Lecoq de Boisbaudran,* who investigated the
spectrum of nitrogen. The conclusions he arrived at suggested
that the luminiferous vibrations of the molecules could be compared
with the laws of sound, but as these were not based on wave-length
determinations of sufficient accuracy, and also were not confirmed by
Thalen, no great weight could be attached to the result.

Stoney,f who followed up these investigations, was more success-
ful ; he showed that the hydrogen lines C, F, and h were connected
by the relationship 20 : 27 : 32.

Several other workers — Reynolds, Soret, &c. — took the subject upr
but it was left for the more thorough work of Schuster J to show that this
theory could no longer be considered as expressing the law connecting
the mutual relationships between the wave-lengths of lines in a
spectrum.

Liveing and Dewar§ next called attention to the fact that the
distance between two consecutive lines of these groupings decreases
with diminishing wave-lengths, so that eventually the lines asymptoti-
cally approach a limit. " Harmonic " was the term they used to
express such a series of similar groups of lines.

It was, however, the work of Balmer which gave the subject the
mpetus by which it has of late years made great progress.

Balmer|| published a formula by which the positions of the hydro-

* Comptes rendus (1869), vol. Ixix, p. 694

f Phil. Mag. (1871), [4], vol. xli, p. 291.

J Brit. Assoc. Report, 1880; Proc. Soy. Soc. (1881), vol. xxxi, p. 337.

§ Phil. Trans. (1883), p. 213, and previously.

|| Wied. Ann. (1885), vol. xxv, p. 8.

x.] THE "SERIES" EVIDENCE. 87

gen lines could be calculated with wonderful accuracy. The formula
is as follows : —

rt-2

x = A-^-i>

n- - 4

in which A is the wave-length in vacuo of' the line to be calculated, A
constant common for all the lines, and n one of the series of numbers
from 3 to 15.

The constant A, according to Cornu's measurements, is 3645*42
Angstrom units, or, using Ames' more correct value, 3647*20 Angstrom
units.

Simultaneously with Balmer's discovery, Cornu* pointed out that
the lines of aluminium and thallium, which are readily reversible, bear
a definite relation to those of hydrogen, while at a later date Des-
landresf published a formula from which could be calculated the wave-
lengths of the lines composing the bands of numerous elements.

The above brief history brings us down to the year 1887, in
which Kayser and RungeJ began their series of minute investigations
dealing with a great number of elements. It was also about this time
that Rydberg§ commenced to take up the subject.

The work of Kayser, Runge ami Rydberg.

I will state generally the ground over which their work has ex-
tended. They have attacked the question mathematically from
different standpoints. In the following table (p. 88) I give th&
formula employed by Kayser and Runge, and that employed by
Rydberg.

The formulae are not by any means identical, but both deal with
wave-frequency, that is to say, the number of waves in a given unit
of length. Both Kayser and Runge, and Rydberg employ certain
signs to represent the successive integers which have to be used to-
define certain of their terms, and in addition to this we get certain
constants which are calculated for each series. The most interesting
consideration from this point of view is that Rydberg found that
there was one constant which he could use in order to search for the-
series of lines in the spectra of all the chemical elements with which

* Complex rendu.1 (1885), vol. c, p. 1181.

t Ibid. (188G), vol. ciii, p. 375; (1887), vol. civ, p. 972.

"t "TJeber die Spectren der Elemente," Alhandlungen d. K. Alcad. Berlin,.
1888, 1889, 3890, 1891, 1892, 1893.

§ SvensJca Vetenslcat. Akad. Handlingar, Stockholm (1890), vol. xxiii No. 11;.
Wied. Annalen (1893), vol. 1, p. 629 ; (1894), vol. lii, p. 119.

INORGANIC EVOLUTION.

[CHAP.

Formula for Calculating Series.

Kayser and Eunge.

Bydberg.

where
<or

A = ware-length

— = wave frequency)

n = 3, 4,5, . .

A, B, C = constants calculated for
each series.

The constants for the principal series
tire different from those used in the
subordinate series.

For sub-series of every element the
constant A is nearly identical. For all
series of all elements the constant B does
not vary by more than 22 per cent.
This constant B corresponds to Byd-
berg's NO.

n = nn —

where

n — wave frequency
m = 1, 2, 3, . . .
NO = 109721'G (a constant ap-
plicable to all series
of every element)

»0 J characteristic constants

M 1 varying with each series.

In the above formula, when m = oo ,

n = n0; or n0 is the limit which the

number of waves n approaches when m

is infinite.

The value of N0 is assumed by Eyd-
berg to be constant, as it varies only
slightly, and this variation may be due
to uncertain data.

he worked. There was no common constant • similar to this used by
Kayser and Eunge, but they found that some of their constants varied
little from element to element. In that way they not only obtained
the first term of a series, but the whole series throughout the entire
length of the spectrum, and where observations had been made in the
case of the different elements they could of course check their calcu-
lations by the actual observations so made, and see how the theory
seemed to be justified as the work was extended. The first line in a
.series must be considered to be comparable to a fundamental note in
music. It represents really the longest light wave in the same way
that the fundamental note in music represents the longest sound
wave. Both series of results, obtained in the way I have described
by Kayser and Runge and by Eydberg, show us that, in many cases,
we may be almost certain to obtain from the higgledy-piggledy arrange
ment of the lines in the spectrum of any one substance two or three
beautiful regular series like those already shown in the case of
the cleveite gases. There is a little difference in the nomenclature
employed by the investigators to whom I have referred, as shown in
the annexed table.

x.] THE "SERIES" EVIDENCE. 89

Series Nomenclature.
Intensity. Kayser and Runge. Rydberg.

Strongest . .
Weaker
Weakest . .

. . ! Principal series . .
. . 1 st subordinate series
, . 2nd subordinate series

. . ' Principal series.
. . Nebulous series.
. . j Sharp series.

The strongest lines which they observed at the temperatures they
worked with, they put into what they call a " principal series," and
then the weaker lines were distributed among other two series.
Kayser and Eunge called them the " first- " and " second-subordinate "
series ; Rydberg calls them the " nebulous-series " and the " sharp-
series." The lines of the principal series almost always reverse
themselves very easily indeed — that is to say, that the absorption is
indicated by them more readily than it is by the other lines. When
we come to the second subordinate or sharp series, it is found that
these sometimes broaden out towards the red end of the spectrum.

This work, of course, has required considerable investigation ;
the first attempts were not quite satisfactory, because the observations
on which they were based had not been of sufficient accuracy. With
greater dispersion it has been found that some of the lines which were
supposed at first to be single are really double ; so that it is quite usual
now when we consider this question of series to suppose that in some
cases the series are composed of single lines, in other cases of doubles,
and in other cases of triplets ; and it was at first, indeed, imagined that
in these differences we were face to face with a very important physical
difference between the various elements, but Rydberg has suggested
that possibly after all it may be a difference merely in the seeing.

He says :*

" The difference between the doubles and triplets is only relative.
This opinion is confirmed by the fact that the triplets appear often
in the form of doubles, the most refrangible component not having
sufficient intensity to become visible. Further, the relative intensity
of the components of the doubles seems equal to that of the two less
refrangible components of the triplets.

" For these reasons I have dared to propose the hypothesis that
the two kinds of component rays are of the same order, or that
the doubles are only triplets of which the most refrangible com-
ponent is too feeble to be seen, or has perhaps the absolute value of
zero. . . ."

* Kon. Sv. Vet. Ale. Hand., vol. xxiii, ii, p. 135.

90 INORGANIC EVOLUTION. [CHAP.

If the lines are more difficult to see, and if the sub-series of
lines get stronger towards either the red end or the blue end, then
we are more likely to see one line than two, and more likely to see
two lines than three.

With regard to this suggestion made by Rydberg, it is interesting
to riote that Professor Kayser is not inclined to hold the same opinion
and does not look upon triplets, doublets, or single lines of the series as
remnants of flutings, whose other members are too weak to be seen.
He points out that we have for the elements of the first vertical column
in Mendeleeff's table, doublets ; for the second column, triplets ; for
the third, doublets. As the first column contains monavalent elements,
the second bivalent ones, the third trivalent, it seems as if the elements
with uneven valencies had doublets, those with even valencies triplets.
This is confirmed by the triplets of oxygen, sulphur, and selenium,
which belong to the sixth column, with even valency. As in every
natural group of elements, the first elements show the series strongest,
and they get weaker as the atomic weight increases (i.e., in the group
of alkalies we cannot see the weaker second series for rubidium
and caesium ; in the group copper, silver, gold we can find no series
in gold ; in the group of magnesium, calcium, strontium, barium, for
strontium the second series is already weak, for barium we cannot
find the series). We should expect to find, according to Rydberg's
hypothesis, in the spectra of every group first triplets, then doublets,
then single lines. But that is not so : so long as we find anything of
the series the members are and remain triplets or doublets.

There is only a very small number of the chemical elements
which give us single lines ; in the principal series, so far, we only
know of helium and •asterium ; in the subordinate series we only know
of asterium. The number of doubles is very much greater, but it is
not so great in relation to the principal series as it is in the case
of the subordinate series ; but although we have nine elements giving
us triplets in the subordinate series, we have only three which give
them in the principal series. These results are shown in the following
table.

It is well that I should indicate the basis of these statements, and
for this purpose I give in Fig. 33 a very small part of the spectra of
three different elements, in order that the way in which the work has-
been done may be followed. In the lower horizon we are dealing with
zinc, and the way in which the triplets have been picked out will be
easily gathered. The triplet in each case has its central line nearer to-
one side of the triplet than the other. All the triplets in the zinc
spectrum are perfectly symmetrical from that point of view. If we
take the upper spectrum — that of calcium — we find also that the

X.]

THE "SERIES" EVIDENCE.

91

Single lines.

Doubles.

Triplets.

Principal Subordinate
series. series.

Principal
series.

Subordinate
series.

Principal

series.

Subordinate
series.

Helium
Asterium

Asterium Hydrogen (?)
Lithium (?)

Helium
Hydrogen
Lithium (?)

Oxygen
Sulphur
Selenium

Oxygen
Sulphur
Selenium

Sodium

Sodium

Magnesium

Potassium

Potassium

Calcium

Rubidium

Strontium

: Caesium

Zinc

Copper

Cadmium

Silver

Mercury

Aluminium

-

Indium

Thallium

triplets are formed in exactly the same way. We can thus appreciate-
the enormous labour which has been faced by the inquirers I have
named in working out from the spectra of a great many substances
and from all the different regions of the spectrum, visible and photo-
graphic, these delicate triplets. In a great many cases they do not
represent the strongest lines, those most easily seen, and some want a.
great deal of looking for.

These investigations show that in some cases the series have repro-
duced the same chemical group, but in some instances the series
groupings, so to speak, are quite different from the chemical group-
ings.

The facts so far ascertained are as follows : —

Grroup 1 .. Lithium, sodium, potassium, rubidium, ca?sium.

,,2 .. Copper, silver, (gold?).

,, 3 .. Magnesium, calcium, strontium.

„ 4 . . Zinc, cadmium, mercury.

„ 5 .. Aluminium, indium, thallium.

Iii the group of lithium, sodium, potassium, the series sequence-
follows absolutely the chemical sequence. But when we come to the
chemical group — calcium, strontium, barium — we find it replaced by a
group, magnesium, calcium, strontium, while barium is not used at all.
That is a very remarkable departure, and it shows that we have to con-
sider the various conditions which we observe in passing from group
to group.

From group to group with increasing atomic weights the series
advance towards the violet. Thus, as the limit of a series is repre-

:92

INORGANIC EVOLUTION.

[CHAP.

* J

X.]

THE " SERIES EVIDENCE.

93-

sented by the first constant of the first subordinate of the four groups,,
the theoretical wave-length limit lies

Between 3498 '2 and 5065' 1 for lithium," sodium, potassium, rubidium, caesium.
„ 3168 '6 „ 3256 '0 „ copper, silver, gold.
,, 2512*8 „ 3222*6 „ magnesium, calcium, strontium.
„ 2328*5 „ 2490 '1 „ zinc, cadmium, mercury.

In each group with the increasing atomic weight the spectrum
advances continually towards the red end ; also the distance between
the components of the doublets and triplets increases with the atomic
weights, so that for every group the distance is approximately propor-
tional to the square of the atomic weight.

The Irregularities observed.

The above account I trust will give a general idea of the new
investigation in its most general aspect.

I have next to point out that we meet with most marvellous-
irregularities. We have some elements with many series, in others no
series have been detected, the numbers of the series varying even in
the gases. With regard to the metals, Kayser has suggested that the
melting point seems to have something to do with the phenomena
observed ; that is, that the higher the melting point the smaller
generally is the percentage of lines which is possible to distribute into-
series. The following table will show this : —

Relation of Series to Melting Points.

Element.

Melting point.
Centigrade.

Percentage of
series lines.

Barium

1600°

0

Gold

1200

4

Copper

1050

6

Silver .

960

26

Strontium .

700

20

Calcium

700

34

Magnesium

600

64

Zinc

410

80

Cadmium

; 320

50

Lithium

180

100

Sodium

90

100

Cffisiuni

62

100

Potassium

58

100

Rubidium

38

100

Mercury

•

-40

27

The accompanying general table will show the facts touching these
various points which are at present known. The metallic elements.

INORGANIC EVOLUTION.

[CHAP.

are arranged in the order of Mendeleeff 's groups, and the irregularities
touching the total number of series, of principal series, the simple or
compound nature of the lines of each series, and percentage of lines
picked up by the various series, can all be gathered from an inspection
of the table.

Mendeleeff groups.

Atomic weights.
No. of series.

.8

1

U

•3
a

'I
•8

£

Principal
series.

1st and
2nd sub-
ordinates.

1

a'i

3 -*

8-4| ad

-3 ^

1 -o^

-S § S
*#•§ .

0 M

i**

* S P,

|JB

Per cent, of lines of in-
tensity 10, picked up
in series.

•

G

T3

!i

^fe

s*»
il

"o O
£

05

'be

1

Double.

43

0

•1
FH

05
I

00

Double.

i

'Ei

g

Hydrogen . .
Helium
Asterium . .
Lithium
Sodium
Potassium . .
Rubidium . .
Caesium
Copper
Silver
Gold
Magnesium..
Calcium
Strontium . .
Barium
Zinc
Cadmium . .
Mercury
Aluminium. .
Indium
Thallium . .
Tin..
Lead
Arsenic
Antimony . .
Bismuth
Oxygen
Sulphur
Selenium . .

i"

k

J-

}-•
},,

}"-
h

|n

1

7*-0
23-0
39-0
85-2
13-3
63-4
107-6
196-7
2-1-3
39-9
87-4
136-8
65-1
111-7
199-8
27-0
113-7
203-7
117-8
206-4
74-9
119-6
207-5
15-88
31-8
78-5

3
3
3
3
3
3
1
1
2
2
0
2
2
2
0
2
2
2
2
2
2
0
0
0
0
0
6
3
3

1
1
1

i

i
i

i

0
0

°0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
(2)

1
1

X
X

?

| ?

X
y

X
X

••

X

X
X

p

X
X

9

••

100
100
100
100
100
100
100
100
6
26
p

100
100
100
100
100
100
100
100

• .

180
90
58
38
62
1080 -5
960
106 L -7
600
700
700
475
410
320
-40
654 -5
176
282
232
326
450
629-5
270

, ,

;;

p

X
X

::

.. ..

. .

••

X
X
X

64
34
20

55
17

7

M

. |

..

X
X
X

80
50

27

43
14
12-5
25
25
17

, i X

i

i

..

1

114
217

M

" I

X

••

• •

X

••

• 4

With regard to the stated absence of " principal series " in the case
of zinc, cadmium, and mercury, it may be pointed out that in each
case a very strong broad reversed line in the ultra-violet may repre-
sent the principal series; and in the case of copper, silver, and gold,
each of these elements contains in the ultra-violet a very strong pair of
lines which may represent the principal series.

x.] THE "SERIES" EVIDENCE. 95

I think it is quite fair to remark at this stage of our inquiry, that
if all the vibratory atoms which produce the spectra of the chemical
elements had all been brought to a similar condition of greatest sim
plicity, in other words, if we were really dealing with the chemical
atom as defined, in each case, the amazing irregularities which we have
found could hardly be expected.

Some Details.

I will next go a little further into detail in the case of some ele-
ments for the sake of instituting comparisons, and seeing whither the
results lead us.

The most remarkable case which I have to refer to is that of
hydrogen. We do not know the meaning of it yet, but it has to be
taken into account in any consideration of these questions. Until a
little time ago only one series was known in the spectrum of this gas,
and reasoning on this basis, it was thought that the atom of hydrogen
was far more simple than that of any other chemical element, and also
that a chemical atom was only competent to produce one series. A
short time ago, however, Professor Pickering, in his magnificent work
on the stars, to which I have already had the opportunity of referring,
pp. 58 et seq., discovered a second series of lines. Not long after, Pro-
fessor Rydberg suggested that one of the most important lines seen in
a large group of stars really represented a line of the principal series of
hydrogen. That conclusion has been generally accepted, although the
evidence is considered doubtful by some ; so that we now assume that
hydrogen has three series like helium and astqrium, and we seem there-
fore to be on solid ground in one direction, at all events, in regard to
some gases. That is, we may assume either that a simple atom may
by vibrating produce three series, or that hydrogen itself is of at least
threefold complexity. We have- another series of metals of low atomic
weight, which therefore chemically are supposed to represent a con-
siderable simplicity ; we find that in the case of lithium and sodium we
also deal with three series, a principal series and two subordinate series.
The same remark applies to potassium. It has recently been found that
sulphur and selenium also give us three series.' We have a principal
series and the first and second subordinates, the suggestion of anything
beyond these three is confined to one or two lines in each case.

But if we pass from the gas hydrogen to the gas oxygen, what do
we find 1

In oxygen we have six series, that is twice as many as we know of in
hydrogen, helium, asterium, lithium, sodium, sulphur, and so on. So
far as that goes, we are in the same condition that we were some

96 INORGANIC EVOLUTION. [CHAP.

time ago when we imagined that the gas obtained from the mineral
cleveite, when exposed to the action of a high tension spark, was really
a single gas with six series. Very many arguments have been employed
to show that that view is probably not an accurate one ; so that some
are prepared to separate the cleveite gas at spark temperatures into two,
calling one helium and the other asterium. That brings these two con-
stituents of the cleveite gas, brought out by high temperatures, to the
same platform as hydrogen with the recent developments, lithium,
sodium, sulphur, &c.

If we consider this extraordinary condition in the case of oxygen a
little further, we find that the six series only after all pick up the oxygen
hues seen at a low temperature, and that if we employ a high temperature
to observe the oxygen spectrum, that is to say, if we use an induction
coil, a jar and an air break, we find a very considerable number of lines
which have no connection whatever with any of the series so far made

•

s£u 1 i

*rS3

&&

Suki.

E

#IG. 34. — Map showing series and residual lines in spectra of calcium and

magnesium.

out. And we are face to face with this very awkward fact, that in the
case of oxygen there are more lines which we cannot get into a series
than there are lines in the six series which we have attributed to that
chemical substance. Here, therefore, on the hypothesis that we are
dealing with the oxygen " atom" we begin certainly to get into diffi-
culties. The inquiry is not straightforward.

The next point is, that in the case of other substances, we have no-
series, but only two subordinate ones. This happens in the

x.] THE "SERIES" EVIDENCE. 97

case of magnesium, calcium and strontium, and also aluminium, zinc,
and tellurium; we have a first and second subordinate series, but
no principal series. I have studied- the lines of calcium and mag-
nesium, in the same way that the lines of oxygen were studied, to see
how many of the lines are picked up by the series, and I proceed to
furnish some details. In the upper part of the diagram (Fig. 34) I
give the lines seen in the arc spectrum of calcium, and in the two next
horizons we have the lines picked up in the first and second subordinate
series. The next horizon gives the residual lines — lines, that is, which
are not distributable among these series. We see that there is a large
number outstanding just as in the case of oxygen, and it is very im-
portant indeed to note that the two lines H and K, which are more
conspicuous in the spectrum of the sun than all the other lines of the
spectrum, have not been caught by any of these researchers into the
series of calcium. Therefore, with a reduced number of series, we seem
to be getting still further from the simplicity we began with in the
case of some of the permanent gases like hydrogen and helium. The
game thing holds with regard to magnesium, the spectrum of which at
the temperature of the arc has not so many lines in it as the spectrum
of calcium. A certain number of these lines has been picked up to
form the series, but we get numerous lines which have been left over
after all attempts to sort them into series have been made.

I have now to refer to another consideration. We have dealt so far
in the case of calcium and magnesium with arc temperatures, but I
showed on pages 35 and 36 that in the case of these metals at spark
temperatures, the spectra are greatly changed, enhanced lines making
their appearance ; and I stated on page 57 that the all-important lines
in the hottest stars are lines seen at the temperature of the spark. I
have added these lines to the diagram, and we see that there is not the
slightest trace of those lines having been picked up in the series. So
that the further we go, the more we seem to get away from that beauti-
ful simplicity with which we began.

I refer next to another group of substances, namely, tin, lead,
arsenic, antimony, bismuth arid gold, and I might mention more. No
series whatever have as yet rewarded the many attempts of those who
have tried to get those metals and non-metals on all-fours with those
previously investigated. As already stated, it remained for Kayser
and Eunge to point out that it looked very much as if this complete
absence of series was connected with the melting points of the substances
with which they had been dealing. So long as the melting point was
low, as in the case of sodium and lithium, the normal three series
would show at low temperatures ; and, further, there were no lines
over. But, when we deal with substances with high melting points,

H

98 INORGANIC EVOLUTION. [CHAP,

there are no series at all. In the case of lithium, sodium, potassium,
&c., all the lines are picked up ; in the case of copper, silver, and gold,
the series pick up only a very small proportion. There seems, there-
fore, to be a progression of complexity with the increasing melting
point with regard to all the metallic substances which have so far been
examined.

In the case of barium with a high melting point, we get no lines at
all represented in "series" : contrasted with 100 per cent, in the case
of lithium. But then again, when we come to mercury, which is also
of low melting point, instead of getting 100 per cent, we only get
about 25 per cent, of the lines represented. The metals then vary as
do the gases.

General Conclusions.

The evidence then seems to indicate that the chemical units in the
case of the elements studied by the movements written out by these
series must possess different degrees of complexity. A little time ago
it was imagined that hydrogen was rendered visible to us by such
simple vibrations that only one series of lines could be produced. If
that is so, then it looks very much as if whenever we see three series
of lines at least three molecules or atoms, three different things, are
in all probability at work in producing them. When we get six
series, that points to a still greater complexity, and when, as in the
case of oxygen, we get six series not accounting for half the lines, then
we should be quite justified, I think, in supposing that oxygen was one
of the most complex things that we were brought face to face with in
our studies of " series " in cases where they are observable. When we
come to metals where there are no series at all, what do we find 1 We
are dealing with substances with high melting points — that is to say,
we cannot bring them down easily to those mobile states represented
by the free paths and collision conditions of a permanent gas ; and it
is quite easy to suppose, on that account alone, that we do not see the
vibrations of any of the more simple forms.

Hence, then, I submit that the evidence presented as to the com-
plex origin of line spectra by the studies of " series " is as clear as that
obtained from high temperature work in the laboratory and a discus-
sion of stellar spectra in relation to that work.

I have already referred to the case of hydrogen.

Professors Pickering and Kayser both concede that the new series,
is due most probably to a high temperature, and Kayser expressly
states, " that this series has never been observed before, can perhaps be
explained by insufficient temperature in our Geissler tubes and most o£
the stars."

X.] THE " SERIES " EVIDENCE. 99>

It seems as if the two series are of the " subordinate " type, and
that the principal series is wanting if Rydberg's conclusion be not
accepted ; because while in subordinate series the lines for large values
of n lie very near to one another, the similar lines of the principal series
on the other hand are always more refrangible. It seems, therefore,
probable that one or two of the many unknown lines recorded in stellar
spectra still awaiting identification may belong to the principal series
of hydrogen.

If we are dealing in this case with a single molecule of hydrogen
vibrating in a previously unknown way in consequence of a higher
temperature, why is it that the molecules of other bodies do not put on-
similar transcendental vibrations and appear in the same stars so that
we shall get new forms of the other chemical elements ? The fact that
we do not do so is, I claim, an argument in favour of the view that the-
principal and subordinate series are produced by molecules of different
complexities, and that the finer molecules can alone withstand the-
action of the highest temperatures, and require high temperatures to-
produce them.

In this way we can easily explain the visibility of the new form of
hydrogen in connection only or mainly with the lines of the cleveite
and other similar gases (for there is already evidence of the existence
of other similar gases) in the hottest stars.

From the admirable work done on such substances as lithium,,
sodium and potassium, which apparently are reduced to their finest
atoms at relatively low temperatures, and more recently on the series of
oxygen seen at low temperatures, we are bound to consider that when the-
research includes the complicated spectrum of iron that that also must
follow suit ; but it is already obvious that a principal and two sub-
ordinate series will never do ; there will be very many series involved.

Now these series must include both the arc and the enhanced lines,
and as these are visible each without the other in stars of different
temperatures, in one case associated with the cleveite gases, in another
without them, we have another argument in favour of molecular com-
plexity.

I may here point out that it is always the hot line which avoids
" series." The argument that lines in series represent the vibration
of one molecule proves that lines not in series are produced by the
vibrations of some other molecule.

Finally then, I stated in 1878 that the spectrum of a substance was
the integration of the spectra of various molecular groupings.

It has now been definitely established that the spectrum of some
substances is the integration of " series."

H 2

100 INORGANIC EVOLUTION. [CHAP. X.

So far there has been no definite pronouncement touching the
possibility that each series may represent vibrations of similar mole-
cules, but the facts as they stand are in favour of this view so long as
we consider a series as representing the simplest result of atomic vibra-
tion. There are facts which suggest that even a series is not a simple
result.

I am glad to be able to complete this chapter, which Professor
Kayser has kindly read over for me, with the following expression of
his opinion, which he allows me to publish.

" I quite agree with your opinion, that the molecules of elements
.are in general very complicated systems of atoms, and that their
complexity is very variable with temperature and perhaps other
•conditions. I think that at the highest temperature every molecule
has the simplest structure ; is perhaps a single atom ; and that in this
condition it will emit a very simple spectrum consisting of one, or
perhaps three, series of doublets or triplets. If the temperature is not
high enough above the melting point to dissociate all the molecules,
nevertheless some will be dissociated, and we shall have always a
mixture of molecules, from the most complex ones that can exist at
this temperature to the most simple ones. When the temperature
gets lower and lower, more and more complex molecules will be added,
while the simplest ones gradually disappear. In the same degree the
simplicity of the spectrum is lost, of the series only the strongest lines
or none remain, and the spectrum is the sum of more or less lines of
a great many different spectra. I expressed the same opinion in the
first publication of Kayser and Runge (Abhandl. d. k. Akad., Berlin,
1888), and I think our researches have shown nothing that contradicts
it."

101

CHAP. XL — EVIDENCE AFFORDED BY THE SHIFTING OF LINES.

RECENT work in America, by means of the great dispersion afforded by
Rowland's concave gratings, has supplied us with results* of the
highest interest, touching small variations in the wave-lengths of
spectral lines and the causes which produce them. These are stated to
have been, in the first instance, established by Mr. Jewell by an exami-
nation of the Rowland series of photographs of the solar and metallic
spectra taken by means of a concave grating of 21 \ feet radius and
20,000 lines to the inch — an instrument of research which, so far as
my own experience goes, is obtained with great difficulty by workers
in this country.

Mr. Jewell's investigations began in 1890. Messrs. Humphreys and
Mohler studied in 1895 the effects of pressure on the arc spectra of the
elements, work suggested by Mr. Jewell's prior researches.

Mr. Jewell, as a basis for his new conclusions, investigated under
modern conditions classes of phenomena which I was the first to-
observe and describe more than a quarter of a century ago.

To show the relation of the new work to the old, it is best to-
begin with a short historical statement, which will have the advantage
of giving an idea of the meaning of some of the terms employed.

I first employed, as stated on p. 22, the method of throwing an
image of a light source on to the slit of a spectroscope by means of a
lens in 1869, and some of the results obtained by the new method
were the following.

(1) The spectral lines, obtained by using such a light source as the
electric arc, were of different lengths; some only appeared in the
spectrum of the core of the arc, others extended far away into the
name and outer envelopes. This effect was best studied by throwing
the image of a horizontal arc on a vertical slit. The lengths of the
lines photographed in the electric arc of many metallic elements were
tabulated and published in Phil. Trans., 1873 and 1874.

(2) The longest lines of each metal generally were wider than the
others, the edges fading off, and they reversed themselves ; by which I
mean that an absorption line ran down the centres of the bright lines.
These results were afterwards confirmed and extended by Cornu.f

*• Astropbysical Journal, February, 1896, vol. iii, p. 111.
f Chemistry of the Sun, p. 379.

102 INORGANIC EVOLUTION. [CHAP.

(3) From experiments with mixtures of metallic vapours and gases
it came out that the longest lines of the smaller constituent remained
visible after the shorter lines had disappeared, the spectrum of each
substance present getting gradually simpler as its percentage was
reduced,* the shorter lines being extinguished gradually. Shortly
.alter these observations were made, I included among some general
propositions :f "In encounters of dissimilar molecules the vibrations of
•each are damped."

(4) The various widths of the lines, especially of the winged longest
ones, were found to depend upon pressure or density, and not tem-
perature. J

(5) The " longest lines " of any one metal were found to vary in
their behaviour in most extraordinary fashion in solar phenomena,
being furthermore differentiated from the shorter ones ; and on this
-and other evidence, I founded my working hypothesis of the dissocia-
tion of the chemical elements at the solar temperature. In 1876 I set
out the facts with regard to calcium.

(6) In 1883, Professor W. Vogel, in a friendly criticism, pointed
out the evidence, then beginning to accumulate, that under certain cir-
cumstances the wave-lengths of lines are changed.§ In 1887, I extended
this evidence,|| and I think it w^as I who coined the word " shift " to
•express these changes. U

I now pass on first to the results which Mr. Jewell claims to have
•established.

With the enormous dispersion produced by the instruments referred
to, it is found that certain metallic lines, but not all, are displaced or
•" shifted " towards the violet when compared with the corresponding
;solar lines. " There was a distinct difference in the displacement, not
only for the lines of different elements, but also for the lines of dif-
ferent character belonging to the same element."

The " different character " above referred to turns out to relate
not so much to the intensity as to the length of the line, and, asso-
ciated with this, its reversibility ; the longest lines are the most dis-
placed, the shortest, least.

Further, in the spectrum of the arc itself, the position of a line with

* Phil. Trans. (1873), p. 482.

f Studies in Spectrum Analysis (1878), p. 140.

t Phil. Trans., 1872, p. 253.

§ Nature, vol. xxvii (1883), p. 233.

|| Chemistry of the Sun, p. 369.

1[ Since the parentage is uncertain, I may say that perhaps "shiftings" would
Tiave been a better word, as shift is otherwise employed, e.g., Love's last shift
^translated by a French author, la derniere chemise de V amour).

XI.] EVIDENCE AFFORDED BY THE SHIFTING OF LINES. 103

but little material present " was approximately the same as the posi-
tion of the line when reversed." Now since the longest lines are most
displaced to the violet, this means that the smaller the quantity of
a substance present the greater is the displacement towards the violet ;
and, therefore, the greater the quantity present, the greater the dis-
placement towards the red.

Mr. Jewell found that " with an increase in the amount of the
material in the arc there was an increasing displacement of the line
towards the red," and then that, " unless the line became reversed, all
further progress in that direction ceased."

Here is an observation regarding the red line of cadmium. " It
was found that if the micrometer wires were set upon it with very
little cadmium in the arc, then as the amount was increased the line
almost bodily left the cross-hairs, always moving towards the red."

Mr. Jewell considers he has established that the vibration-period
of an atom depends to some extent upon its environments. " An
increase of the density of the material, and presumably an increase of
pressure, seemed to produce a damping effect upon the vibration
period." My result of 1872 with regard to pressure was endorsed,
" the new results are found to be due to pressure and not temperature."
We seem, then, now to be in presence of two damping effects in the
case even of metallic lines, one which extinguishes lines when we deal
with dissimilar molecules, and one which changes their wave-length
towards the red when we deal with similar molecules.

A carefully prepared table showed the origin, intensity and
character of the solar lines considered, the intensity and character of
the corresponding metallic lines, the wave-lengths of both, and the
observed displacement.

Many references to solar phenomena were made by Mr. Jewell in
relation to his work, but I do not propose to discuss them here. There
is one point, however, I must refer to. He considers that the conclu-
sions to be drawn from a study of the new shifts " effectually disposes
of the necessity of any dissociation hypothesis to account for most
solar phenomena." I have already pointed out that this was Professor
W. Vogel's conclusion with regard to possible shifts, so far back as
1883.

It is quite easy. " Two adjacent lines of iron, for instance, may
show the effects of a violent motion of iron vapour in opposite direc-
tions, in the neighbourhood of spots, or one line (the smaller one cor-
responding to one of Lockyer's * short lines ') may show a broadening
and increase of intensity in the spectrum of a sun-spot, while the
other line (the larger one corresponding to one of Lockyer's ' long
lines ') is unaffected. But this does not prove that iron vapour is dis-

104 INORGANIC EVOLUTION. [CHAP.

sociated in the sun. It merely shows that the apparently similar
portions of the two lines in the solar spectrum are produced at dif-
ferent elevations in the solar atmosphere. The stronger iron line
will be affected in a sun-spot as much as the other one, but it is
the portion of the line produced at the same level as the other line,
and may be masked completely, or very largely, by the emission line
produced at a higher level, while the second absorption line in the
solar spectrum may be entirely unaffected, being produced at a still
higher altitude."

" This also explains why some of the lines (the short lines generally)
of an element may be most prominent in sun-spot spectra, while others
(generally the long lines) are those most frequently seen in promi-
nences or in the chromosphere."

My thirty-three years' work at solar physics leaves me with such
an oppressive feeling of ignorance that I willingly concede to Mr.
Jewell a knowledge so much greater than my own as to give him a
perfect right to dismiss all my work in two lines ; but I am compelled
to point out that he has not carefully read what I have published.

A comparison of the facts brought together on page 26, for instance,
drives his last paragraph into thin air ; it is distinctly shown that we
have to do with the short lines in the chromosphere and with the long
lines in spots, the exact opposite of his statement. Mr. Jewell does not
run counter to my views in supposing that different phenomena are pro-
duced at different elevations. I thought I had abundantly proved in my
eclipse observation of 1882 (Chemistry of the Sun, p. 363), and the later
evidence will be found on p. 41, et seq., that the iron lines, to take a
concrete instance, are produced at different heights in the solar
atmosphere ; and that was one among many reasons which compelled
me to abandon the thin reversing layer suggested by Dr. Frankland
and myself in 1869 in opposition to KirchhofFs view. But surely the
more we consider the solar atmosphere as let out in flats, with certain
families of iron lines free to dwell in each and to flit a discretion, the
more a dissociation hypothesis is wanted. And beyond all this, we
have to take into account that at the sun-spot maximum no iron lines
at all are seen amongst the most widened lines, while at the minimum
we have little else.

The real bearing of the new work on the dissociation hypothesis
has been accurately caught by Professor Hale, as I shall show later.

Another very interesting part of Mr. Jewell's work refers to the

phenomena of absorption. There is room for plenty of work here.

As I pointed out in 1879, we get unequal widenings, " trumpetings,"

and a \vhole host of unexplained phenomena.* It is clear that the

* Chemistry of the Sun, pp. 380—387.

XL]

EVIDENCE AFFORDED BY THE SHIFTING OF LINES.

105

dispersion at Mr. Jewell's command will largely help

enormous
matters.

I now pass to Messrs. Humphre/s and Mohler's researches.

These investigators used an electric arc enclosed in a cast-iron-
cylindrical vessel, which enabled them to vary the pressure up to four-
teen atmospheres. One hundred photographs of metallic spectra
were taken, and the shifts of some lines of twenty-three elements have
been measured. The accompanying rough diagram, bringing together
specimens of their observations, will indicate the kind of result they
have obtained.

/O

35. — Changes of -ware-length produced by pressure, showing the different
behaviours of the lines of calcium (H and K and the blue line).

The pressures in atmospheres are shown to the left. The shift
towards the red in thousandths of an Angstrom unit are shown below.
The shifts have been reduced to what they would be at A, 4000, in the
neighbourhood of which most of the work was done.

The displacement or shift varied greatly for different elements. It
was always towards the red, and directly proportional to the wave-
length and the excess of pressure over one atmosphere.

106 INORGANIC EVOLUTION. [CHAP.

Only one exception to this general statement was noted at the
beginning of the inquiry ; it refers to calcium. " The lines H and K,
.among others, shift only about half as much as g (the blue line at
A 4226-91), and the group at A 5600. That.? should differ in this
respect from H and K is not very surprising, since it is known to differ
greatly from them in many other respects."

On this exceptional behaviour of these lines of calcium, I quote the
following, from a note by Professor Hale.*

" The difference in behaviour of H and K and the blue line of
calcium discovered by Messrs. Jewell, Humphreys, and Mohler, seems
to support Lockyer's views as to the dissociation of calcium in the arc
and sun. The remarkable variations of the calcium spectrum with
temperature have long been known principally through the investiga-
tions of Lockyer. The writer has shown that the H and K lines are
produced at the temperature of burning magnesium and in the oxy-
coal-gas flame. They could not be photographed in the spectrum of
the Bunsen burner, though an exposure of sixty-four hours was given.
Since these experiments were made, I have been informed by Professor
Eder that his own efforts to photograph the lines in the Bunsen burner
were no more successful, though an optical train of quartz and fluor-
spar was employed. It would thus appear that the temperature of the
dissociation of calcium is between that of the Bunsen burner and that
of the oxy-coal-gas flame. The high molecular weight of calcium has
hitherto conflicted with our belief in the presence of this metal in
prominences. If, however, it be granted that dissociation can be
brought about by temperatures even lower than that of the arc, the
difficulty is very greatly lessened."

In an article which I wrote in Nature on this work,f I pointed out
that " it would be very interesting to see if the strontium line at
A 4607*52 behaves like the calcium g in relation to the lines at A 4077-88
and A 4215-66, representing H and K."

This prediction was subsequently confirmed by Mr. Humphreys, J
who gave a table of the shifts measured on the strontium lines mentioned
above. When working with pressures varying from 6 to 12 atmo-
spheres, the shift of the line at A 4077*88 was always approximately
half that at A 4607-52.

There can be little doubt after this successful prediction that other
enhanced lines will follow suit as this new attack is carried further.

* Astrophysical Journal, loc. cit.
t Nature, vol. liii, p. 416, March, 1896.

J " The Effect of Pressure on the Wave-lengths of lines in the Spectra of certain
Elements," Asirophysical Journal, vol. iv, p. 249.

XI.] EVIDENCE AFFORDED BY THE SHIFTING OF LINES. 107

Artificial Shifting of Lines.

The " shifts " we have so far referred to are real, depending upon the
environment of the molecules the vibrations of which build up the spectra.

But there are also what we may term artificial shifts, the observa-
tion of which has recently led Dr. Schuster and Mr. Hemsalech to
conclusions of great importance almost equalling those noted by
Messrs. Jewell, Humphreys and Mohler from our special point of view.

To see the point of this new work, let us consider a strong jar spark
taken between two different metallic poles in air. What happens is
thus described.

u The initial discharge of the jar takes place through the air ; it
must do so because there is at first no metallic vapour present. The
intense heat generated by the electric current volatilises the metal
which then begins to diffuse away from the poles ; the subsequent oscilla-
tions of the discharge take place through the metallic vapours and
not through the air."*

Next let us assume that the vapours produced at each pole take time
to pass to the other. If we observe by means of a revolving mirror,
the spark qud air will give us a straight line, the spark gud each vapour
will give us curved lines.

Next suppose that instead of observing the sparks thus produced
by the three different sources, we observe their spectra. This has been
done by Dr. Arthur Schuster and Mr. Hemsalech, who thus refer to it : —

" The method of the rotating mirror tried during the course of
several years in various forms by one of us, did not prove successful.
On the other hand good results were obtained at once on trying the
method used by Professor Dixon in his researches on explosive waves.
This method consists in fixing a photographic film round the rim of a
rotating wheel. All that is necessary for its success is to have sparks
so powerful that each single one gives a good impression of its spec-
trum on the film. Were the sparks absolutely instantaneous, the
images taken on the rotating wheel would be identical with those
developed on a stationary plate, but on trial this is found not to be
the case. The metal lines are found to be inclined and curved when
the wheel rotates, and their inclination serves to measure the rate of
diffusion of the metallic particles. The air lines, on the other hand,
remain straight, though slightly widened.

" To avoid the tendency of the film to fly off the wheel when fixed
round its rim, as in the original form of the apparatus, a spinning disc
was constructed for us by the Cambridge Scientific Instrument Com-
pany. The film is placed flat against the disc, and is kept in place by

* Free. Roy. Soc., vol. 64, p. 331.

108 INORGANIC EVOLUTION. [CHAP. XI.

a second smaller disc, which can be screwed lightly to the first. The
diameters of the two discs are 33 and 22-2 cm., the photographs being
taken in the annular space of 10'8 cm., left uncovered by the smaller
disc. An electric motor drives the disc, and we have obtained velocities
of 170 turns per second, though in our experiments the number of
revolutions was generally about 120, giving a linear velocity of about
100 metres/second for that part of the film on which the photograph
was taken."

Now the curvature of the metallic lines must depend upon the
rate of diffusion of the vapours in opposite directions from the metallic
poles ; and if the spectrum of each metal used as a pole be due to the
vibrations of one set of molecules, there will be equal curvature in
all the lines of that metal.

The photographs however, so far taken, show that the curvature
is not equal ; so in this work as in the other I have referred to in the
previous chapters, and shall refer to in subsequent ones, we are driven
to the conclusion that the spectrum has a complex origin. The results
of the investigation, so far as it has gone, have not yet been completely
published, but Dr. Schuster in a letter to me states that he has " no
doubt as to great differences in inclination [curvature] of the bismuth
lines. I also believe the difference to be real in the case of the zinc lines
(the green doublet being different from the blue triplet), but this I do not
consider established with the same certainty as in the case of bismuth."

In order to give an example of the magnitude of the differences in
velocity determined by the unequal curvature of the lines, Dr. A.
Schuster allows me to print the following numbers :— -

Wave- Velocity

Metal. length. metres/ second.

Zinc 4925 1 41g

4912J

481H 545

4722 /

Cadmium 5379 1 435

5339 J
5086 1

4800 L 559

4416 |

3613-1

Bismuth 5209 i

4561 I- 1420

3696 J

43021 533

4260 J

3793 394

Mercury . 4359 481

3663 383

109

CHAP. XII. — EVIDENCE AFFORDED BY THE MAGNETIC PERTURBA-
TIONS OF LINES.

LONG before the present electro-magnetic theory of light was formu-
lated in its present shape, several observers endeavoured to see if any
spectrum change was to be noted when the light source was placed in
a strong magnetic field.

Of these, Professor Tait seems to have been the earliest. He made
the attempt in 1855 :* it led to no result. The same thing happened to
Faraday in 1862. Indeed, his experiment on this question was the
last he ever made. I extract the following account of it from his life
by Dr. Bence Jones : — t

" 1862 was the last year of experimental research. Steinheil's appa-
ratus for producing the spectrum of different substances gave a new
method by which the action of magnetic poles upon light could be
tried. In January he made himself familiar with the apparatus, and
then he tried the action of the great magnet on the spectrum of chloride
of sodium, chloride of barium, chloride of strontium, and chloride of
lithium."

An experiment made on March 12 is thus recorded : —

" The colourless gas flame ascended between the poles of the magnet,
and the salts of sodium lithium were used to give colour. A Nicol's
polarizer was placed just before the intense magnetic field, and an
analyzer at the other extreme of the apparatus. Then the electro-
magnet was made and unmade, but not the slightest trace of effect on
or change in the lines in the spectrum was observed in any position of
polarizer or analyzer.

" Two other pierced poles were adjusted at the magnet, the coloured
flame established between them, and only that ray taken up by the
optic apparatus which came to it along the axis of the poles, i.e., in the
magnetic axis, or line of magnetic force. Then the electro-magnet was
excited and rendered neutral, but not the slightest effect on the polar-
ized or unpolarized ray was observed."

About the year 1872, Professor Clifford and myself made some
experiments with the large Steinheil spectroscope then in use in my
laboratory at the School of Science ; the only magnet available was a

* Proc. Roy. Soc. Edin., vol. ix, p. 118, 1875-6.
t Vol. ii, p. 449, 1870.

110 INORGANIC EVOLUTION. [CHAP.

feeble one, and nothing came of them. In 1885 M. Fievez* was more
fortunate. He made a set of experiments which may be said to be the
first recorded success, or at least partial success, of the solution of this
problem which now concerns us. M. Fievez observing with a flame in
a magnetic field as Faraday had previously done. He noticed a widen-
ing, and apparently a doubling of lines, but the doubling he attributed
to absorption. He wrote : —

" Les phenomenes qui se manifestent sous Faction du magnetisme
sont identiquement les memes que ceux produits par une elevation de
temperature."

In spite of this, however, Dr. Preston has expressed the opinion
that if Fievez " had known the theory, the whole question would have
been settled in 1885."

The subject remained unfruitful until 1897, when Dr. Zeeman
made known the results! of an important series of observations which
he had been quietly carrying out.

In a course of measurements concerning the phenomena first ob-
served by Dr. Kerr, Dr. Zeeman was led to reopen the inquiry whether
the light of a flame submitted to the action of magnetism really did
undergo any change. He remarked : "If a Faraday thought of the
possibility of the above-mentioned relation, perhaps it might yet be
worth while to try the experiment again with the excellent auxiliaries
of spectroscopy of the present time. . . ." And his observations
established that the bright lines of spectra are modified considerably
when a strong magnetic field is used. It was at once seen why pre-
vious experimenters had failed : the effect is small, so that besides a
strong field, high dispersion is necessary.

No sooner had Dr. Zeeman made his discovery public, than Pro-
fessor Lorentz, and subsequently Dr. Larmor, investigated the subject
theoretically. They showed that dealing with the theory in its simplest
form, not only mere broadening of the lines should be expected, but
that each line should really consist of three separate lines, or in other
words, form a triplet.

According to the simple theory, each element of matter which
carries an electric charge proper to it — the complex being called an
ion — has its movements affected by the magnetic field.

If we consider these ions to be the elements of matter the move-
ments of which produce light, it is certain that in a magnetic field the
movements will be affected ; there will not only be the normal move-
ment in the orbit, but an added precessional movement, or spin, round

* Bulletin de VAcad. des Sciences de Belgique, 3e Serie, tome ix, p. 381, 1885.
t Phil. Mag., [5], vol. xliii, p. 226.

XII.]

THE MAGNETIC PERTURBATION OF LINES.

Ill

the lines of magnetic force. If we represent the electric charge of the
ion by e, and its inertia by m, the ratio e/m in a field of given strength
is proportional to the precession, or spin, of the orbit of the ion.

By using specially constructed electro-magnets, and arranging
special conditions of the experiment, it was not long before a magnetic
field was produced which was sufficiently strong to completely separate-
the components of the lines previously thought to be only broadened.

MOO
§ £ *

* i S

2 1 A

K*

S'C *

sir

P tC

"o

fe

S1 2 •&

t^ * r-

H c g^.S

<d 1 5|:

CO .5 _C S3

o"

£

It was found that while ,<?0m0 of the spectral lines were converted
into triplets, others were resolved into quartets, sextets, octets, or other

112 INORGANIC EVOLUTION. [CHAP.

complex types, while others again remained almost if not altogether
unchanged.

And then there was another revelation.

Not only do lines in the spectra of different substances vary in this
respect, but lines in the spectrum of any one substance are differently
changed ; while some spectral lines of an element show a considerable
resolution in the magnetic field, others are scarcely affected at all.
This important fact was first stated by Dr. Preston in 1897.*

This brings us to the connection between this line of work and my
own, for we now find lines of the same substance behaving differently
qua magnetic -perturbations, as I found iron lines behaving differently
in the spectra of sun spots qua velocity. About this different behaviour
qua perturbation there is no question. I will refer to some of the work
since done in this connection.

M. Cornu was the next to note the importance of this. He
writes : —

" The effect of the magnetic field on the period of vibration of the
radiations of the luminous source seems to depend, not only upon the
chemical nature of the source, but also upon the nature of the group of
spectral lines to which each radiation belongs, and on the part which it plays
in this group."! Somewhat later MM. H. Becquerel and Deslandres
gave details with regard to .the spectrum of iron in the ultra-violet
region, calling attention to these observations as being of great import-
ance " physically, chemically, and astronomically. "J

Dr. Zeeman§ subsequently published the statement that observing
across and along the lines of force, although the vast majority
of iron lines were, with the field used, resolved into doublets,
triplets, quadruplets, &c., three or four lines seemed unaffected. In the
case of a few lines he further found inequality between the outer com-
ponents of a triplet across, and of the corresponding doublet along,
the lines of force.

Messrs. Ames, Earhart, and Reese next noticed further peculiarities
about the behaviour of some of the iron lines. ||

When the radiation at right angles to the magnetic field was studied,
each line in the spectrum was found in general to be broken up into
three, the central component being plane polarized with its vibrations
-along the line of force, the two side components being plane polarized

* Trans. Hoy. Dub. Soc., vol. vi, p. 385 (1898), and vol. vii, p. 7 (1899).
f Astrophysical Journal, vol. vii, p. 163, 1898.
J Comptes Rendus, vol. cxxvi, p. 997 ; vol. cxxvii, p. 18.

§ Proc. of Roy. Acad. of Sciences, Amsterdam, June 25, 1898, and Astro-
physical Journal, vol. ix, p. 47.

|j Astrophysical Journal, vol. viii, p. 48.

XII.] THE MAGNETIC PERTURBATION OF LINES. 113

at right angles to these, their vibrations being at right angles to the
field of force.

Exceptions to this rule, however, were found in the lines having
wave-lengths 3587-13, 3733*47, and 3865'67, which behaved in exactly
the reverse manner. Two other lines at wave-lengths 3722*72 and
3872-64, were quadruplets, the central component, which had its vibra-
tions along the line of force, being a close double. Some of the lines
which showed no modifications whatever, were those at AA 3746 '06,
3767-34, 3850-12, and 3888-67.

These observers further noticed that the separation of the side com-
ponents of the triplets seemed to be irregular ; they found that there
were certain lines in which the separation was nearly the same, but
much greater than that of other lines where separations seemed to be
quite closely alike. On this basis they separated the lines in the iron
spectrum into two classes, in each of which the " magnetic separation "
was the same, but in the one set much greater than in the other.

The lines belonging to these two sets were found to be practically
identical with those sets of lines into which the iron spectrum breaks
up when studied with reference to the shift produced by pressure, but
this conclusion is not accepted by Dr. Preston.

I have already stated that on the simple theory we should get
triplets only, as on the simple theory of thirty years ago we should have
got motion of a solar vapour indicated by all the lines in a spectrum.
The facts are equally against the simple theory in both cases.

The magneticians can now, however, by extending their theory,,
embrace and explain all the new, and at first sight extraordinary,
phenomena. To show how they have done it, I cannot do better than
quote from a lecture recently given by Dr. Preston, who is among the
most successful investigators of this new branch of science.*

" According to the simple theory, every spectral line, when viewed
across the lines of force, should become a triplet in the magnetic field,
and the difference of the vibration frequency between the side lines of
the triplet should be the same for all the spectral lines of a given sub-
stance. In other words, the precessional frequency should be the same
for all the ionic orbits, or the difference of wave-length 8 A between
the lateral components of the magnetic triplet should vary inversely
as the square of the wave-length of the spectral line under consider-
ation. Now, when we examine this point by experiment, we find
that this simple law is very far from being fulfilled. In fact, a very
casual survey of the spectrum of any substance shows that the law
does not hold even as a rough approximation ; for, while some spectral

* Nature, vol. 60, p. 178.

114 INORGANIC EVOLUTION. [CHAP.

lines show a considerable resolution in the magnetic field, other lines
of nearly the same wave-length, in the same substance, are scarcely
affected at all. This deviation is most interesting to those who con-
cern themselves with the ultimate structure of matter, for it shows
that the mechanism which produces the spectral lines of any given
substance is not of the simplicity postulated in the elementary theory
of this magnetic effect.

" According to the prediction of the simple theory, the separation
8A. should be proportional to A2, and although this law is not at all
obeyed if we take all the lines of the spectrum as a single group, yet
we find that it is obeyed for the different groups if we divide the lines
into a series of groups. In other words, if the lines of a given
spectrum be arranged in a series of groups, the lines of the first
group being denoted by the letters AI, BI, Ci, . . . , those of the second
group by A2, Bo, C2, . . ., and so on, then the corresponding lines Ab
Ao, A3, &c., have the same value for the quantity ejm, or, as we may
say, they are produced by the motion of the same ion. The other
corresponding lines, Bb B^ B3, &c., have another common value for
•e/m-, and are produced therefore by a different ion, and so on. We. are
thus led by this magnetic effect to arrange the lines of a given spectrum
into natural groups, and from the nature of the effect we are led to
suspect that the corresponding lines of these groups are produced by
the same ion, and therefore that the atom of any given substance is
really a complex consisting of several different ions, each of which
gives rise to certain spectral lines, and these ions are associated to form
an atom in some peculiar way which stamps the substance with its
own peculiar properties."

The general law announced by Preston states the further remark-
able fact that if we consider a group of chemically related metals such,
for example, as magnesium, zinc, and cadmium, then the sets of lines
into which the spectrum of any one of these may be divided as
above, correspond set for set with those into which the lines of any
other of these metals are divided, in such a way that the magnetic
change of frequency (or e/m) for any one set is the same as that for the
corresponding set in each of the other metals. This seems to point to
the conclusion that the metals of the. same chemical group are built up,
in part at least, of ions which are the same in all the metals of the
group.

It will be abundantly seen, then, that these new inquiries have
presented exactly the same difficulties as the old ones, and that they
have been met in exactly the same* way, by establishing the fact that

XII.] THE MAGNETIC PERTURBATION OF LINES. 115

the spectra of elementary substances are not produced by the vibra-
tion of similar " atoms " or " ions," but by a series of different ones.

It is already pretty obvious that .when ordinary spectroscopic obser-
vations, and the evidence supplied by " series," and these magnetic per-
turbations are completely correlated, we shall have taken a long step
forward.

I 2

116

CHAP. XIIL— " FRACTIONATION " EVIDENCE.

IN the three previous chapters I have endeavoured to show that new
methods of inquiry in the physical field all support the dissociation
hypothesis. I have next to show that similar confirmation may be
expected when the present ineffective chemical methods of analysis and
determination are replaced by more stringent ones, such as those exem-
plified and foreshadowed by Sir William Crookes's patient fractionation
work on yttria.

For the first definite chemical confirmation of my work I had to
wait till 1883. In that year Sir William Crookes gave an account, in
a Bakerian Lecture to the Royal Society, of his beautiful researches
on yttria. In the lecture he gave a sketch of the train of reasoning by
which he had been led to the opinion that systematic fractionation had
split up this stable molecular group into its " constituents," and these
were not yttrium and oxygen, as they should have been.

Subsequently in an address to the British Association at the Bir-
mingham meeting in 1886, he gave an account of the method of frac-
tionation which had led to these results.

The importance of the work on yttria in relation to the question of
dissociation lies in the fact that by the variation in intensity of the
various lines of the phosphorescence spectrum of yttria, Sir William
Crookes was led to the view that more elements than one were in
question — that the ordinary chemical processes had been quite unable
to make anything but an element out of a mixture. As a result of his
work he found five components "by a veritable splitting up of the
yttrium molecule." This obviously strengthens the view that if our
chemical resources were much greater than they are, the demonstration
that other similar changes of intensity in the spectra of other elements
would also be achieved.

I now quote Sir William Crookes on his method, which constitutes
a veritable new engine of chemical research.

" Broadly speaking, the operation consists in fixing upon some
chemical reaction in which there is the most likelihood of a difference
in the behaviour of the elements under treatment, and performing it
in an incomplete manner, so that only a certain fraction of the total
bases present is separated : the object being to get part of the material

CHAP. XIII.] " FRACTIONATION " EVIDENCE. 117

in the insoluble, and the rest in the soluble, state. The operation
must take place slowly, so as to .allow the affinities — which, by the
nature of the case, are almost equally balanced — time to have free
play. Let us suppose that two earths are present, almost identical in
chemical properties, but differing by an almost imperceptible variation
in basicity. Add to the very dilute solution dilute ammonia in such
amount that it can only precipitate half the bases present. The dilu-
tion must be such that a considerable time elapses before the liquid
1 Kjgins to show turbidity, and several hours will have to elapse before
the full effect of the ammonia is complete. On filtering we have thfc
earths divided into two parts, and we can easily imagine that now
there is a slight difference in the basic value of the two portions of the
earth, the portion in solution being, by an almost imperceptible amount
more basic than that which the ammonia has precipitated. This
minute difference is made to accumulate by a systematic process until
it becomes perceptible by a chemical or physical test."

With reference to the result to which this most laborious research
had led him, I will quote his own words,* remarking in the first
instance that crude yttria from samarskite, gadolinite, cerite, and
other similar minerals, is the raw material. The first operation is to
free it roughly from earths of the cerium group, which is effected by
taking advantage of the fact that the double sulphates of the potasi-
sium and the yttrium metals are easily soluble in saturated potassium
sulphate solution, while the corresponding double sulphates of the
cerium group of metals are difficultly soluble-.

" No longer than twelve months ago the name yttria conveyed a
perfectly definite meaning to all chemists. It meant the oxide of the
elementary body, yttrium. I have in my possession specimens of
yttria from M. de Marignac (considered by him to be purer than any
chemist had hitherto obtained), from M. Cleve (called by him ' purissi-
Mium '), from M. de Boisbaudran (a sample of which is described by
this eminent chemist as * scarcely soiled by traces of other earths '),
and also many specimens prepared by myself at different times and
purified up to the highest degree known at the time of preparation.
Practically these earths are all the same thing, and up to a year ago
every living chemist would have described them as identical, i.e., as
the oxide of the element yttrium. They are almost indistinguishable
one from the other, both physically and chemically, and they give the
phosphorescent spectra in menu with extraordinary brilliancy. This is
what I formerly called yttria, and have more recently called old yttria:
Now these constituents of old yttrium are not impurities in yttrium
any more than praseodymium and neodymium (assuming them really

* Chemical Neu-s, vol.liv, p. liOO.

118

INORGANIC EVOLUTION.

[CHAP.

to be elementary) would be impurities in didymium. They constitute
a veritable splitting up of the yttrium molecule into its constituents."

FIG. 37. — Showing how by the method of fractional oil yttria is separated into fire
different substances, defined spectroscopically by the different intensities of
the phosphorescent lines.

•" The final result to which I have come is that there are certainly
five, and probably eight, constituents into which yttrium may be split.
Taking the constituents in order of approximate basicity (the chemical
analogue of refrangibility) the lowest earthy constituent gives a deep
blue band Ga (X 482) ; then there is the strong citron band GS (A. 574),
which has increased in sharpness till it deserves to be called a line ;
then come a close pair of greenish-blue lines, G/3 (A 549 and A 541,
mean X 545); then a red band, Gf (A 619), then a deep red band,
Grj (y 647); next a yellow band, Gt (A 597); then another green band,
Gy (A 564) ; this (in samarskite and cerite yttria) is followed by the
orange line So (A 609). The samarium bands remain at the highest
part of the series. These, I am satisfied, are also separable, although
for the present I have scarcely touched them, having my hands fully
occupied with the more easily resolvable earths. The yellow band,
Ge, and green band, Gy, may in fact be due to a splitting up of
samarium."

So far as I know, Sir William Crookes has not" yet named the

XIII.] " FRACTION ATIOX EVIDENCE." 119

elements differentiated by the lines of the wave-lengths to which he
refers ; but more recently, still ^dealing with yttria, he has made
another research, having for its objective the separation of the element
characterised by a group of lines in the neighbourhood of A 3110.
The discovery of victorium with an atomic weight near 117 ha»
rewarded his efforts.

By following up the spectroscopic evidence, then, Sir William
Crookes has already "split up" one "element" into five; another
argument, if one were needed, that the spectrum of an element is pro-
duced not by similar but by dissimilar molecules.

120

BOOK IV.— OBJECTIONS TO THE DISSOCIATION
HYPOTHESIS.

#
CHAP. XIV. — THE CHEMISTRY OF SPACE.

I HAVE now to refer to certain objections which have been urged
against the views to which I have been giving expression during the
last thirty years, views which seemed to me to indicate a way out of
the tangle in which both laboratory, solar and stellar spectroscopic
work have one after the other landed me, and as I have shown in the
immediately preceding chapters, others after me, engaged in somewhat
similar inquiries.

The objections to which it is of most importance to refer in this
place have to do with the stellar evidence. I have supposed, and I
think legitimately, until the contrary is proved — that is, the onus pro-
hindi lies with objectors — that the materials out of which, on the
Meteoritic Hypothesis, worlds are eventually formed, are similar in all
parts of space. Neither Kant nor Laplace thought of differentiating
the ultimate chemistry of the material, and indeed the only view of
special differences which has been put forward to my knowledge in
recent years was a subtle one suggested by a learned divine to account
for miracles. On this theory, in certain parts of space miracles might
happen, in others not ; and the movement of the solar system through
space provided us with the necessary changes of this condition.

But quite recently this view has been extended to the chemical
conditioning of space, and the first most general objection I have to
meet is that the various spectral differences which it has been my duty
to chronicle as defining the various groups of stars, are not brought
about by temperature, but are due simply to the fact that the chemistry
of space varies, so that in consequence of their locus of origin and
their environment, some stars may be composed chiefly of the cleveite
gases, others of hydrogen, others of calcium, others of iron, others of
carbon, and so on.

But it is assumed that there may be some cases, not so extreme
as these, so that only the relative composition may vary from star to
star. This view of space divided into chemical parishes is supposed to
be supported by the alleged localisation of stars of the same type in
particular parts of space (as indicated by proper motions, &c.).

.CHAP. XIV.] OBJECTIONS TO THE DISSOCIATION HYPOTHESIS. 121

A possible vcra caum of such chemical differentiation wa$, I believe,
suggested by Dr. Wolf, who in 1866,* misled by Sir Wm. Huggins'
statements concerning the chemistry of the nebulae, endeavoured to
explain their spectra, and therefore their chemical constitution as
distinguished from that of stars, in a way that will be gathered from
the following extract : —

" If we admit the data of spectrum analysis as to the gaseous state
of these singular bodies (the nebulae), and the simplicity of their com-
position, one is led to see in them only the residuum of the primitive
matter after condensation into suns and into planets has extracted the
greater parts of the simple elements which we find on the earth, and
chemically in some of the stars."

It will be seen that Wolf considers only the differentiation of
nebulae from suns by the "extraction of matter" by some previous
local action. The chemistry was general to begin with, then the resi-
duum was worked up.

Dr. Schuster, however, has more recently gone further, still start-
ing however from a general chemistry : —

" We have no reason to believe that the nebulae of the present day
resemble our sun's ancestor. Some of the stars which are now in an
early stage of development, may be forming through the condensation
of matter which has been left over by others ; and it would not be
surprising if the youngest star did not agree in constitution with its
aged companions.''!

Let us suppose then that the number of different chemical parishes
in space is : legion to begin with, and that by such actions as those
suggested by Drs. Wolf and Schuster more differences are established,
surely the stellar differences must be legion too. I would submit that
the more such causes as these be added to a hypothetic irregular dis
tribution of different kind of matter in space, the more differences in
the chemical, constitution of stars should be found. But this is not so
according to the facts.

While the number of chemical elements known at the present time
is over seventy, the number of well-marked groups of stars is only ten,
if we take one side of the temperature curve ; that is, if we deal with
stars increasing or stars decreasing their temperature. We are justi-
fied in using one side only, because the spectra of stars on opposite
sides of the temperature curve indicate precisely the same elements,
though the percentage composition of effective absorbing regions is
different in the two cases. At the same temperature on opposite sides,

* Hypotheses Cosmogoniques, p. 7.
f Proc. Roy. Soc., vol. 61, p. 209.

122 INORGANIC EVOLUTION, [CHAP,

the chief difference is in the inversion of the intensities of the hydrogen
and the metallic lines.

Hence the facts are distinctly against the view of different chemical
parishes in space ; they also suggest that we are not justified in even
conceding possible variations in the percentage composition. On this
ground an infinite variety of spectra might be expected, but, as already
stated, the number of well-marked groups is ten.

The Sun, Capella, arid Arcturus, and other cooling stars, enormously
separated in space, contain the same spectral lines with almost identical
intensities, so that not only do they contain the same " elements," but
they contain them in absolutely identical proportions. The earlier and
hotter stages of such stars could not therefore have consisted of different
mixtures.

Again, all the blood-red stars, which it is generally acknowledged
are near the point of extinction, have practically identical spectra.

Another strong argument against the objection now under discussion
is that each particular kind of star spectrum is always associated with the
same degree of stellar temperature as determined by other considera-
tions, chiefly the extension of the spectrum into the ultra-violet. With
differences of chemical composition, different spectra would occur with
equal temperatures.

We are therefore justified in the conclusion that the differences
recorded in stellar spectra do not come from a different percentage
composition of the elements present, but arise from the action of dif-
ferent temperatures in the same molecules • and until the above facts
are explained, I must hold that the argument is complete that we do
get the same elements represented by different spectral lines in different
stars when the apparent differences are such as to suggest the objection
to which I am now referring. It is not a question of the absence of
elements, but of the absence of certain molecular complexities of each ele-
ment, which separates the spectrum of the sun from those of the stars
of various orders.

Having said so much regarding the objection generally, I must
now proceed to discuss the only piece of evidence which has been
brought forward in support of it, namely, the alleged localisation of
certain chemical groups of stars in particular parts of space, arising
from the fact that certain of the chemical elements are only to be
found in certain regions. This localisation is not held to be a quantita-
tive one merely, that is, depending upon varying proportions of elements,
but upon their absolute absence here and there.

I propose to discuss this question in the following way.

Since we can only deal with the masses of matter in space which
are visible, it is obvious that any inquiry mto the distribution of the

XIV.] OBJECTIONS TO THE DISSOCIATION HYPOTHESIS. 123

chemical conditionings, as revealed by spectra, of these masses must be
preceded by an inquiry into the distribution of the visible masses, con-
sidered merely as masses, and quite independent of chemistry.

We must therefore first deal with the general distribution of the
stars and nebulae, independently of their chemistry. That will give us
a general idea of our stellar system.

Having this as a basis, we can next see whether stars of the same
chemistry are seen along the same radius (taking our solar system as
the centre) or the same direction in space. Next, taking distances into
account, we can see if there be any proof of different chemical shells,
so to speak.

It must be borne in mind that a greater or less proportion of stars of
the same chemical quality in certain regions will not touch the question.
We can only deal with demonstrations of the absence of certain
chemical elements in certain regions, so far as the stars supply us with
evidence.

124

CHAP. XV.— THE GENERAL DISTRIBUTION OF STARS.

THE labours of three or four generations of astronomers have con-
clusively proved that the distribution of the stars well within our ken
is dominated by the Milky Way. Although the Milky Way to the
naked eye looks very unlike the other parts of the heavens, we have
known since the time of Galileo that the difference arises from the fact
that it is composed of a tremendous multitude of stars, a very large
percentage of the masses of matter which compose our system lying in
its plane ; it does not merely represent a fiery or igneous fluid, as dif-
ferent schools thought it did in the olden days. A small opera-glass
or telescope easily shows us that we are in presence of an innumerable
multitude of stars.

The Milky Way is a great circle inclined at an angle of about 62
to the earth's equator or to the equatorial plane extending to the stars.
"V\'e know nothing, of course, of the reason for that angle of 62°, but
it has its importance, because not only must the belt cross the equator
at two opposite points, as it does in two opposite constellations, Aquila
and Monoceros, but the poles of the Milky Way must lie at the points
of greatest distance from the junction with the equator in certain
constellations. These are Coma Berenices and Sculptor, and the posi-
tion of the N. galactic pole, as the north pole of the Milky Way is
called, is in R.A. 12 h. 40 m. Dec. + 28°.

When we come to look at the Milky Way a little more closely, we
find that from two points in it branches are thrown out, so that over
some part of its orbit, so to speak, it is double. The great rift which
separates these two parts of it begins near a star in the southern
hemisphere, a Centauri, and it continues for more than six hours in
right ascension until the two branches meet again in the constella-
tion Cygnus, which is well within our ken in the northern heavens.
The distance apart of the middle lines of these two components of the
Milky Way. where the split is most obvious, is something like 17°, so
that, in addition to the angle of 62° from the ecliptic, in some part of
the Milky Way, there is another offshoot springing out of it at an
angle of something like 17°. The regions of greater brilliancy corre-
spond approximately to the places where the branches intersect each
other. In short, there are sundry indications that the whole pheno-
mena of the Milky Way may become simplified by treating it as the

CHAP. XV.] THE GENERAL DISTRIBUTION 01* STARS. 125

resultant of two superimposed galaxies. The general view till
recently was that the Milky Way is not a great circle, because it was
thought the sun was not situated in its plane. The whole mass of
stars was likened to a millstone split along one edge, which was Sir
William Herschel's first idea. But the recent work, chiefly of Gould
in Argentina, has shown that it practically is a great circle. However
that may be, in one part of the heavens this wonderful Milky Way
appears as a single, very irregular, stream, and in another part it
appears to be duplicated.

This galaxy of stars is full of wonderful majesty and complexity.
We find in it indications of delicate markings going out into space,
apparently coming back strengthened ; of streams in all directions ; of
clusters clinging to those streams, and so on. In other parts it is
curdled, which is the only term which I can use to express my mean-
ing. In one region we may find it absolutely free from any important
stars ; in another we may find it mixed with obvious nebula ; and in
another we may find it mixed not only with obvious nebula, but with
a great number of bright-line stars involved not only in the Milky
Way, but in the nebula itself.

We have now, fortunately for science, priceless photographs of
these different regions which give us an idea of the enormous number
of stars in some parts, and of the streams of nebulous matter which
are seen in the Milky Way from region to region. Here we find a
regular river of nebulous matter rushing among thousands of stars,
elsewhere the galaxy seems to tie itself in knots. There is an indivi-
duality in almost every part of it, which we can study on our photo-
graphic plates ; practically there are no two parts alike. Other
photographs bring before us the curdled appearance which is visible in
different regions, and finally the connection of the infinite number of
stars with obvious nebulous matter. In this way, then, we are enabled
to form an idea of the general conditioning of things as we approach
the Milky Way.

The next important point is that the enormous increase of stars in
the Milky Way is not limited to the plane itself, but that there is
really a gradual increase from the poles of the Milky Way, where we
find the smallest number of stars. It is not very easy to bring together
all the information, for the reason that different observers give different
measures ; they take different units for the space they have determined
to be occupied by stars from the pole towards the galactic plane ; and
also the number of stars in the northern hemisphere is not the same
as the number in the southern hemisphere. But roughly speaking we
may say, if we represent the number of stars at the galactic pole by
four, the number of stars in the galactic plane will be about fifty-four.

126

INORGANIC EVOLUTION.

[CHAT.

The following table will show the gradual increase in the number
of stars from the pole to the plane, as seen by the Herschels with a
reflecting telescope of 18 inches aperture and 20 feet focal length : — *

Average number of stars per field

of 15'.

Galactic polar distance.

Northern.

Southern.

0°-15°

4-32

6-05

15-30

5-42

6-62

30-45

8-21

9-08

45-60

13-61

13 -49

60-75

24-09

26-29

75-90

53-43

59-06

A consideration of the distribution of stars in right ascension
between declinations 15° N: aud 15° S., led Struve to the conclusion
that there are well marked maxima in K.A. 6 hrs. 40 mins. and 18 hrs.
40 mins., and minima in E.A. 1 hr. 30 mins. and 13 hrs. 30 mins.;
he remarks that the maxima fall exactly on the position of the Milky
Way in the equator, and further states that " the appearance of the
close assemblage of stars or condensation is closely connected with the
nature of the Milky Way, or that this condensation, and the appear-
ance of the Milky Way, are identical phenomena."

Although the Milky Way dominates the distribution of stars, and
especially of the fainter stars, it does not appear to be the only ring
of stars with which we have to do. Sir John Herschel traced a zone
of bright stars in the southern hemisphere, which he thought to be the
projection of a subordinate shoot or stratum. That was the first
glimpse of a new discovery, which was subsequently established by
Dr. Gould in' his work in the southern hemisphere at. Cordova. He
found that there was a stream of bright stars to be traced through the
entire circuit of the heavens, forming a great circle as well defined as
that of the galaxy itself, which it crossed at an angle of about 25°.

Gould, while in the southern hemisphere, had no difficulty in
observing that along this circle, which we may call the Star Way, in
opposition to the Milky Way, most of the brighter stars in the southeni
heavens lie.

When he subsequuently came home he made it a point of study to
see whether he could continue this line of bright stars completely
through the northern hemisphere, and he found no difficulty. So that

*' Outlines of Astronomy, Herschel, pp. 535, 536.

XV.] THE GENERAL DISTRIBUTION OF STARS. 127

we may now say that the existence of this supplementary Star Way,
indicated by the line of extremely bright stars, is beyond all question.

I quote the following from what Gould has written on this
subject : — *

" Few celestial phenomena are more palpable there than the ex-
istence of a stream or belt of bright stars, including Campus, Sirius,
and Aldebaran, together with the most brilliant ones in Carina, Puppis,
Columba, Canis Majw, Orion, &c., and skirting the Milky Way on its
preceding side. When the opposite half of the galaxy came into view,
it was almost equally manifest that the same is true there also, the
bright stars likewise fringing it on the preceding side, and forming a
stream which, diverging from the Milky Way at the stars a and ft
Centauri, comprises the constellation Lupus, and a great part of Scorpio,
and extends onwards through Ophiuchus towards Lyra. Thus a great
circle or zone of bright stars seems to gird the sky intersecting with
the Milky Way at the Southern Cross, and manifest at all seasons,
although far more conspicuous upon the Orion side than on the other.
Upon my return to the North, I sought immediately for the northern
place of intersection ; and although the phenomenon is by far less
clearly perceptible in this hemisphere, I found no difficulty in recog-
nising the node in the constellation Cassiopeia, which is diametrically
opposite to Cnw. Indeed it is easy to fix the right ascension of the
northern node at about 0 hr. 50 mins., and that of the southern one
at 12 hrs. 50 mins.; the declination in each case about 60°; so that
these nodes are very close to the points at which the Milky Way ap-
proaches most nearly to the poles. The inclination of this stream to
the Milky Way is about 25°, the Pleiades occupying a position midway
between the nodes."

Gould also had no difficulty in showing that the group of the fixed
stars to which I have just referred, at all events of fixed stars brighter
than the fourth magnitude, is more symmetrical in relation to this new
star line than to the Milky Way itself, and that the abundance of
bright stars in any region of the sky is greater as the distance from
this new star line becomes less. Practically 500 of the brightest
stars can be brought together into a cluster, independent of the
Milky Way altogether — a cluster he points out of somewhat flattened
and bifid form.

Connection of the Milky Way with Nebulce.

Not only do we find that the stars are very much more numerous
near the Milky Way than elsewhere, but that the same thing happens

* Amer. Jour. Sci., vol. Tiii, p. 332.

128 INORGANIC EVOLUTION. [CHAP. XV.

with regard to the planetary nebulae. Nebulae generally we cannot at
present discuss with any advantage, because there are very many bodies
classed as nebulae in the different catalogues about the physical natures
of which we know absolutely nothing. I shall only call attention to
those points about which we can be most certain.

Not only do we find stars and planetary nebulae increasing in
number as the Milky Way is approached, but the undoubted star clus-
ters also increase towards the Milky Way in a marvellous manner.

BauschingerJ (1889) in a review of Dr. Dreyer's "New General
Catalogue" (7,840 objects), discussed the distribution of different
classes of objects and found that star clusters, by which he means of
course resolved clusters, and planetary nebulae congregate in and near
the galaxy.

Mr. Sydney Waters some four years later, in 1893, brought
together the nebulae and the star clusters on maps which showed, in a
most unmistakable manner, that the star clusters, like the planetary
nebulae and stars generally, are very much more numerous in the plane
of the Milky Way than they are in any other part of the heavens.

It is striking to note the fidelity with which the clusters follow
not only the main track of the Milky Way, but also its convolutions
and streams, while the remarkable avoidance of the galaxy by the
nebulae, excluding the planetary nebulae, is obvious ; it was indeed
noted by Sir Wm. Herschel.

We have seen, then, that the greatest number of stars congregate
in the plane of the Milky Way, and the greatest number of planetary
nebulae and the greatest number of star clusters.

* F. J. S. Ast. Ges., vol. xxiv, p. 43.

129

CHAP. XVI.— THE DISTRIBUTION OF CHEMICAL GROUPS OF STARS.

A. In Relation to Direction.

THE most convenient way to consider the distribution of the various
chemical groups of stars, is to take the plane of the Milky Way as a
base, as we have already done regarding the stars merely as masses
of matter independently of all chemistry, and to note whether any
particular chemical species of stars congregates in the Milky Way or
avoids it. In this way the new molecular inquiry will be on all fours
with the older molar one.

I will begin by leaving distances out of consideration.

At present it will be sufficient for our purpose to deal with the
more generalised classification (already given on p. 72), which is as
follows : —

Highest Temperature.

f Proto-hydrogen stars.
Gaseous stars < J 5

I Cleveite-gas stars.

Proto-metallic stars.

Metallic stars.

Stars with fluted spectra.

Lowest Temperature.

In discussing the work of other observers I have, as far as possible,
transposed the different notations employed into the chemical one given
above, and in some cases the two arms of the temperature curve will
require to be considered.

The first attempt at such an inquiry as this was made in 1884, by
Duner,* who had made himself famous by his admirable observations
on two different classes of stars — those which I have referred to as
being defined by carbon flutings in one case, and metallic flutings in
the other. His work was practically the only research on the carbon
stars — the stars, that is, with carbon flutings. He was, naturally,
anxious to see how they were distributed, and he gave the number of
these stars in varying parts of the heavens in relation to the Milky
Way. He found that the numbers increased towards the Milky Way.
The table. I give will show the general result at which he arrived.

* fitoiles de la troisieme Classe, p. 126.

K

130

INORGANIC EVOLUTION.

[CHAP.

We saw in the case of the ordinary stars that a very rapid pro-
gression in number is to be noticed from the pole of the Milky Way
to the plane ; we had three stars at the pole when we had fifty-three in
the plane.

Distance
from galactic pole.

Number.

Mean magnitude.

0°-35° 3

6-6

35-60

8

6-6

60-70

8

7'2

70-80

13

7'4

80 -90 29

8-3

Duner found, with regard to his carbon stars, that there was dis-
tinctly an increase from the pole towards the plane, but we observe
that the rate of increase is very much less in this case ; so that,
starting with three at the pole, he only found twenty-nine in the plane.
Although then it was true that the number of stars did increase
towards the Milky Way, they did not increase so rapidly as the stars
taken as a whole ; still, from his observations, we are justified in
stating that there is an increase as we approach the plane of the Milky
Way. They are, therefore, not limited to the plane.

POLE90

LKT

!£. to so vo 5b 60 >a-

Fia. 38. — Comparison of relative numbers of stars generally and carbon stars.

That I was in 1884. In 1891 Professor Pickering, when he found
that he had collected something like 10,000 stars in the Draper cata-
logue, began to consider their distribution in different parts of space
in relation to the then classification, which was practically one founded
on hieroglyphics, since we knew very little about the chemistry of the
different bodies at that time.

XVI.] THE DISTRIBUTION OF CHEMICAL GROUPS OF STARS. 131

He found that the Milky Way was due to an aggregation of white
stars, by which he meant, as we now know, very hot stars, and the
hottest of them, that is the gaseous ones, exist more obviously in the
Milky Way than do the others. The proportional number of proto-
metallic stars in the Milky Way was greater for the fainter stars than
for the brighter ones of this kind, and that at once suggests a possi-
bility that in the Milky Way itself there is a something which
absorbs light ; so that the brightest stars are apt not to be really the
brightest, but apparently bright because they have not suffered this
absorption, and that those which have suffered this absorption may be
very much further away from us than the others of a similar chemistry.
He also arrived at this extremely important conclusion, namely, that
the metallic stars, that is, stars like our sun, stars more or less in their
old age, had no preference for the Milky Way at all, but are equally
distributed all over the sky. With regard to the group of stars known
by metallic flutings in their spectra, he has no information to give us
any more than Duner had, for the reason that their number is small,
and they have not yet been completely studied.

Only last year this inquiry was carried a stage further by
Mr. McClean, who not only photographed a considerable number of
stellar spectra in the northern hemisphere, but subsequently went to
the Cape of Good Hope in order to complete the story with reference
to the stars down to the third or fourth magnitude which he could
observe there. He was very careful to discuss, in relation to the Milky
Way and certain galactic zones, the distribution of the various kinds
of stars which he was fortunate enough to photograph.

He found that if we deal with the gaseous stars the numbers in the
north and south polar region are small, and that the numbers nearer
the Milky Way are greater, so that finally we can see exactly how these
bodies are distributed. If we take the gaseous, that is to say the
hottest stars, we find the smallest number in the polar regions ; but if
we take the metallic stars we find practically the largest number, at all
events a considerable number, in the polar regions. The general result,
therefore, is that the gaseous stars are mostly confined to the galactic
zones, the proto-metallic stars, that is those down to about 3J mag-
nitude, are not so confined. What is also shown is that the metallic
fluting stars are practically equally distributed over the polar regions
and over the plane of the Milky Way itself ; so that, in that respect,
we get for these stars very much the equivalent of the result arrived
at by Duner for the carbon stars, that is to say, they have little pre-
ference for the Milky Way.

K 1

132 INORGANIC EVOLUTION. [CHAP.

Bright-line Stars.

These, then, are the results with regard to the stars having obviously
dark lines in their spectra, but besides these there are many so-called
bright-line stars.

I should say that there has necessarily been a change of front in our
views with regard to these bright-line stars since they were first classi-
fied with nebulae. The nebulae are separated generic-ally from the stars
by the fact that in their case we have to deal with bright lines, that is
to say, we deal only with radiation phenomena, and not with absorp-
tion phenomena, as in the case of the stars so far considered ; and in
the first instance it was imagined that the bright-line stars were, from
the chemical point of view, practically nebulae, although they appeared
as stars, because the brightest condensations of them were so limited
or so far away that they gave a star-like appearance in the telescope.

Since that first grouping of bright-line stars, by the work chiefly of
the American astronomers, it has been found that in a large number
of cases they hove also dark lines in their spectra, and that being so we
must classify them by their dark lines instead of by their bright ones ;
and the bright-line stars thus considered chiefly turn out to be gaseous
stars, with a difference. What is that difference 1 It is this, I think:
in the case of the bright-line stars we are dealing with the condensa-
tions of the most disturbed nebulae in the heavens, together with the
light which we get from the nucleus of that nebula which appears as a
star, and can be spectroscopically classified with the other dark-line
stars, inasmuch as the surrounding vapours close to the star produce
absorption, and therefore give us dark lines ; other parts of the nebulae,
probably those further afield, give us bright lines which mix with the
dark ones. Therefore we get both bright lines and dark lines under
these conditions. So far as the result goes up to the present moment,
it looks as if we have now to consider that these bright-line stars, instead
of being nebula? merely, are gaseous stars at a very high temperature,
in consequence of the fact that the nebula which is surrounding them,
which is falling upon them, is increasing the temperature of the central
mass by the change of vis vim into heat. Pickering,* in his discussion
of these stars, had thirty-three to deal with, and he found that there
was a wonderful tendency among these to group themselves along the
Milky Way : that very few of them, in fact, lay outside its central
plane ; the galactic latitude, the distance in degrees from the plane
being limited in the generality to only 2°, and the greatest departure,
the greatest galactic latitude, was something within 9°. That was the
story in 1891. Two years afterwards Campbell, another distinguished

* Astr. Nach., No. 2025.

XVI.] THE DISTRIBUTION OF CHEMICAL GROUPS OF STARS. 133

American astronomer, also interested himself in this question of the
bright-line stars, and he discussed them, his catalogue containing fifty-
five as opposed to Pickering's thirty-three. He found also that they were

•J

2

vo

•

•

c

—

c r

collected almost exclusively in the Milky Way, and that outside the
Milky Way . practically none had ever been observed. The importance
of this result I will indicate by and by. The . central line of the map

134 INORGANIC EVOLUTION. [CHAP.

(Fig. 39) represents the galactic zone, the plane of the Milky Way, and
along it the different galactic longitudes are indicated, above and
below the plane a few degrees of galactic latitude north and south
are shown, sufficient to enable all the bright-line stars which Campbell
discussed to be plotted. The map shows that all the bright-line stars

FIG. 40.— Photograph of a glass globe showing the relation of the Milky Way to
the Equator and to Gould's belt of stars.

really are close to the central plane of the Milky Way. Only one out
of the fifty-five is more than 9° from it, and this lies in a projecting
spur, so that we cannot really say that that is out of the Milky Way.

It is remarkable that these bright-line stars are not equally dis-
tributed along the Milky Way. They are chiefly condensed in two oppo-
site regions, and there is one region in which they are markedly absent.

XVI.] THE DISTRIBUTION OF CHEMICAL GROUPS OF STARS. 135

Figs. 40 and 41 are photographs of a glass globe, on which are
indicated the Milky Way; the secondary Milky Way, which starts
from it at one point of the heavens-and meets it again, is also shown ;
together with Gould's Star Way and the equatorial plane. The dark
wafers indicate the positions of the bright-line stars.

Fia. 41. — The Milky Way, where double in relation to the Equator and Gould's
belt of stars, showing that the bright-line stars (dark wafers) and new stars
(white wafers) are limited to the Milky Way.

We find that these stars begin just before the doubling commences.
They continue along the plane, and are sometimes very numerous, and
they end just after the doubling ends ; and we notice there is a long
range of the Milky Way where it is single in which there is absolutely
no bright-line star at all. It looks, therefore, very much ae if there is

136 INORGANIC EVOLUTION. [CHAP.

a something connected with this doubling of the Milky Way which
produces the conditions which generate these bright-line stars.

By the labours of Duner, Pickering, McClean and Campbell, we
are beginning to get very definite notions as to the distribution of the
various chemically different stars in relation to the Milky Way. As
I have already noticed, there can be no question as to the intimate
association of the bright-line stars with nebulae. We must next then
consider the nebulae from the point of view of chemical distribution,
but here we are somewhat in a difficulty.

I have already stated that with regard to the general question of
the nebulae it is impossible to speak with certainty, because at present
there has not been sufficient time and there has not been a sufficient
number of observers at work to classify the thousands of " nebula? "
which we now know of into those which give us the gaseous spectrum
and those which are entirely different, apparently, in their constitution,
and only give us what is called a continuous spectrum. Still we can
go a little way in this direction by means of some figures which I have
noted. The point is to see whether there is any difference in the dis-
tribution of those nebulae which are undoubtedly masses of gas, which
give us the so-called nebulous spectrum, and those other nebulae about
which at present we know very little, which give us so-called continuous
spectra. It is clear that on this point undoubtedly, at some future
time, a great deal will be learned. The figures I give bring the results
up to the year 1894. If we take the region near the Milky Way, the
region bounded by 10° galactic latitude north and south, and consider
the planetary nebulae, we find that there are forty-two ; but if we deal
with those which are further than 10° from the Milky Way, that
number drops to five. If we take other nebulae, not necessarily
planetary but gaseous like planetary nebulae, inasmuch as they give us
a spectrum of bright lines, we find that there are twenty-two in or near
the Milky Way, and only six outside. If we take the so-called nebulae
known to have continuous spectra, which need not be nebulae at all—
we only imagine them to be nebulae because they are sa far away that
we cannot get a really true account of them — we find that the condi-
tions are absolutely reversed. There are only fourteen of them in the
plane of the Milky Way, but there are forty-three lying outside it ; so
that the percentage within 10°- of the Milky Way comes out to be
eighty-four in the case of the planetary and the other nebulae which
give us bright lines, and in nebulae with continuous spectra only twenty-
five. Therefore we get an absolute identity of result with regard to
the bright-line stars and the other objects which give us bright-line
spectra.

There is another class of bodies of extreme interest. In fact, to

XVI.] THE DISTRIBUTION OF CHEMICAL GROUPS OF STARS. 137

some they are more interesting than all the other stars in the heavens,
because they are the mysterious " new stars," which have been supposed
to be new creations. When we come to examine these so-called new
stars we find that they also are almost absolutely limited to the Milky
Way. Our information begins 134 years before Christ, and it ends
last year. The number of stars thus reported as new stars is thirty-
one, and of these only three have been seen outside the Milky Way.
Fig. 40 shows what the facts are with regard to the new stars. The
bright-line stars being distinguished by dark wafers, the new stars are
shown by white wafers. We notice that where we get practically the
greatest number of dark wafers we get a considerable number of white
ones. That means that these new stars take their origin in the same
part of space as that occupied by the bright-line stars, and it is also
interesting to point put that the void indicated where the Milky Way
is single, where there were no bright-line stars, is equally true for the
new stars; only one new star has been recorded in this region
(Fig. 41).

As I have said, a great deal of interest has been attached by many
people to the question of the new stars, for the reason that whenever a
new star appeared in a part of the heavens where no star was seen
before, it was imagined that something miraculous and wonderful had
happened. That was justifiable while we were ignorant, but recent
work has shown, I think almost to a certainty, that the real genesis of
a new star is simply this. We have near the Milky Way a great
number of nebulae, planetary or otherwise • we have more planetary
nebulae near the Milky Way than in any other part of the heavens ;
the nebulous patches also observed in it may include streams of
meteorites rushing about under the influence of gravity ; the origin of
a new star is due to the circumstance that one of these unchronicled
nebulae suddenly finds itself invaded by one of these streams of meteor-
ites. There is a clash. These meteorites we know enter our own
atmosphere at the rate of thirty-three miles a second, and we may
therefore be justified in assuming that any meteoritic stream in space,
even in the Milky Way, would not be going very much more slowly.
If we get this rapidly-moving stream passing through a nebula, which
is supposed to be a mass of meteorites more or less at rest, of course we
must get collisions ; of course, also, we shall get heat, and therefore
light. When the stream has passed through the nebula the luminosity
will dim and ultimately, attention having been called by this cataclysm
to that particular part of space, we shall find that there is a nebula
there. This has always been so ; and therefore in the case of new stars
we must always expect to get indications of the existence of two bodies,
the intruder and the body intruded upon.

138 INORGANIC EVOLUTION. [CHAP.

We must also expect, if we are dealing with small particles of
meteoritic dust, .that the action will be very quick, and that the war
will be soon over. All this really agrees with the facts. In the case
of the new star we were fortunate enough to have the opportunity
of observing in the northern hemisphere, not very long ago, the new star
in the constellation Auriga, we obtained undoubted indications of the

K H h G- F

Fia. 42. — The spectrum of Nova Aurigae, showing both bright and dark lines.

fact that we were dealing with two different masses of matter ; for the
reason that if we take the chief spectral lines marked G, h, H and K
(Fig. 42), that is to say, the lines of hydrogen and of calcium, we find
both bright lines and dark lines, which being interpreted means that
hydrogen and calcium were both giving out light and stopping light.
We cannot imagine that the same particles of calcium and of hydrogen
were both giving out light and stopping light ; there must have been
some particles of hydrogen and calcium giving light and others stopping
light ; and if we look at the photograph carefully we find that the
bright lines and the dark lines are side by side, and we know that that
means a change of wave-length in consequence of movement, and we
also know from the change of wave-length indicated that the differential
velocity of the particles which gave us the bright hydrogen and calcium,
and the dark hydrogen and calcium, must have been something like
500 miles a second. In that way we obtained indisputable proof that
we were really dealing with two perfectly different series of particles
moving in opposite directions, and that that was the reason we got that
sudden illumination in the heavens which as suddenly died out until
finally a nebula previously undiscovered was found to occupy the place.
The nebula is really not the result, the nebula was the cause, but we
did not know of its existence until our special attention had been
drawn to that part of the heavens.

B. In relation to Distance.

So much, then, for the first statement of facts relating to the dis-
tribution of the various star groups and nebular groups in the most
general form. The next question is, can we say anything about the
distances of these bodies *?

The way in which an astronomer attempts to determine the dif-
ferent distances of the various stars from the earth, may be very well

XVI.] THE DISTRIBUTION OF CHEMICAL GROUPS OF STARS. 139

grasped by considering what happens to any one, travelling in a
railway train. If the train be going fairly quickly, and we look
at the near objects, we find that they appear to rush by so rapidly that
they tire the eye ; the more distant the object we look at is the more
slowly it appears to move, and the less the eye is fatigued. Now, sup-
pose that instead of the train rushing through the country and passing
the objects which we regard under these different conditions, the dif-
ferent objects are rushing past us at rest. Then, obviously, those
things which appear to be moving most quickly will be those nearest,
and the more distant objects, just because they are distant, will appear
to move more slowly ; that is to say, we shall get what is called a large
" proper motion " in the case of the objects nearest to us, and a small
" proper motion " in the case of the bodies which are further away.

This question has been attacked with regard to the stars in mag-
nificent fashion by a great number of astronomers.

It was Mr. Monck who was the first to show in 1892* that the
gaseous stars had the smallest proper motion ; that is to say, that the
hottest stars were further away from us than the cooler ones. He
next found that the proto-metallic stars — that is to say, the stars not
so hot as the gaseous, but hotter than the metallic stars — had the next
smaller proper motion. This, of course, indicates that the metallic
stars are the nearest to us unless proper motion does not depend upon
distance, but rather upon a greater average velocity in space. It has
been shown, however, by considering the sun's movement in space, that
this view probably may be neglected. The first discussion of proper
motion, then, went to show, roughly, that the hotter a star is the
further away from us it is ; and it made out a fair case for the conclu-
sion that the sun forms one of a group or cluster of stars in which the
predominating type of spectrum is similar to its own.

Kapteyn carried the inquiry a stage further.! Working upon the
idea that stars with the greatest proper motion are on the average the
nearest, the part of the piroper motion due to the sun's translation in
space he considered must depend strictly upon the distance, and he
determined this by resolving the observed proper motion along a great
circle passing through the point of space towards which the sun is
moving, which is called the apex of the sun's way, and reducing to a
point 90° from the apex. His results were practically the same as
those obtained by taking the individual proper motions. He also
found that stars with the greatest proper motion are mainly metallic,
and have no regard at all to the Milky Way ; that stars with the
smallest and no observable proper motion are gaseous and proto-

* Astronomy and Astro-Physics, vol. xviii, 2, p. 876.
f Amsterdam Academy of Science, 1893.

HO

INORGANIC EVOLUTION.

[CHAP. xvi.

metallic, including a few metallic ones which have collected in the
galactic plane. In this he agrees with the prior observations to which
I have drawn attention. In the table which I now give the mean
proper motion is shown.

Relation between Spectra and Proper Motions of Stars (Kapteyn).

Mean proper
motion.

Gaseous and \
proto -metallic
stars.

Metallic

stars.

Metallic
flutings.

Ratio,
metallic to
gaseous.

1-39

3

51

17'0

0-52

12

66

]

5'5

0-35

14

66

—

4-7

0-24

34

124

—

3'6

0-18

35

67

3

1-9

Inappreciable

79

35 1

0-44

j

We find that the gaseous and proto-metallic stars increase in num-
ber as the proper motion decreases. We find also the ratio of the
metallic to the gaseous and the proto-metallic. We begin with a ratio
of 17, and end with something like a ratio of O4; so that the results
may be considered to be pretty definite. These results were obtained
by Kapteyn with 591 stars which were common to Stumpe's catalogue
of proper motions and the Draper catalogue dealing with spectra.
The general result may, therefore, be stated that at the nearest dis-
tance the metallic stars are seventeen times more numerous than
gaseous stars, and at the greatest distance they are not half the
number.

Here again the question arises, how far the intrinsic brightness of
these bodies, in relation to their distance from us and the possible
greater or less extinction of light in space, has to be taken into con-
sideration. That is a problem which will require a considerable
amount of work in the future. It is rather remarkable that if we
take the stars with very great proper motion, very much greater than
the average, we find with regard to four that three of them are
undoubtedly metallic, but it is possible that the star 1830Groombridge,
which is always looked upon as the star which beats the record in
velocity seeing that it would travel from London to Pekin in about
two minutes, is not a metallic star.*

* These stars are —

1830 Groombridge
2 2758 . .
S 578 . .
B.C. 583 . .

7 '04 . . Gaseous or proto-metallic.

5 -196 . . Metallic.

4-0<!9 .. Probably metallic.

3-7 - . Metallic.

141

CHAP. XVII. — THE RESULT OF THE INQUIRY.

WE are finally in a position to make a general summary of the dis-
tribution of the various chemical groups of stars not only in relation
to their direction in space, as seen from the solar system, a direction
most conveniently considered in relation to galactic latitudes and
longitudes, but also in relation to their distance from us.

The results arrived at in the two previous chapters may be sum-
marized as follows. First we will consider the stars studied by their
absorption phenomena.

Group.

Kelation to Milkv Way.

Proper motion.

Gaseous stars
Proto-metallic

Metallic

Metallic flutings.
Carbon .

Condensed in Milky Way

(Pickering and McClean)
Brighter ones not notably

condensed in Milky Way

(McClean)
Tend to collect in Milky Way

more especially the fainter

stars (Pickering)
Not condensed in Milky Way

(Pickering and McClean)
Collected in Milky Way

(Kapteyn)

Smallest* (Monck).
Intermediate (Monck).

Div. 1. Greatest (Kapteyn).
Div. 2. Small (Kapteyn).

We find that the gaseous stars are chiefly in the Milky Way and
are far away from us ; that the proto-metallic stars are not so confined
to the Milky Way, and they are not so far away from us. But when
we come to the metallic stars and the carbon stars they have not much
obvious connection with the Milky Way, and they are close to us.
Unfortunately, with regard to the metallic fluting stars the informa-
tion is not so complete. Mr. McClean has dealt with a very small
number, and he shows that they, like Duner's stars, the carbon stars,
have very little relation to the Milky Way. We thus obtain a tre-
mendous separation between the hot stars with their great distance
and the cooler stars with their smaller distance.

* Kapteyn finds small proper motions for gaseous and proto -metallic stars, but
does not separate them into two groups.

142 INORGANIC EVOLUTION. [CHAP.

Although this discussion of the distribution of different types of
stellar spectra indicates a collective tendency of some types, it proves
at the same time that the chemical substances represented in such
types are distinctly not limited to the regions in which they pre-
dominate. Thus we know of hydrogen in all stars except the carbon
stars ; although the stars showing strong indications of helium are most
numerous in and about the Milky Way, stars of this kind do appear
in other parts of space remote from the Milky Way, among them
being the bright stars Spica and 77 Ursse Majoris. Besides this direct
evidence of the wide diffusion of helium there is the indirect evidence
based upon the fact that helium is known to be present in the sun
although it is not represented among the Fraunhofer lines. By
analogy then we must allow that helium is also present in Arcturus
and the thousands of other stars which have spectra like the sun which
have no special connection with the Milky Way. Helium must, there-
fore, be practically like hydrogen, distributed in all directions as seen
from the sun.

Another illustration of this general diffusion of a particular kind
of matter is afforded by carbon. In the hottest stars, stars like the
sun, and the coolest stars, we alike find indications of this substance,
so that a localisation of any particular type of star does not imply the
restriction of carbon to such localities. Again, if we take iron, we
find its indications, either as iron or proto-iron, through a great variety
of stellar types, while we may say that calcium and magnesium show
direct evidence of their presence in almost every star.

Thus we are led to conclude that there is no localisation of the
chemical elements so far as direction in space is concerned. While the
discussion of proper motion indicates that particular types of stars
tend to congregate at distances peculiar to themselves, the condensa-
tion is by no means absolute. Some stars of each type have proper
motions widely different from the average. Hence at all distances
from us we find similar chemical types of stars and therefore evidence
of similar chemical substances.

We have already seen that the chemistry is the same in all direc-
tions, so that, finally, we must grant that the chemistry of all parts of
space is the same. In other words the chemical parishes required by
the view that the stellar types represent different chemical conditions
as regards the presence or absence of certain substances do not exist.
In no direction from our system, in no shell surrounding it, is any
chemical element found which is not present in other directions and in
other shells.

The major objection then against the stellar evidence in support
of the dissociation hypothesis, upon inquiry, vanishes into thin air.

XVII.] THE RESULT OF THE INQUIRY. 143

Our lengthened consideration of this question has really led us to
a firm support not only of the dissociation hypothesis but of the meteo-
ritic hypothesis as well.

As on the latter hypothesis the stars become hot in consequence
of meteoritic collisions, we should expect to find nebulous conditions
following suit ; seeing that nebulae are masses of meteorites, we should
expect to find especially the gaseous nebulae and results depending
upon their presence in the region where the hottest stars exist in
which dissociation has been studied.

The planetary nebulae consist of streams of meteorites moving
generally in spirals or in circular paths. There iis no very great dis-
turbance; we get a bright line spectrum from them, and we know
they are practically limited to the Milky Way. We have found that
the bright-line stars are limited to the Milky Way ; they are simply
stars involved in nebulae. There again we get a connection between
the Milky Way and nebulae. The new stars are due to relatively
fixed nebulae driven into by moving nebulae comet fashion, and they
are also limited practically to the Milky Way ; there again we have
the nebulous touch. The nebulous regions, which Sir William
Herschel was the first to chronicle, are more prevalent near the Milky
Way than elsewhere.

It will be seen that we have a strict association of nebulae, possible
dissociation conditions, and the hottest stars in which that dissociation
has been studied; and we are at length face to face with a simple
explanation of the close contiguity of these apparently very diverse
phenomena.

144

CHAP. XVIIL— REPLIES TO SPECIAL OBJECTIONS.

I NOW proceed to consider some less general objections. When I
brought the question of dissociation before the Royal Society in 1897,
in a discussion which I was requested to initiate, I pointed out that it
had been proposed to explain the spectral differences between such stars
as Bellatrix with its hydrogen and cleveite gases ; Sirius with its
tremendous development of hydrogen ; and our own sun and stars like
it with an atmosphere chiefly metallic ; by supposing that " the hydro-
gen and cleveite gases may from some reason or other escape from
among the metallic vapours and form an upper special atmosphere of
their own, in which, in consequence of its greater chemical simplicity,
the lines of these substances will become more important,"* and I
added, " But this argument is not philosophical, because we have no
right to assume such a change."!

This remark, referring to a very special point, was .unfortunately
misheard, and Dr. Schuster in the discussion stated : —

" Had Mr. Lockyer confined himself to bringing forward his hypo-
thesis as one which is legitimate, consistent, and deserving of attention,
many of us would I think have agreed that he had made out a good
case. But he claims his theory as the only one which can explain the
facts, and dismisses as unphilosophical the only alternative which he
discusses."

In spite of this misapprehension, however, Dr. Schuster's criticisms
are of great value, and I propose to consider them in this place and
reply to them as best I can. I may add that he expresses his concur-
rence with my system of classification ; and the necessity of a constant
appeal to laboratory experiment is insisted upon ; at the same time
he acknowledges that the investigation of the enhanced lines is "a
very material advance."

In my paper I pointed out, in relation to stellar atmospheres, that
what we might expect to observe if we assumed the sun's temperature
to be increased would be vastly different according as dissociation did
or did not take place (see pp. 78-9). I said : —

•' The only change which we can imagine on the usual hypothesis, as
resulting from the increase of temperature, is tJmt with the increase in

* Proc. Soy. Soc., vol. Ixi, p. 202.
f Loc. cit.

CHAP, XVIII.] REPLIES TO SPECIAL OBJECTIONS, 145

'volume ikere, will be a reduction in density, and all the lines will be equally
tufecbled. But this is exactly what does not happen."
With regard to this statement Dr. Schuster writes :—
" With this remark I cannot agree. The main fact to be explained
is the gradual displacement of hydrogen, which is predominant in the
hottest stars, by calcium, iron, and other metals. There are in my
opinion several causes at work which might produce that effect. A
glowing mass of gas may be either in thermal or in convective equili-
brium, and the spectroscopic appearances in the two cases will be pro-
foundly different. In reality an intermediate state probably is arrived
at, but there is good evidence to show that the state of convective
equilibrium is more nearly approached in our sun than in the hydrogen
stars. We know as a fact that there are powerful convection currents
near the sun's surface. There is, in consequence, an approach to a
uniform distribution of matter and enormous differences of temperature
in layers which are comparatively close together. Those who have not
given much attention to this subject will hardly realise the differences
of temperature brought about by convection currents. On the surface
of the sun the temperature gradient produced by convection currents
would be equal to 20,000° for each 100 kilometres difference in level,
so that an angular distance of one second of arc would correspond to a
difference of 100,000°. Radiation and condensation will diminish this
gradient, but that it is very large is sufficiently proved by the spectro-
scopic evidence. Thus, according to the results of Messrs Jewell,
Mohler, and Humphreys,* the pressure in the reversing layer for hot
calcium giving the H and K lines is about six atmospheres, while that
for the cooler calcium vapour is about three atmospheres. With a
gravitational constant twenty-seven times as large as that of our earth,
a difference of three atmospheres can only mean a comparatively small
difference in level ; while, then, in the sun we must admit a more or less
effectual stirring up of the constituents together with an accompanying
rapid temperature gradient, the evidence is just the other way in the
case, of stars like 7 Lyrae. The spectrum of that star, according to
Professor Lockyer, contains only the high temperature lines of iron.
This means not only that the reversing layer is very hot, but also that
there are no rapid changes of temperature at different levels. It is
impossible to imagine this hot layer of gas ending abruptly ; it must
be surrounded by cooler matter, which cannot be iron, as the low tem-
perature lines of iron do not appear. In such a star there cannot be an
effectual mixing up of the constituents, and hence the layers of gas will
arrange themselves according to the laws of diffusion. It would follow
that hydrogen, being a lighter gas than iron, will be chiefly represented
* Astropkytical Journal, yol. iii, p. 138.

c.

146 INORGANIC EVOLUTION. [CHAf».

in the cooler and outer layers, while iron will be found more particu-
larly in the inner and hotter parts. The relative proportion of different
elements in different layers will be regulated partly by their density,
but to a great extent also by the total quantities present in the star ;
for the different gases will not float on each other as liquids might, but
the density of each gas will increase steadily from the surface to the
centre. The chief difference, according to this view, between a hydro-
gen and a solar star lies in the more or less effectual mixing up of the
constituents. If we could introduce a stirrer into 7 Lyrse there can be
no doubt whatever that the low temperature lines of iron would make
their appearance, while, on the other hand, if we could stop all convec-
tion currents on the surface of the sun the hydrogen which now lies
under the photosphere would gradually diffuse out and give greater
prominence to its characteristic absorption lines."

" In the face of the direct evidence of the absence of convection
currents in the hotter stars, it is not necessary for the purpose of my
argument to discuss why this is the case, but it can be seen that
diminished gravity, diminished density, and consequently increased
viscosity, will contribute to the effect, while effectual radiation will,
owing to the smaller density, take place more evenly through a thicker
layer of the envelope, so that the principal cause of convection currents
will also be much diminished."

In replying to this objection of Dr. Schuster's I will first deal with
the convection currents and the tremendous temperature gradient
which Dr. Schuster postulates. In the sun, the seat of such convection
currents, according to him, while they are absent from y Lyrse they
are sufficiently powerful to cause a difference of 20,000° C. for each
100 kilom. in difference of level, or, as he otherwise puts it, a
difference of 100,000° for one second of arc.

The eclipse photographs give no evidence of the rapid temperature
gradient in the sun supposed by Dr. Schuster. In the Indian series,
two successive photographs taken at intervals of about one second near
the beginning of totality differ inasmuch as the first includes a
stratum about 150 miles above the photosphere, which would be
covered by the moon when the second was taken (except for the effect
produced by irregularities in the moon's limb). Yet there is no great
difference in the spectra ; both contain arc and enhanced lines about
equally, and therefore indicate that the temperature changes can only
be small in a depth of 150 miles. In fact throughout a distance of
500 miles above the photosphere the spectrum indicates no change of
temperature of importance.

We have got the facts then in the eclipse photographs, and find no
large spectral changes in a region where Dr. Schuster postulates a dif-

XVIII.] REPLIES TO SPECIAL OBJECTIONS. 147

fererice of 100,000° 0. Are we to take this value as the temperature
of the sun's photospheric level ? If so, how does Dr. Schuster reconcile
it with the values obtained by all the recent workers who make it less
than 10,000° C. 1 and even with 'Homer Lane's 28,000° 1

Surely the facts show that there are not-, in the sun, such tremen-
dous convection currents as are demanded on Dr. Schuster's view.

Professor Schuster refers to the conclusion drawn by Messrs. Jewell
and others as to the pressure' of hot and cold calcium in the reversing
layer. His reference shows that he agrees with my view that we are
dealing with different molecules, but I wish to 'remark that I think we
must not be too hasty in accepting the conclusions to which he refers,
for the reason that the eclipse photographs do not tally with them at
first sight. In these photographs (1898) the K layer reached a height
of 6,000 miles ; the A 4226'96 layer only 2,000 miles. This suggests
that cool calcium falls and is dissociated at the bottom. It certainly
does not mean that there is a layer of cooler calcium at a higher eleva*
tion and at less pressure surrounding a hotter one at a lower elevation
and higher pressure.

The evidence on which it is assumed that convection currents are
absent from the hotter stars likejy Lyrse of decreasing temperatures
does not appear to be conclusive. But let us assume it.

The absence of cool iron lines only shows that we are in a region
of higher temperature than in the sun. May there not still be a rapid
temperature gradient, from " high " to " very high " temperature
instead of from " low " to " high " as in the sun ? But in any case, a
mere stirring up of 7 Lyrse would not make its spectrum like that of
the sun. Such stirring up could only introduce the cooler lines of iron
if the proto-iron were by that process driven out into the cooler
regions, where it might become iron and so produce cool iron absorp^
tion lines in the spectrum of the star. But it by no means follows
that these cool iron lines would be as strong as in the solar spectrum,
for we know that the amount of absorbing proto-iron is only small.
Moreover, this process of stirring would hardly reduce the intensity of
the hydrogen lines.

A reduction of temperature, however, furnishes us with a sufficient
explanation of the changes observed in passing from such a star as
7 Lyrse to one like the sun ; the cool lines of iron would appear as a
matter of course, and such lines would become stronger if iron can be
formed at the expense of the hydrogen.

If we take the converse view, and suppose the postulated convec-
tion currents in the sun to be stopped, I do not see how such a condi-
tion of things would result in changing the present spectrum of the
sun into a spectrum like that of 7 Lyrae. We have not only to ex-

L 2

148 INORGANIC EVOLUTION. [CHAP.

plain the increased intensity of the lines of hydrogen, but the appear-
ance of the enhanced lines of iron as absorption lines. Now these
enhanced lines are already in the sun's chromosphere, and are pre-
sumably absent from the Fraunhofer spectrum, because the vapour
producing them approaches the temperature of the photosphere. Is it
possible that a state of quiescence in the sun would so increase the
temperature of the photosphere as to make visible the absorption of
these high temperature vapours ? And, if this be possible, there would
still be no apparent reason for the disappearance of the cool Jines of
iron. The change, however, from y Lyrae is readily explained if we
grant that there is an increase of temperature, producing proto-iron
from the previously cool iron vapour, and a dissociation capable of
producing the observed increase of hydrogen absorption at the expense
of proto-iron.

How the increased absorption of hydrogen can be accounted for
otherwise is not clear. The idea of hydrogen being set free for this
purpose from beneath the photosphere does not seem to me probable.

The final discussion of such subjects as these is very difficult,
because we learn from the sun that the absorption recorded is onl}r
that of a middle region. Neither helium nor coronium writes its
record among the Fraunhofer lines. Surely everybody will agree that
there are hundreds of substances in the higher cooler reaches of the
solar atmosphere which write no record. How then can we say that
under the conditions assumed by Dr. Schuster " there can be no doubt*
whatever that the low temperature lines of iron would make their
appearance."

Dr. Schuster also refers to hydrogen " imprisoned beneath the
photosphere " ; is there any justification for this view ? The complete
history of hydrogen, including proto-hydrogen in stellar atmospheres,
is simply and sufficiently explained on the dissociation hypothesis. I
question whether an explanation which requires such an imprisonment
of hydrogen is more satisfactory.

I now proceed to give another quotation from Dr. Schuster :—

" There is especially one question which Professor Lockyer must be
prepared to answer. Amongst the heavier metals, tellurium, anti-
mony, mercury, are not represented in the sun, but they are found in
Aldebaran. To be consistent, we must, if we adopt the theory of dis-
sociation, assert that these metals are decomposed in the sun. But, if
I understand Professor Lockyer right, he believes that with our
strongest sparks we can exceed the state of dissociation which exists
in the reversing layer of the sun. Take such a strong spark, then,
from a pole of mercury, do you get lines of helium, or of calcium, or
* The italics are mine. — X, L.

XVIII.] REPLIES TO SPECIAL OBJECTIONS. 14$.

of hydrogen 1 This seems to me to be almost a crucial experiment.
Possibly, of course, we should get high temperature lines not hitherto
looked for, but present in the sun. If so, the objection would fall to
the ground, but if this is not the case, and if mercury at a high tem-
perature refuses to be dissociated into simpler elements, a most serious
objection to the theory would have to be answered."

In reply to this I may state that in recent large dispersion photo-
graphs the differences pointed out by Dr. Schuster between the spectra
of the sun and Aldebaran do not exist. I quite agree that such experi-
ments as he describes should be made, and I have made many, but the
work which is necessary has been interrupted, since I have no longer
at my disposal the Spottiswoode coil, the superiority of which, over all
others, for such a general inquiry as this I have amply demonstrated.
I may say here, however, that so far as the observations have gone
there is apparently an agreement between the laboratory and stellar
results, but there are possible sources of error which require to be
studied, and also in a matter of such high importance the experiments
must be repeated many times before a final statement is made.

Dr. Schuster next states : —

" While I think that we shall all admit that different stars are in
different stages of development, and that hydrogen stars will ulti-
mately approach more nearly to the state of our sun, it would be
unwise to push the argument of uniformity too far, and to say that
every star will pass exactly through the same stages. Ritter, who is
favourably inclined to the dissociation hypothesis,* gives good reason
to believe that the sun's surface was never much hotter than it is
now, and that the higher temperature of hydrogen stars is connected
with their greater masses. It is, in fact, impossible to admit that the
process of development should be quite independent of the total mass
of the star. It may be urged that Arcturus must have a mass much
larger than that of our sun, and its spectrum, according to Professor
Lockyer, is identical with that of the sun. But I suppose that that
statement only refers to the blue and violet region, for, according to
Dr. Huggins, to whose early stellar photographs we owe so much, the
spectrum of Arcturus in the ultra-violet approaches that of Sirius."

Although the masses of very few white stars have been determined
with trustworthy results, one case in which a white star can be shown
to have a smaller mass than the sun will be sufficient to show a weak-
ness in Hitter's conclusions. For /3 Persei (Algol) Vogel states the
mass as four-ninths that of the sun; so that the sun, on Hitter's
theories, may be supposed to be of sufficient mass to reach a tempera-
ture as high as that of ft Persei — a result which does not accord with
* Wied. Annalen, vol. xx, p. 152.

150 INORGANIC EVOLUTION. [CHAP.

his statement that the sun has probably never been, and never will be,
much hotter than at present.

Sir William Huggins's statement as to the ultra-violet spectrum of
Arcturus is most interesting, if confirmed. The Kensington series of
large .dispersion photographs show an : almost perfect similarity of
spectrum with that of the sun, extending to X 3880.

It is difficult to see any objection, on the ground of unequal
masses, even if we grant the similarity of the two spectra. It is only
necessary to suppose that Arcturus, like the sun and other solar stars,
has passed its hotter stages, and that it may have commenced its
condensation before the sun.

To take another case, f Ursse Majoris and ft Aurigse have spectra
which are almost identical, although the masses of the two systems,
according to Pickering, are respectively 40 times and 4 '6 times that of
the sun. Another very hot star, Spica, has a mass only 2 '6 times that
of the sun.

Dr. Schuster further suggests that it if, not known to me that
Ritter has long studied the question of gaseous masses contracting
under their own gravitation. In my work which has consisted in the
discussion of spectroscopic observations, I was at the outset led to the
view that it was not a question of gaseous masses at all, originally,
and therefore I did not refer to Bitter's conclusions on this point.
Again, I had to face the spectroscopic evidence of a chain of obviously
cooling bodies, arid it was a detail to consider the fact that "a radiat-
ing and contracting mass is not necessarily a cooling mass," because
in spite of this truism a time must certainly come when all bodies will
find their temperature reduced. I am aware that Hitter's conclusions
regarding the first rise and subsequent fall of temperature of gaseous
bodies, are similar to those supported by the spectroscopic evidence of
what I have considered to be condensing swarms of meteorites, but it
would not have been fair to claim Ritter's conclusions as supporting my
own, because the bases of the phenomena considered by us were so
different.

I, perhaps, may be allowed to point out that where Ritter's conclu-
sions dp not seem to harmonise with the spectroscopic facts, it may be
that, as Professor Perry has pointed out,* a stellar atmosphere is a
more complicated thing than the theory of a gaseous mass implies.
Even the. spectroscope deals generally only with the reversing layer.
. , Professor Perry writes : —

; " He (Bitter) assumes that the radiating layer on the outside of a

s.tar is of constant mass. He also assumes that the rate of radiation

is proportional to the fourth power of the average temperature of

. » Nature; vol. Ix, p. 247, 1899."

XVIII.] REPLIES TO SPECIAL OBJECTIONS. 151

this layer. He is dealing with temperatures which are so much
greater than the temperatures with which we work in the laboratory,
that such assumptions must be regarded as quite arbitrary.

" Mr. Homer Lane, in his classical paper on the theoretical tem-
perature of the sun,* makes the assumption that Dulong and Petit 's
law of radiation is true for solar radiation, and he uses it to calculate
the temperature of the radiating layer, which he finds to be 28,000° F.
That is, he uses an empirical law, obeyed possibly at laboratory tem-
peratures in radiation from hot solids, to express the radiation at
enormous temperatures from a hot layer of gas which has layers of
gas of all sorts of temperatures above and below it.

"It seems to me that we know too little about the phenomenon of
radiation from layers of gas with denser and hotter layers below and
rarer and colder layers above to allow of any weight being placed upon
these assumptions of Bitter or Homer Lane. In a star we have layers
of fluid at all sorts of temperature and density. We have no labora-
tory knowledge of radiation that is applicable. We know very little
about any star except our own sun. *.**•,-** Assumptions like
those of Homer Lane and Ritter may lead to results which are
altogether wrong."

Finally, I may refer to two more objections from another quarter,
the first relates to the connection which I have insisted upon between
the length of the continuous spectrum and the temperature of the light
source, and I have stated that this is based upon KirchhofF s law. To
this it is objected that rays far up in the ultra-violet can be emitted
from bodies not at a high temperature. The inference is that the stars
with the longest spectra may be cold. But they are connected with
the sun by an unbroken chain of sequences in the phenomena. Then
is the sun also cold 1

Again, it is urged that the phenomena of the gaseous stars instead
of being due to high temperature, are caused by phosphorescence.
Where then are Crookes's phosphorescent spectra 1 If this objection
implies that hydrogen can be made to phosphoresce so as to give us
Pickering's spectrum, the objector should have made the experiment
before he committed himself to such an objection..

* American Journal of Science anl Arts, 2nd series, vol. i, p. 57, 1870.

152

BOOK Y.— INORGANIC EVOLUTION.
CHAP. XIX. — WHAT EVOLUTION MEANS : ORGANIC EVOLUTION.

IN the previous chapters I have endeavoured to correlate all the facts
which have been obtained during the last, let us say, thirty years, in
relation to the sun, with more recent facts that have been gathered
with regard to the stars. In this we were, by hypothesis, watching
the effects of dissociation as the temperature rose higher and higher ;
we have found that the dissociation hypothesis, the view, namely, that
at high temperatures the chemical units with which we work at low
temperatures are broken up into smaller masses, explains the spectral
phenomena observed not only in our laboratories but in the sun and
stars.

I have also shown that in the opinion of many investigators suck a
dissociation is necessary to explain the phenomena observed in physi-
cal inquiries other than those which directly concern us here.

In these concluding chapters I propose to change the point of view,
to consider the phenomena no longer from the point of view of dis-
sociation but from that of evolution.

What is evolution 1 To answer this question I can refer to
another line of work in which the word is frequently used and
thoroughly understood. It is important that I should do this for
another reason, which will be gathered later. That line of work has
to do, not with inanimate forms, like the chemical elements and the
stars, but with living things, with so-called organisms. Most of my
readers know that what we now recognise as one of the greatest
triumphs of the century just ending was the determination of the truth
of a so-called " organic evolution " in which we have, I suppose, the
most profound revolution in modern thought which the world has
seen.

That evolution tells us that each kind of plant and animal was not
specially created, but that successive changes of form were brought
about by natural causes, and that the march of these forms was from
the more simple to the more complex. Organic evolution, in fact,
may be defined as the production of new organic forms from others
more or less unlike themselves ; so that all the present plants and

<3HAP. XIX.] WHAT EVOLUTION MEANS : ORGANIC EVOLUTION. 153

Animals are the descendants, through a long series of modifications or
transformations, or both, of a limited number of an ancient simpler
type. We must not suppose that this change has gone on as if things
were simply mounting a ladder ; the truth seems to be that we have
to deal with a sort of tree with a common root and two main trunks
representing animal and vegetable life ; each of these is divided into a
few main branches, these into a multitude of branchlets, and these
into smaller groups of twigs.

This new view represents to us the evolution of the sum of living
beings; shows that all kinds of animals and plants have come into
existence by the growth and modification of primordial germs. Now I
want just to say that this is no new idea, it is the demonstration which
is new to us in our present century and generation ; we have really to
go back to the seventeenth century, if indeed we must not go as far back
as Aristotle, for the first germs of it ; but with regard to the history,
however, I have no time to deal with it. There are two or three
points, however, to be considered in regard to this evolution. The
individual organic forms need not continuously advance, all that is
required is that there shall be a general advance — an advance like that
of our modern civilisation — while some individual tribes or nations, as
we know stand still, or become even degenerate. With this reserva-
tion, the first forms were the simplest. It may be that as yet we know
really very little of the dawn of geological history; that the fossili-
ferous rocks are nowhere near the real base. This conclusion has been
derived by Professor Poulton* from the complexity of the forms met
with in them ; still we find that we have not to deal with such a vast
promiscuous association of plants and animals of lowest and highest
organisation as we know to-day; we deal relatively only with the
simplest. The story both with regard to plants and animals is alike
in this respect.

Let me deal with the plants first. The first were aquatic — that is
to say, they lived in and on the waters. So far as we know, the first
plant life was akin to that of the algae, which include our modern sea-
weed, moss-like plants followed them, and then ferns, and it is only
very much later that the forms we know as seed plants with gaily
coloured flowers living on the land made their appearance. The
general trend of change amongst the plants has been in the direction of
a land vegetation as opposed to one merely in or on the surface of the
waters, and some present seaweeds exhibit the initial simplicity of
plant-structure which characterised the beginning of vegetable life,
while the seed plants I have mentioned are of comparatively late de-

* Presidential Address, Section D, British Association Meeting at Liverpool,
1896.

154 INORGANIC EVOLUTION. [CHAP.

velopment ; but we still have our seaweed ; so that with all the change
in some directions, some forms like the earlier survive,.

After this explanation, relating to work in an apparently different
direction, there should be no difficulty, in understanding the meaning I
attach to the word " evolution " so far as the history of plant change
is concerned, in relation to the chemical elements ; but we are not
limited to plant life. The same conceptions apply to animal life, and
it is important for my subject that I should refer to that also. What
do we find there ? We are brought face to face with the same pro-
gression from simple to complex forms. This is best studied by a
reference to the geological record.

Stratigraphical geology is neither more nor less than the anatomy
of the earth,* arid . the history of the succession of the formations is
the history of a succession of such anatomies ; or corresponds with
development as distinct from generation. In Stratigraphical geology,
as can be gathered from any book on the subject, we find the names
of certain beds which contain certain different forms of animal and
vegetable life. We begin with the Laurentian and Algonkian and
then pass to the Cambrian, then to the Ordovician, the Silurian and
Devonian, and so on through a long list of beds and geological
strata until we come eventually to the Eecent, that is to say,
the condition of things which is going on nowadays on the surface
of the earth. And if we prefer to map those many different
beds into more generic groupings, we begin with the Primary or
Palaeozoic, we pass on to the Secondary or Mesozoic, and then we
finally reach the Tertiary or Cainozoic. The deposition of these beds
and of the animal life which has been going on continuously on the
surface while those beds have been deposited, gives us the various
changes and developments which have taken place with regard to
animal forms.

It is worth while to go a little more into details and to indicate
the changes in these forms which have taken place, in the most general
way. Beginning with the Lower Cambrian, we find that the animal
forms were represented by Irrvertebrata such as Sponges, Corals,
Echinoderms, Brachiopods, Mollusca, Crustacea with many early Trilo-
bites; not to mention true Fucoids and other lowly plant-remains.
When we come to the Silurian, we find a large accession of the above
forms, especially of Corals, Crinoids, and Giant Crustaceans (such as
Pteryyotus) and armoured animals (Ostracodermi) without a lower jaw,
or paired fins ; the beginnings of Vertebrate life, ,not yet fully evolved,
and one lowly organised group of armoured fishes named Cyatluispis
(without bone-cells in their shelly-shield). Here, too, we meet with
* Huxley, Q.J.G.S., vol. xxv, p. 43.

XIX.] WHAT EVOLUTION MEANS : ORGANIC EVOLUTION. 155

the first air-breathers ; the wing of a Cockroach, and several entire and
undoubted Scorpions ! Thus in addition we get vertebrates as
opposed to invertebrates, and the -first traces of the fishes. In the
advance to the Devonian the fishes (associated with giant Crustacea)
predominate ; it has been called the age of fishes. In the next series,
the Carboniferous, we find the first certain traces of amphibians, of
which the early existence is like that of a fish : a state of things illus-
trated by the frog, which the majority of us in our early days have,
I am sure, studied as a tadpole in its early stages ; and some of these
amphibians still retain fish-like characters. It is not until we arrive
at the Permian that the true reptiles are met with, but in the next
great series, the Triassic, we meet with a remarkable evolutionary
group of Keptiles, the Theriodontia, or beast-toothed animals, because
(unique among reptiles) they possess a dentition like a dog or a lion,
with incisors, canines and cheek-teeth ; the precursors, doubtless, of the
succeeding mammalian type. We pass easily thus from the reptiles
to mammals which are related to them ; for instance, the ornitho-
rhyrichus and the echidna are both Australian mammals which bring
forth their young within the egg as do the reptiles. After that we
begin to deal with birds. The early birds were strikingly reptilian in
some of their characters ; and the pterodactyle, remains of which exist
in many museums, was really a winged reptile and not a bird. From
that we gather that mammals and birds are variants of reptiles. When
we progress from the Jurassic to the Recent, we find man making his
appearance as a direct descendant of all those early forms.

When we come to study the life-history of the various forms
brought before us by the geological beds, we find it to vary consider-
ably, a fact indicated by the presence or absence of the different
genera in the various strata. We find that the trilobites, for instance,
only appear in the very early geological formations ; there is no trace of
them in the recent, but of the annelids and Brachiopods we note that they
are continuous from the earliest to the latest formations ; we still have
our worms. Again we learn that certain other organic forms made
their appearance very low down in the time scale, forms which were
not represented at all in the earlier Cambrian and Silurian, and that
some of these are continuous to the present day.

Let us take the story of the fishes. A great many fishes made
their appearance at the Devonian stage, there were few in the Silurian ;
some of these stopped there, whereas others have been continued from
the Devonian times to our own. Take, for instance, the Australian
mudfish Ceratodus ; to judge from the teeth this fish might well have
lived on unchanged from late Palaeozoic times until the present day !
We see there is a tremendous variation of possible life-range, so to

156 INORGANIC EVOLUTION. [CHAP. XIX.

speak, with regard to these different forms. In that way, then, the
geologist has been able to bring before us the continuity of life in
various forms, from the most ancient geological strata to the most
recent. The record may be incomplete, but is complete enough for my
purpose.

But that is not the only evidence of evolution to which I can refer.
The teachings of embryology confirm the argument based upon the
study of geology, and suggest that the life-history of the earth is
reproduced in the life-history of individuals. The processes of organic
growth or embryonic development present a remarkable uniformity
throughout the whole of the zoological series ; and although knowledge
is still limited, some authorities hold that there is the closest possible
connection between the development of the individual and the develop-
ment of the whole series of animal life. There are others, however,
who do not regard the argument derived from embryology as a very
convincing one. However this may be, if we study the embryos of the
tortoise, fowl, dog, and man, we find that there is a wonderful simi-
larity between them at a certain stage. At a further stage of develop
ment the similarity is still borne out. This does not mean that a
vertebrate animal during its development first of all becomes a tortoise,
and then the various animals which are represented by these embryos ;
it simply means that they are all related, inasmuch as there is con-
tinuity.

After these references to plants and animals it should be clear
what organic evolution really is, and therefore what evolution is
generally.

157

CHAP. XX.— THE STELLAR EVIDENCE EEGARDING INORGANIC

EVOLUTION,.

JUST as plants and animals compose the organic or living world, so do
the so-called chemical elements (either single or combined) compose the
inorganic or non-living world.

Formerly plants and animals and the chemical elements were all
considered to represent special creations — " manufactured articles " ;
we now know that plants and animals do not ; that they have been
continuously evolved from simpler forms.

What we have now to consider is whether the facts set out in the
preceding chapters do or do not indicate that we must accept the
chemical elements, like plants and animals, as products of evolution.

Taking plants and animals as we know them, the more we dive
into past times the more differences in form are noted, though the
temperature at which the vital processes were and are carried on have
certainly not been widely different.

Taking the chemical elements as we know them here, we find differ-
ences in composition continuously indicated as stars of successively higher
temperature are studied. It is obvious that this is a very important
point. In inorganic evolution we are dealing with a great running down
of temperature ; how tremendous no man can say. We know the tem-
perature of our earth, but we do not know, and we cannot define, the
temperatures of the hottest stars. So that how great the temperature
of the earth may onee have been, supposing it to be represented by the
present temperature of the hottest star, no man knows with certainty.

With regard to organic evolution, however, which has to do with
the plant world and the animal world, there can have been no such
running down of temperature at all. The temperature must have been
practically constant within a very few degrees.

The differences then depend upon time in organic, and upon tem-
perature in inorganic, nature.

It is for this reason that in the inorganic evolution which now
concerns us the chemical changes brought about by changes of tem-
perature must be our chief guide, and the earliest and simplest forms
must be sought in regions where the highest temperature is present.

The effect of high temperature in producing simplifications is known
to everybody. If we deal, for instance, with well known chemical

158 INORGANIC EVOLUTION. [CHAP.

compounds, say chloride of sodium, that is common salt, and oxide of
iron, that is iron-rust, we produce the simpler substances of which they
are composed by heat, and we further have no difficulty in recognising
the fact that chlorine and sodium in one case, and oxygen and iron in
the other, must have existed before their compounds, common salt
and iron-rust, could be formed or associated. Water is split into
hydrogen and oxygen at a high temperature, so that there is a tem-
perature above, which the two. gases would remain in contact but
uncombined; when the temperature falls water is produced. Disso-
ciation, therefore, in all its stages must reveal to us the forms the
coming together of which has produced the thing dissociated or broken
up by heat. If this be so, the final products of dissociation or breaking up
by heat must be the earliest chemical forms. Hence if the various stars
behave like the various geological strata in bringing before us a pro-
gression of new forms in an organised sequence, we must regard the
chemical substances which visibly exist in the hottest stars which, so
far as we know, bring us in presence of tsmperatures higher than any
we can command in our laboratories, as representing the earliest
evolutionary forms.

I have said if. Now do the stars from the hottest to the coldest
present us with a progression of new forms as the geological strata do
from the. oldest to the newest ?

The preceding pages enable us to answer this question fully. On
p. 47 I indicated how, in cosmical evolution, we deal with a continuity
of effects accompanied by considerable changes of temperature ; from
the gradual coming together of meteoritic swarms until eventually we
have a mass of matter cold and dark in space. The various stars
which represent the different changes have been got out and have, in
fact, been arranged along a so-called temperature curve. As we
ascend one branch of this curve the stars get gradualty hotter and
hotter till ultimately at the top we find the hottest stars that we know
of. Then on the descending branch are represented the cooling bodies,
and finally they come down in temperature until we reach that of a
dark world like the companion of Sirius, of our own moon, and the
planet in which we dwell.

Thanks to the recent work, we can now deal with all these bodies
in special relation to their chemistry. No doubt the record will be
made more complete as time goes on and other workers come into the
field ; but it is already complete enough for my present purpose, for
the story is one of changes of chemical forms from one end to the
other.

When the photography of stellar spectra work was begun our
knowledge was so incomplete that a continuous chain of chemical facts

XX.] STELLAR EVIDENCE REGARDING INORGANlO EVOLUTION. 159

was out of the question ; but, thanks to the recent advances, we can
deal with this inorganic evolution from a chemical stand-point, and
what we have now to do is to consider tlie result of this inquiry.

Chapters VI and VII give the evidence on which the statement
can now be firmly made, that in the hottest stars we are brought in
presence of a very small number of chemical elements. As we" come
down from the hottest stars to the cooler ones the number of spectral
lines increases, and with the number of lines the number of chemical
elements. I will only refer to the known substances— it looks as if at
present we have still many unknowns to battle with. In the hottest
stars of all, we deal with a form of hydrogen which we do not know
anything about here (but which we suppose to be due to the presence
of a very high temperature), hydrogen as we know it, the cleveite
gases-, and magnesium arid calcium in forms which are difficult to get
here ; we think we get them by using the highest temperatures avail-
able in our laboratories. In the stars of the next lower temperature
we find the existence of these substances continued in addition to the
introduction of oxygen, nitrogen, and carbon. In the next cooler
stars we find silicium added ; in the next we note the forms of iron,
titanium, copper, and manganese, which we can produce at the very
highest temperatures available in our laboratories ; and it is only when
we come to stars much cooler that we find the ordinary indications of
iron, calcium, and manganese and other metals. All these, therefore,
seem to be forms produced by the running down of temperature. As
certain new forms are introduced at each stage, so certain old forms
disappear.

The salient features of the organic record are thus exactly reproduced,
to such an extent indeed that the most convenient way to present the
results was to define the various star-stages by means of the chemical
forms which they reveal to us in exactly the same way as the geologists
have done in regard to organic forms ; so that we may treat these
stellar strata, so to speak, as the equivalent of the geological strata.

From the hottest to the coldest stars I have found ten groups so
-distinct from each other chemically that they require to be dealt with
separately as completely as do the Cambrian and the Silurian forma-
tions. Imitating the geologist still further, I have given names ending
in ian- to these groups or genera beginning with the hottest, that is the
oldest dealing with the running down of temperature : — These are
Argonian, Alnitamian, Achernian, Algolian, Markabian [a " break in
strata "], Sirian, Procyonian, Arcturian (solar), Piscian.

I have also defined the chemical nature of these stellar strata as
the geologist defines the nature of any of his various beds ; we can say,
for instance, that the Achernian stars contain chiefly h5Tdrogen,

160 INORGANIC EVOLUTION. [CHAP,

nitrogen, oxygen, and carbon, and to a certain less extent they con-
tain proto-magnesium, proto-calcium, silicium, and sodium,* and pos-
sibly chlorine and lithium ; so that at last, by means of this recent
development of spectrum analysis, we have been able really to do for
the various stars what the biologist, a good many years ago, did for
the geological strata.

It will be seen, then, that the answer to the question : " Do the
stars show a progression of chemical forms as the geological beds show
a progression of organic forms ? " is clear and precise. There is a.
progression.

We are justified, therefore, in considering the matter further from
the evolution point of view. There are several points which merit
detailed consideration.

Obviously we cannot expect to get much help by thinking along
several obvious lines, for the reason that in the stars we are dealing
with transcendental temperatures. For instance, we must not make
too much of the difference between gases and solids, because at high
temperatures all the chemical elements known to us as solids are just
as gaseous as the gases themselves ; that is to say, they exist as gases ;
at a high temperature, everything, of course, will put on the nature
of gas. Those substances with the lowest melting points, such as-
lithium and sodium, will, of course, under our present conditions put
on the gaseous condition very much more readily than other substances
like iron and platinum, but those are considerations which need not be
taken into account in relation to very high stellar temperatures ; of
course, there would be no solids at a temperature of 10,000° C., and
there will be no gases in space away from the stars if the temperature
of space be taken at absolute zero.

Then with regard to metals and non-metals. Here again we really
are not greatly helped by this distinction. The general conception of
a metal is that it is a solid, and that, therefore, a thing that is not a
solid is not a metal : but the chemical evidence for the metallic nature
of hydrogen has been enlarged upon by several very distinguished
chemists, and mercury is generally known as a liquid. With regard to-
non-metals, there are certainly very many. Carbon is supposed to be
a non-metal, and it is remarkable that, so far as the stellar evidence has
gone as yet carbon seems to be the only certain representative of that
group.

I must point out specially that the table of the chemical defini-
tions of the various stellar genera (given on pp. 70 and 71), which
contains nothing but hard facts, is perhaps, like the geological record,,

* Campbell, Astronomy and Astro-physics, 1894, Tol. xiii, p. 395.

XX.] STELLAR EVIDENCE REGARDING INORGANIC EVOLUTION. 161

more important on account of what it indicates as to the presence of
the chemical elements in the stars than it is for what it omits.

There are a great many reasons why some of the substances which
may exist in these stars should not make their appearance. I wish to
enlarge upon the fact that, seeing the very small range of our photo-
graphs of stellar spectra, seeing also that it doesjnot at all follow that
the crucial lines of the various chemical substances will reveal them-
selves in that particular part of the spectrum which we can photo-
graph, the negative evidence is of very much less importance than
the positive evidence. I think it is possible, for instance, that^we must
add lithium to the substances which we find in the table on pages 70
and 71, we must certainly include sodium and also aluminium, and
chlorine possibly, but about sulphur at present I have no certain
knowledge. At all events, we can with the greatest^confidence point
out the remarkable absence of substances of high atomic weight, and
the extraordinary thing that the metals magnesium, calcium, sodium
and silicium undoubtedly began their existence in the hottest stars long
before, apparently, there is any obvious trace of many of the other
metals which a chemist would certainly have been looking out for.

162

CHAP. XXL— THE SIMPLEST ELEMENTS APPEAR FIRST.

WITH regard to the substances which appear in the hottest stars,
the all important, the first point to make, is that the chemical forms
we see are amongst the simplest.

How can this be determined 1 In two ways. The chemist will
acknowledge that an element of low atomic weight is simpler, that is,
has less mass than an element of high atomic weight. If we rely upon
spectrum analysis we can say, when dealing with the question of
*' series," that the elements which most readily give complete series
are in all probability simpler than those which give none, and this is
still truer when we find that all the lines in the spectrum of a sub-
stance can be included in those rhythmical series, as happens in the
case of hydrogen and the cleveite gases. Judged then by these
standards it is certain that the first stage of inorganic evolution, if
there has been such an evolution, is certainly a stage of simplest forms
as in organic evolution, whatever view we take of the nature of the
" atom."

It is worth while to compare in detail the results obtained by this
newest form of spectrum analysis relating to "series," with the
earliest stellar forms, because it is evident that we are here in presence
of the beginning of a new method of study of the nature of the
so-called chemical elements.

We found that the hottest stars contained hydrogen, helium and
asterium. We have also found (Chap. X) that those substances
have the simplest series ; that is to say, one set of three. It is more
than probable, although it is not absolutely established, that the
lithium group of metals is also represented in stars of very high
temperature. There, again, we have the simple series of one set of
three. About sulphur we do not yet know positively, but it is
probable, I think, that sulphur may exist in the hot stars. There,
again, we get another simple set of three ; so that for three perfectly
certain constituents of the hottest stars, together with one present in
all probability and one doubtful, we are dealing with the simplest
series.

But now comes the remarkable fact that side by side with these
simple substances we get in the hottest stars magnesium, calcium and
silicium. Of the " series " conditions of the last we know nothing. Of
magnesium and calcium only subordinate series have been determined.

CHAP. XXI.] THE SIMPLEST ELEMENTS APPEAR FIRST. 163

We cannot suppose that the absence of the principal series means
a greater simplicity, because I have shown that only about half the
lines in the spectrum of each of these substances has yet been picked
up in the series, and if the series represent the vibrations of. a single
particle, of course the lines which are not represented in the series, by
theory must represent the vibrations of some other particles. So that
there we are face to face with the possibility of a greater complexity of
the particles which produce the series than of those which in the
stars give us the lines not in the series. These then are other simple
forms.

Coming further down in stellar temperatures we find oxygen ;
here we deal with six series instead of three, or two, as in the case of
magnesium and calcium ; and even then, as I have pointed out, we do
not deal with above half the lines of the gas as we can see them at a
higher temperature. This then, seems to suggest that in the hottest
stars there are very various stabilities of very various forms : in short,
there seems to be there as here distinctly the survival of the fittest ;
otherwise how can we account for the fact that certainly in the hottest
stars we get three metals, magnesium, calcium and silicium, before we
have indication of any other, and that where we have those metals and
bring our series touch-stone to them, we find that instead of being very
simple they are really very complex as they exist here. However this
may be, we are now assured that there is a much greater quantity of
some apparently more complex forms in the hotter stars than of the more
simple ones ; and that is a matter which the chemists, when they come
to inquire into these questions which we are now considering, will
certainly have to face. This suggests, too, another very interesting
question. A great many simple organic forms appear in the strati-
graphic series at a late period ; some of the simplest forms died out,
others remained. Now, it may be that some of the more simple forms
in inorganic evolution, as in organic evolution, really represent later
introductions ; but, however this may be, it is perfectly certain that
we have not an absolute parallel between the results of the spectro
scopic observations of series and the spectroscopic observations of
stars.

In all these changes we seem to be brought into presence of succes-
sive complications, due to reduction of temperature, but there is a
longer series of complications in some substances than in others.
Of the origin of proto-magnesium and proto-calcium the stars as yet
tell us nothing ; but it is difficult to believe that the earliest forms of
the other metals are not built up of some of the constituents of the
heat ranges represented by those between y Argus and a Crucis, and
that their other complications began later.

M 2

164 INORGANIC EVOLUTION. [CHAP.

The next point is that the astronomical record, studied from the
evolution point of view, is in other ways on all-fours with the geological
record in relation to increasing complexity. We note the same
changes of forms, sudden breaks in forms, disappearances of old,
accompanied by appearances of new, forms ; and with these we have
to associate, whether we consider the atomic weight point of view or
the series point of view, a growth of complexity.

Although in this chapter I have chiefly referred to the stellar
evidence, I must not neglect to point out that over a restricted range
of temperature solar evidence can be utilized as well. We have
brought the sun and the stars together into line in all matters relating
to the discussion of the effects of higher temperatures. The photo-
graphs taken during the recent solar eclipses show that when we deal
with the hottest part of the sun that we can get at, which is hotter
than that part of the sun which produces the well-known absorption
spectrum marked by the so-called Fraunhofer lines, we are not in an
unknown territory at all, but are brought face to face with similar
phenomena to those in the atmospheres of stars which are hotter than
our sun. The bright-line spectrum of the sun's chromosphere seen
during an eclipse shows us the effects produced by heat in the hottest
part of the sun that we can reach ; these we can compare with the dark
lines of a star which contains absorption lines very different from those
represented by the Fraunhofer lines, and we find that they correspond
almost line for line.

Such an inorganic evolution was suggested by me many years ago
now, to explain the few stellar facts with which we were then familiar.
I must point out, however, that we are now in a very much better
condition to consider this problem than we have ever been before,
because at the present moment we have tens of thousands, I might
almost say hundreds of thousands, of co-ordinated facts to go upon,
and it is not a little remarkable that now the gaps in our knowledge
have been filled up, we find ourselves in the presence of evidences of
an evolution which is really majestic in its simplicity.

It is proper that I should say that jus^ as the work of Darwin in
the nineteenth century was foreshadowed by seventeenth century sug-
gestions, so the stellar demonstration with which we are dealing has
been preceded by hypotheses distinctly in the same direction. The
first stage of chemistry was alchemy ; alchemy concerned itself with
transmutations, but it was found very early that the real function
of the later science of chemistry was to study simplifications, and, of
course, to do this to the utmost we want precisely those enormous
differences in temperature which it appears the stars alone place at our
disposal.

XXI.] THE SIMPLEST ELEMENTS APPEAR FIRST. 165

With regard to the general question of inorganic evolution, the
first idea was thrown out in the year 1815 by Prout, who, in conse-
quence of the low atomic weight of hydrogen, suggested that that
substance was really the primary element, and that all the others,
defined by their different atomic weights, were aggregations of hydro-
gen, the complexity of the aggregation being determined by the
atomic weight ; that is to say, the element with an atomic weight of
20 contained 20 hydrogen units ; with an atomic weight of 40 it
contained 40, and so on. The reply to that was that very minute
work showed that the chemical elements, when they were properly
purified and examined with the greatest care, did not give exactly
whole numbers representing their atomic weights. They were so and
so plus a decimal, which might be very near the zero point, or half-way
between, and that was supposed to be a crushing answer to Prout's
view. The next view, which included the same idea — that is to say, a
physical connection between these different things as opposed to the
view that they were manufactured articles, special creations, each
without any relation whatever to the other, was suggested by Dobe-
reiner in 1817, and the idea was expanded by Pettenkofer in 1850.
Both pointed out that there were groups of three elements, such as
lithium, sodium, and potassium, numerically connected; that is, their
atomic weights being 7, 23, and 39, the central atomic weight was
exactly the mean of the other two, 7 + 39 = 46, divided by 2, we
get 23. Another way, however, of showing that is that 7 + 16 =23,
and 23 + 16 = 39 ; the latter method suggests a possible addition of
something with an atomic weight of 16.

In 1862 de Chancourtois came to the conclusion that the relations
between the properties of the various chemical elements were really
simple geometrical relations. It is not till 1864 that we come to the
so-called " periodic law," which was first suggested by Newlands, and
elaborated by Mendeleef in 1869. According to this law, the chemical
and physical properties of the elements are periodic functions of theii
atomic weights. Lothar Meyer afterwards went into this matter, and
obtained some very interesting results from the point of view of
atomic volumes. He showed that if we plot the atomic volumes of
the different elements, arranged according to their atomic weights from
left to right, there is a certain periodicity in the apices of the curve
indicating the highest atomic volumes.

So far there was no reference to the action of temperature in rela-
tion to this, but in 1873 I suggested that we must have a fall of tem-
perature in stars, and that the greater complexity in the spectra of
certain stars was probably due to this fall of temperature. This idea
was ultimately utilised by Sir William Crookes in an interesting varia-

166 INORGANIC EVOLUTION. [CHAP.

tion of the periodic law, in which he assumes that temperature plays
a part in bringing about the changes in the characters of the ele-
ments. Brodie, in 1880, came to the conclusion that the elements
were certainly not elementary, because in what he called a " chemical
calculus " he had to assume that certain substances, supposed to be
elements, were really not so ; and he then threw out the very preg-
nant idea that possibly in some of the hotter stars some of these ele-
ments which he predicted might be found. Nine years afterwards,
Rydberg, one of the most industrious investigators of the question of
" series " to which I have referred, stated that most of the phenomena
of series could be explained by supposing that hydrogen was really the
initial element, and that the other substances were really compounds
of hydrogen ; so that he came back to Front's first view in 1815. All
these ideas imply a continuous action, and suggest that there was some
original stuff which was continuously formed into something more
complex as time went on. That is to say, that the existence of our
chemical elements as we know them does not depend upon their having
been separately manufactured, but that they are the result of the
working of a general law, as in the case of plants and animals.

It will be gathered from the above statement that the stellar facts
are entirely in harmony with the highest chemical thought, and indeed
establish the correctness of its major contention. We may be said to
pass from chemical speculation to a solid chain of facts, which doubt-
less will be strengthened and lengthened as time goes on. In all these
changes we seem to be in the pre&ence of a series of complications, the
possibility of which depends upon 'a reduction of temperature. There
may have been roughly, a series of doublings, or the greater complexi-
ties may also have been brought about by the union of different
substances. In either case, as temperature falls, we get a possibility
of combinations which was not present before ; so that more and more
complex forms are produced.

In discussing the idea of evolution, both organic and inorganic, we
are driven to the consideration of a first form, from which all subse-
quent ones are derived.

The method of inorganic evolution must depend upon the way in
which complications are brought about. Although in this chapter I
have dealt with the received chemical view, I shall show subsequently
that it is not the only one we have to consider.

' It is well to point out that the inquiries referred to in this book
are now not the only ones which suggest the evolution of inorganic
matter from some primordial element such as I suggested in 1873, to
explain the spectroscopic facts then available.

I have already referred to the work recently accomplished on the

XXI.] THE SIMPLEST ELEMENTS APPEAR FIRST. 167

perturbations of spectral lines. Mr. Preston, in discussing the bear-
ings of his results, thus writes : — *

" We have, I think, reasonable hope that the time is fast approach-
ing when intimate relations, if not identity, will be seen to exist
between forms of matter which have heretofore been considered as
quite distinct. Important spectroscopic information pointing in this
same direction has been gleaned through a long series of observations
by Sir Norman Lockyer, on the spectra of the fixed stars, and on the
different spectra yielded by the same substance at different tempera-
tures. These observations lend some support to the idea, so long
entertained merely as a speculation, that all the various kinds of
matter, all the various so-called chemical elements, may be built up in
some way of the same fundamental substance."

In the same way Professor J. J. Thomson, in his important investi-
gations of the cathode rays, after describing a new series of facts,
writes : — f

" The explanation which seems to me to account in the most simple
and straightforward manner for the facts is founded on a view of the
constitution of the chemical elements which has been favourably enter-
tained by many chemists : this view is that the atoms of the different
chemical elements are different aggregations of atoms of the same
kind. In the form in which this hypothesis was enunciated by Prout,
the atoms of the different elements were hydrogen atoms ; in this
precise form the hypothesis is not tenable, but if we substitute for
hydrogen some unknown primordial substance X, there is nothing
known which is inconsistent with this hypothesis, which is one which
has been recently supported by Sir Norman Lockyer, for reasons
derived from the study of the stellar spectra."

On these points we must now go more into details.

* Nature, vol. Ix, p. 180.
f Phil, Mag., 1897, p. 311.

168

CHAP. XXII. — THE EELATIONS OF THE ORGANIC AND INORGANIC

EVOLUTIONS.

IT may be of interest to briefly consider the processes of inorganic
evolution in relation to those of organic evolution. I have already
referred to the fundamental difference in the conditions ; we found
evidence of a running down of temperature which no one can define
in the case of the stars ; in the case of the organic evolution going on
at the present time, we cannot be very much removed from the tem-
perature conditions of the Cambrian formations. That is a point
which I have made before, and it is important to insist upon it. Clearly
there cannot have been any very great change of temperature during
the whole cycle of organic life. Previous to it we have found com-
plexity brought about, possibly by doublings, and certainly by com-
binations, the result being, as I have already mentioned, more com-
plex forms. Of course, at the dawn of organic life on the surface of
the earth there may have been residua of the earlier chemical forms;
that is to say, not all the elements which we found in the hottest stars
had combined to form the substances of which the earth was com-
posed. However this may have been, although the work of organic
evolution, unlike that of inorganic evolution, must have been done
under widely different temperature conditions, the result has been the
same ; it has since provided us with another succession of forms getting
more complex as time has gone on, and there is still a residuum of early
forms.

We are led, then, to the conclusion that life in its various forms
on this planet, now acknowledged to be the work of evolution,
was an appendix, as it were, to the work of inorganic evolution
carried on in a perfectly different way. Although the way was differ-
ent, still nature is so parsimonious in her methods — she never does
a thing in two ways that can be as well done in one — that I have
no doubt that when these matters come to be considered as they
are bound to be considered with the progress of our knowledge,
we shall find a great number of parallels ; but I am not concerned
with parallels now. I wish to refer to a chemical point of view
which I think of some importance in relation to what has gone
before ; it is a point which I wish to make depending upon the
existence of those elements which make their appearance in the hottest
stars.

€HAP. XXII.] ORGANIC AND INORGANIC EVOLUTIONS. 169

In inorganic forms, in those represented to us in the hottest stars
.and the stars of gradually lower temperature, \ve have forms pro-
duced by a method by which complication is brought about • what this
method may probably be, we shall consider later on. Now the more of
these complications the more the early forms must have disappeared,
unless we may take it that they may have been made occasionally
to reappear by the destruction of the later forms ; that is a point to
bear in mind. If the simpler forms must go on combining to provide
the more advanced forms, then if all the simpler forms are so used up,
the only chance of getting the simpler forms again is to destroy some-
thing which had been previously made ; and we can quite understand,
of course, that there were many conditions of this destruction possible
at the time when the crust of the earth was being formed. But how-
ever that may be, the gaseous elements, together with the non-gaseous
elements first formed, would be the chief chemical substances on the
surface and over it. Now the substances over the crust, of course,
would be the gases, hydrogen, oxygen, nitrogen, and dealing with the
stellar evidence we may suggest carbon combined with them ; that is
to say, hydrocarbons, carbonic acid, and so on. On the surface, whether
the surface be one of land or water, we should expect, in addition to the
low melting point metals lithium and sodium, those three metals which
we know existed in the hottest stars long before the rest, magnesium,
calcium, and silicium. Lithium probably and sodium certainly exist in
some of the relatively hot stars ; the evidence also suggests sulphur,
and this is rendered more probable because of the simplicity of its
spectrum-series. Now these are very remarkable associations, and
seem far away from ordinary chemical considerations. Is it a mere
coincidence that they are the important substances in sea water 1

Constituents of Sea-water.

Chloride of sodium 77'75

„ magnesium 10'87

Sulphate of „ 473

„ lime ... ... ... 3-60

„ potash :.. 2-46

Bromide of magnesium 0*21

Carbonate of lime .«.. O34

The most easily thinkable organic evolution under these circum-
stances would be that of organisms built up of these chemical forms,
chiefly because they would represent the more mobile or the more
plastic materials ; we should not expect organic evolution to have
begun in iron, but rather in something the most mobile and the most

170 INORGANIC EVOLUTION. [CHAP.

plastic at the time. The available matter then for this evolution
would be those gases plus those metals and those non-metals to which
I have referred. Now, supposing such an evolution, if the forms so
composed were to be multiplied indefinitely, the available material
would be used up and organic evolution would be brought just as
certainly to a dead-lock as the inorganic evolution was brought to
a dead-lock when there was no possibility of any considerable reduc-
tion of temperature. We should expect a tendency to growth
among the organic molecules, I dare not call it an inherited tendency,
but I feel almost inclined to do so, having the growth of crystals
in mind. If when these new organic forms had been produced, the
results instead of being stable were emphatically unstable, and still
better if a dissolution or the destruction of parts or wholes could be
induced, progress would always continue to be possible, and indeed
it might be accelerated.*

The new organic molecules would ultimately not have the first
user of the chemical forms left available by the inorganic evolution,
but they would have the user of the gases and other substances pro-
duced by the dissolution of their predecessors. They would be
shoddy chemical forms, it is true, but shoddy forms would be better
than none. Under these circumstances and in this way, the organic
kingdom could go on ; in other words, the dissolution of parts or wholes
of the new organisms would not merely be an advantage to the race, but
might even be an essential condition for its continuance.

It therefore looks very much as if we can really go back as far as
these very early stages of life on our planet to apply those lines of
Tennyson : —

" So careful of the type she seems,
So careless of the single life."

* My friend and colleague, Professor Howes, has called my attention in this
connection to Professor Weismann's views (Welsmann on Heredity, vol. i, p. 112),
who seems to have arrived at somewhat similar conclusions though by a vastly
different road. He says, in his Essay on Life and Death, " In my opinion life
became limited in its duration not because it was contrary to its very nature to be
unlimited, but because an unlimited persistence of the individual would be a
luxury without a purpose."

The general view I have put forward, however, suggests that perhaps it was not
so much a question of luxury for the living as one of necessity in order that others
might live ; it was a case of morsjanua mtae.

The whole question turns upon the presence or absence, in all regions, of an
excess of the early chemical forms ready to be Used up in all necessary proportions.
Hence it may turn out that the difficulty was much greater for lard- than for sea-
forms, that is, that dissolution of parts or wholes of land-forms proceeded with
greater rapidity. It is a question of the possibility of continuous assimilation (see
Dantec, La Sexualite, p. 11), and the word "parks" which I have used refers to
the somatic cells, and not to the " immortal " part of living organisms.

XXII.] ORGANIC AND INORGANIC EVOLUTIONS. 171

We have arrived, then, at a condition in which the same material
may be worked up over and over again. In this way ultimately higher
forms might be produced. Now, if to this dissolution, as a means of
giving us new material, we add reproduction, then we can go a stage
very much further. If we take bi-partition, which was the first
method of multiplication, as we know, both in the vegetable and
animal world, and then obtain a multiplication of forms by halving
instead of the inorganic multiplication of forms by complicating, then
we can have a very much increased rate of advance.

These, then, roughly, are the ideas touching organic evolution
which are suggested by the stellar evidence as to inorganic evolution,
and the collocation of the simplest forms noted in the hottest stars.

Let us turn finally to the facts. Biologists are very much more
happy than astronomers and chemists, because they can see their units.
A chemist professes to believe in nothing which he does not get in a
bottle, although I have never yet seen the chemist who was ever happy
enough to bottle an atom or a molecule as such ; but the superstition
still remains with them, and they profess to believe in nothing that
they cannot see. Now, the organic cell, the unit of the biologist, is
itself a congeries of subordinate entities, as a molecule is made up of
its elementary atoms, manifesting the properties common to living
matter in all its forms.

The characteristic general feature of the vegetable activity of
plant forms is their feeding upon gases and liquids, including sea-
water. The progress of research greatly strengthens the view that
there was a common life plasm, out of which both the vegetable and
the animal kingdoms have developed. Be that as it may, it is found
that the vegetable grows upon these chemical forms to which I have
referred, and the animal feeds either upon the plant or upon other
animals which have in their turn fed upon plants ; so that there we
get the real chemical structure of the protoplasm, of the real life unit,
in our organic evolution.

Here another question arises. Is there any chemical relation
between the chemical composition of the organic cell and the reversing
layers of the hottest stars— the reversing layer being that part of a
star's anatomy by which we define the different genera 1

When we study the chemical composition of this cell we find it
consists of one or more forms of a complex compound of carbon,
hydrogen, oxygen, nitrogen, with water, called protein ; and proto-
plasm, the common basis of vegetable and animal life, is thus com-
posed. This substance is liable to waste and disintegration by oxidation,
and there may be a concomitant reintegration of it by the assimilation
of new matter.

172 INORGANIC EVOLUTION. [CHAP.

The marvellous molecular complexity of the so-called simple cell
may be gathered from the following formulae for haemoglobin : —

Man ... ... CoooH96oNi54FeiS3Oiv9.

Horse C-^Hnso^wFeiSaOo^-*

Various different percentage compositions have been given of this
protoplasm, but I need not do more than refer to them. It is more
important to consider the other chemical substances which go to form
it, for there are others besides which it is of interest to study from our
stellar point of view. I quote from Mr. Sheridan Lea.f

" Proteids ordinarily leave on ignition a variable quantity of ash.
In the case of egg-albumin the principal constituents of the ash are
chlorides of sodium and potassium, the latter exceeding the former in
amount. The remainder consists of sodium and potassium, in combina-
tion with phosphoric, sulphuric, and carbonic acids, and very small
quantities of calcium, magnesium, and iron, in union with the same
acids. There may be also a trace of silica"

Have we here more coincidences 1 or is it that the more one inquires
into the chemistry of these things the more we are brought back to
our stellar point of view, and to the fact that, taking the simplicity of
chemical form as determined by the appearance of these different
chemical substances in the hottest stars as opposed to the cooler ones,
and in relation to the "series" of spectra which they produce, we
come to the conclusion that the first organic life was an interaction
somehow or other between the undoubted earliest chemical forms ?
Not only have we hydrogen, oxygen, and nitrogen among the gases
common to the organic cell and the hottest stars, but those substances
in addition which I have indicated by italics.

Is it possible that we have here a quite new bond between man and
the stars 1

There is still another point regarding this question of the relation
of the two evolutions, inorganic and organic. I refer to the place of
organic evolution in regard to inorganic evolution in the scale of time.
I do not wish to call too much attention to this diagram, because
it is entirely hypothetical ; but it is constructed on the simplest prin-
ciples, so that it shall go as little wrong as may be. I begin by
drawing a line at the bottom, to represent the zero of temperature ;
certain temperature values are indicated on the left-hand side of the
diagram. Then we have the assumption that a star loses an equal
amount of heat in an equal period of time. In that way, then,
at the bottom we have relative times, at the side we have tempera-

* Verivorn, p. 104.

f The Chemical Bases of the Animal Body, p. 5.

XXII.] ORGANIC AND INORGANIC EVOLUTIONS. 173

tures, in centigrade degrees. Water freezes at a certain tempera-
ture above absolute zero, and boils at a certain other point ; these are
marked on our temperature scale. Then we have to remember that
about half way between the boiling point and the freezing "point, all

30OOO*

ARCONIAN

AL.NSTAM5AM

25OOO*

ACHERNIAM

ALCOLIAN

2OOOO*

MARKABIASO

SIRIAN

PROCYONIAN

10000°

ARCTURIAN

SOOO*

PISCIAN

OSM.ftlRID.MELT
IRON MELTS

ORC.EVOL

WATER OOILS
ICE MELTS I

Fig. 44. — Diagram showing that organic evolution occupies only a point in the
line representing the time and ternparature range required by inor-
ganic evolution.

the organic life with which we are familiar on this planet, from the
geological evidence and our own experienca, must have gone on at a
temperature of somewhere about, let us say, from 50° to 40° C. There,
then, we get the limit of organic life in relation to the possible

174 INORGANIC EVOLUTION. [CHAP. XXII.

inorganic life represented by the various chemical changes in the
stars.

We know from laboratory results that the stars of lowest tem-
perature are about the same temperature as that of the electric
arc, which is about 3,500° C., and so we put the Piscian stars there.
It has also been stated by Mr. Wilson lately, that the temperature of
the sun, measured by several physical methods, is something between
8,000° and 9,000° C., so that we put there the Arcturian stars. Of
course we have no means of determining the temperatures of the
hotter stars, so 1 have ventured to make a very modest supposition
that possibly we get about half the difference of temperature between
those stars as we have found between the Piscian and the Arcturian
stars from experiments on the earth. That will give us, roughly, some-
thing like 5,000° C. We find, then, that if we assume equal increments
of temperature for each of the different genera of stars that I have
brought together in Chapter VII, we get a temperature at the top of
the diagram of something like 28,000° C. All we have to do, then, is
to draw a diagonal line on which to mark the various temperatures
considered. On this the organic evolution, which represents every-
thing which has taken place with regard to living forms on the surface
of our planet from the pre-Laurentian times to our own, is represented
by a small dot.

It looks, therefore, very much as if these recent results of spectrum
analysis, may probably be of greater value in the future, because they
deal with a multitude of changes and a period of time compared
with which all the changes discussed by the geologists are almost
invisible on a diagram of this size. Not only shall we have probably
some help in determining this scale, but I think that, as I have already
indicated, the wonderful similarity between the substances contained
in the organic cell and those which would most likely be free when
the greatest amount of chemical combination had taken place on
the surface of the cooling world, will throw some light on the basis
of organic evolution itself.

In this way, then, we have really been only continuing a train of
thought, which has to do with Man's Place in Nature, in relation
to the Sun's Place in Nature ; and finding fresh grounds for thinking
that the more different branches of science are studied and allowed to
react on each other, the more the oneness of Nature impresses itself
upon the mind.

175

CH^P. XXIII. — INORGANIC EVOLUTION FROM A CHEMICAL
STANDPOINT.

IN the study of the -facts of inorganic evolution presented to us by
stellar spectra, there is one point of paramount importance to be in-
quired into. In the problems of inorganic evolution which we have
now to face, it is sufficiently obvious that we have to deal with a con-
tinuously increasing complexity of forms, precisely as in organic
evolution the biologist has had to deal, and has dealt successfully with,
a like increase of complexity of organic forms.

So far the processes by which complexity has been brought about
have only been referred to generally ; it is time now to endeavour to
gain a more .detailed insight into the methods by which inorganic com-
plexity has been arrived at. I will discuss this question first in rela-
tion to chemical theory.

If we ask the question How has complexity been brought about in
the case of .known chemical compound bodies 1 an easy answer is given
by analysis. -Chloride of sodium, for instance, is thus found to be
formed by the combination of chlorine and sodium. But when we wish
to deal with the formation of the so-called " elements " themselves, no
such easy solution of the question is open to us.

If in order to investigate the problem we take the analogy furnished
by compound bodies as our guide, we should say that the molecules of
the elements themselves were produced by the combination of unlike
forms.

But as a matter of fact, this method of producing complexity is
not the only one known to chemists. There are bodies of the same per-
centage composition which differ in molecular weight; the methane
series of hydrocarbons is a case in point ; the higher molecular weights,
or greater complexes, are produced by additions of the unit CH2, so
that these higher complexes are produced by the combination of
similar lower complexes. This process is termed polymerisation.

We are then familiar with two methods of increasing complexity,
which we may represent by a + a (polymerisation) and x + y (combi-
nation), producing a form A.

This, then, is the problem from the purely chemical side. On which
of these methods have the elements themselves been formed, now that
we are justified in considering them as compound bodies ? I suppose

176 INORGANIC EVOLUTION. [CHAP.

that chemists when hypothetically considering the possible dissociation
of the chemical elements would favour the view of depolymerisation ;
that is, the breaking up of a substance A into finer forms (a) weighed
by A/2 (or A/3), rather than a simplification of A into x and y.

The method of attacking this problem from the chemical point of
view in the first instance, must be a somewhat indirect one.

The Stars and the Periodic Law.

In Chap. XXI I referred to the important hypothesis put forward
by Newlands, Mendeleef and others in relation to the so-called " periodic
law," which law indicates that certain chemical characteristics of the
elements are related to their atomic weights.

It will be well now to study this question with a view of discussing
it more fully in the light of all the facts known to us, among which
the stellar evidence and that afforded by the study of series are, I
think, of especial importance ; since it may be said that we are now
absolutely justified in holding the view that of the lines which make
their appearance in the spectra of chemical substances when exposed
to relatively high temperatures, a varying proportion is produced b)/
the constituents of the substance, whether it be a compound like the
chloride of magnesium, to take an instance, or of magnesium itself.

Now the periodic law based upon atomic weights deals with each
" element " as it exists at a temperature at which the chemist can
handle it ; that is, if it be a question, say of magnesium, the chloride or
some other compound of the metal must have been broken up, and1
the chlorine entirely got rid of before the pure magnesium is there to
handle, and of this pure magnesium the atomic weight is found, andr
having also regard to its chemical characteristics, its position in the
periodic system determined.

But if the magnesium be itself compound, the position thus assigned
for the element is cei'tain not to tally with the stellar evidence if the
temperature of the star from which information relating to it is obtained
is high enough to continue the work of dissociation ; that is, to break
up magnesium itself into its constituents as certainly as the chloride of
magnesium was broken up in the laboratory in the first instance.

It is now known that dealing with this very substance magnesium,
high electric tension brings us in presence of a spectrum which con-
sists of at least two sets of lines, numerous ones seen also at the
temperature of the arc, and a very restricted number which make their
appearance in the spark.

If this be the work of dissociation— and, as I have shown, the
proofs are overwhelming — the " atomic weight " of the -particle, mole

XXIII.] FROM A CHEMICAL STANDPOINT. 177

cule or mass, call it what we will, which produces the restricted number
of Iine3— the enhanced lines — must be less than that of the magnesium
by the breaking up of which it is brought into a separate existence.

And now comes .the chief point in relation to the periodic law.
Seeing that the smaller masses which produce the enhanced lines have not
been yet isolated, their " atomic " weights and their chemical characteristics
have not been determined, and so of course their places in the periodic table
cannot be indicated as it at present exists.

My contention, therefore, is that some, at all events, of the ap-
parent discrepancies — for there are discrepancies — between the stellar
evidence and the " periodic " hypothesis arise from this cause.

The magnesium, and I will now add calcium, which the chemist
studies at relatively low temperatures have atomic weights of 24 and
40 respectively, and the stellar evidence would be in harmony with
the periodic law if magnesium (24) made its appearance after sodium
(23), and calcium (40) after chlorine (39), and generally each substance
should make its appearance after all other substances of lower atomic
weight than itself.

But, and again for the sake of simplicity I shall confine myself to
magnesium and calcium for the moment, in the stars we find lines in
the high temperature spectrum of magnesium and calcium appearing
before known lines in the spectrum of oxygen which has an atomic
weight of 16.

How are these results to be reconciled 1 I suggest that the expla-
nation is that the substances revealed by the enhanced lines of mag-
nesium and calcium and noted in the hottest stars have lower atomic
weights (smaller masses) than the oxygen of the periodic table.

Let us next, then, see what these atomic weights may possibly be.
Assuming A/2, the atomic weight of proto-magnesium would be 24/2 =•
12 ; of proto-calcium 40/2 = 20, supposing only one depolymerisation
has taken place. If we assume two depolymerisations, we get 6 and 10
as the " atomic " weights of the simpler forms of magnesium and
calcium which make their appearance in the hottest stars.

In this way we can explain the appearance of those finer forms of
magnesium and calcium before oxygen, with a small number of depoly-
merisations, and the stellar record of the order of atomic weights
would be the same :—

Hydrogen . . . . . . 1

Profco-calcium .. .. .. 10

Proto-magnesium . . . . 12

Oxygen . . . . . . . . 16

So much, then, for a possible reconciliation. The next point to be
considered is, is depolymerisation on such a small scale sufficient I

N

178 INORGANIC EVOLUTION.- [CHAP.

To do this we have to see the basis of the atomic weight of oxygen
16, and consider the series question in relation to oxygen. This
necessitates a reference back to Chapter X, in which I pointed out that
the simplest case presented in series phenomena is that placed before
us by sodium and other elements which run through all their known
spectral changes at a low temperature. Dealing with the line
spectrum stage we have three "series/' one principal and two sub-
ordinate (first and second). The former contains the orange line D,
constantly seen at all temperatures, the first subordinate the red line,
the second subordinate the green line, representatives of two series of
lines which are best seen both in the flame and arc.

The two subordinate series of sodium, like those of all other
elements so far examined, have the peculiarity that they end at nearly
the same wave-length, while the end of the principal series occurs at
a different, sometimes widely different, wave-length. This is a touch-
stone of the highest importance, as we shall see ; it points to a
solidarity of the two subordinate series, and to a difference between
them and the principal series.

Although the original idea was that all three series were produced
by the vibrations of the same molecule, observations of the sodium
phenomena alone are simply and sufficiently explained by supposing
that we have three different masses vibrating, and that two of them,
producing the subordinate series, can be broken up by heat, while that
producing the principal series cannot. The series represented by the
red and green lines seen best at the lower temperatures have been seen
alone, and it is a matter of common experience that the yellow line
representing the principal series is generally seen alone ; it is not
abolished at high temperature as the others are. Because the mass, the
vibrations of which give us the yellow lines, is produced by the break-
ing up of more complex forms at a low stage of heat, and it cannot
•be destroyed by the means at our command, it is the common representa-
tive of the element sodium. Because the masses, the vibrations of
which produce the two subordinate series represented by the red and
green lines, are easily destroyed by heat, they are more rarely seen ;
scarcely ever at high temperatures when the quantity is small, since,
as I pointed out years ago, " the more there is to dissociate, the more
time is required to run through the series, and the better the first
stages are seen."

This view is greatly strengthened by considering another substance

which, if we accept Pickering's and Kydberg's results, has like sodium,

three series, one principal and two subordinates in quite orthodox fashion.

I refer to hydrogen. The facts concerning which are given on p. 95.

Till a short time ago we only knew of one " series" of hydrogen,

XXIII. ] FROM A CHEMICAL STANDPOINT. 179

and on this ground Rydberg assumed it to represent the finest form of
matter known, regarding the other substances which give three normal
series as more complex. This idea is in harmony with the view ex-
pressed above.

If we accept the recent suggestions, we must regard hydrogen as
identical with sodium in its series conditions. But there is this tre-
mendous difference. In sodium we easily at low temperatures — the
Bimsen is sufficient — see all three series, while in the case of hydrogen
even the Spottiswoode coil can show us nothing more than one of the
subordinate series. At the same time, the other subordinate and the
principal series are visible in stars which we have many reasons for
believing to be hotter than the spark produced by the Spottiswoode
coil.

The argument for the existence of at least three different masses pro-
ducing the three different series, derived from the sodium observations,
is therefore greatly strengthened by what we now know of hydrogen.

I shall therefore assume it in what follows, which has reference to
more complicated phenomena.

Oxygen, instead of having three series like metals of low melting
point such as sodium, and the gas hydrogen, has six. These six have
been divided by Runge and Paschen into two normal sets of three,,
each set possessing one principal and two subordinate series.

There is evidently a new problem before us ; we require to add the
series of hydrogen to the series of sodium to get a " series " result
similar to that obtained from oxygen.

Before we go further it will be well to consider the possible order of
simplifications. Let us take the simplest case represented by sodium
and hydrogen in the first instance. The facts are shown in the follow-
ing table : —

High temperature.
Sodium. Hydrogen.

p Celestial

C Principal I and ter-
I Subordinate J restrial
Line stage <j [_ vapour.

f Terres-

^ Subordinate -< trial
Flutings . . . . [ vapour,

f Solid and

Continuous

" \ liquid.

fPrincipal "1 Celestial!

frmcipal 1 L'elest
Line stage J Subordinate / gas.

^Subordinate ")
Structure spectrum . }. Terrestrial

Continuous

• J

Low temperature.

We may now bring these results to bear upon oxygen. We learned
first from Egeroff that this gas at ordinary temperature and pressure
is so molecularly constituted that it produces a fluted absorption in
the red part of the spectrum. On account of the constancy of the

N 2

180 INORGANIC EVOLUTION. [CHAP.

results obtained by chemists we cannot be dealing with a mixture of
molecules, the fluting absorption therefore must be produced by mole-
cules of one complexity having an " atomic weight " of 16.

If we subject it to an induced current at low pressure (at which
the action of such a current is feeblest), it at once breaks up into two
normal sets of three • series, that is six series altogether ; it is almost
impossible to consider this state of things in the light of what happens
in the case of sodium and hydrogen without assuming on the ordinary
chemical view that the " molecule " with the fluted spectrum is broken
up into two, until finally we get —

High temperature.

SET A. SET B.

f Principal series. Principal series.

Line spectrum ..^Subordinate. Subordinate.

^Subordinate. Subordinate.

Fluted spectrum.
Low temperature.

But if we accept this, we give up depolymerisation, for the mole-
cules of the subordinate series of sets A and B thus produced cannot
be identical, because their spectra are not identical.

If we hold to depolymerisation we must arrange matters thus —

(^Principal.

I
Set B or A . . \ Subordinate.

I
^Subordinate.

fPrincipal,

I
Set A or B . . •{ Subordinate.

^Subordinate.

Fluted spectrum,
and we get six depolymerisations.

The number of lines measured by Runge and Paschen in the spec-
trum of oxygen at low temperature was 76 ; of these the six series
referred to contain 56, leaving 20 residual lines. Now if we employ a
strong induced current at atmospheric pressure, we practically extin-
guish these six series of lines and produce a new spectrum altogether,
containing a still greater number of lines : 114, according to Neovius.
Only one line is common to his table and that of Runge and Paschen.
About the series conditioning of these new lines we are at present
profoundly ignorant.

XXIII.] FROM A CHEMICAL STANDPOINT. 1*1

Let us take the simplest course in harmony with the principle of
continuity, and suppose that the great number of new lines is due to
the breaking up of the molecules of the upper principal series given in
the previous table into representatives of a still finer form, as hydrogen,
as we know it, is broken up into a finer form at the highest stellar
temperatures.

Have we, on the line of reasoning we are pursuing, any means of
estimating the number of finer forms which may be at work to produce
the 113 new lines 1

One possible way — a statistical way — seems open to us. Taking
the number of lines already recorded in the spectra roughly between
A. 7000 and A 2600 of the following substances, which give us three
series — lithium, sodium, potassium, helium, asterium, hydrogen — we
find that the number of lines in each series and the total numbers are
as follows : —

Maximum
number.

Minimum
number.

Average
number.

Principal series
First subordinate

.. lOAst
.. 37He

1 H

6Na

7
9

Second subordinate

.. 12 He

4 Li

8

Totals 39 11 24

This indicates that in oxygen we are slightly above the average
with — = 28 lines per set. If we take the facts for oxygen itself,

which give us 56 lines for two sets of three, the 113 lines will give
almost exactly four additional sets of three series, and therefore the
possibility of twelve more depolymerisations if this method of simpli-
fication is considered.

Of course we can halve the number of depolymerisations by assum-
ing that the fluting molecule instead of being depolymerised is broken
up into x and y, the bases of the two systems of series.

Now it is this last crop of new lines alone which is represented in
the hottest stars, and no one, I think, will now urge that some kind of
simplification which may include depolymerisation has not taken place
before they were brought into evidence.

Our base of 16 then vanishes, and with it the previously considered
possible atomic weights of the forms of magnesium and calcium which
precede the appearance of oxygen in the hottest stars. We must
therefore assume further depolymerisations in the case of these metals
beyond those considered in the first instance.

I now come to another point. How do the above considerations
bear upon hydrogen with its atomic weight of 1 ? Of this hydrogen
we know nothing spectroscopically. There is evidence that it is broken

182

INORGANIC EVOLUTION

[CHAP.

up into something which gives the complicated structure spectrum
with hundreds of lines not yet sorted into series, again into the one
series seen in our laboratories and in the cooler stars, still again into
two other forms we cannot get here.

Let us apply the statistical method we employed in the case of
oxygen.

In the region included in these inquiries the number of hydrogen
lines in the three series referred to is 17. Hasselberg has measured
454 lines in the structure spectrum between XX 642 and 441. Now if
this spectrum is built up of series similar to those observed at the
highest temperatures, we must have more (seeing that Hasselberg's

work was limited) than — . = ± 27 series or 9 sets of 3 each. We deal

then altogether with 12 depolymerisations.

But to be on the safe side, let us assume 6 on the ground that the
lines in the series may be more numerous, and that some of Hassel-
berg's lines may be due to flutings. It will be clear that the masses
or " atomic weights " we arrive at must be very small. Here is the
story : —

Where existent. Series, &c. Mass.

[Celestial /PrinciPal O'0019

. <p \ Subordinate 0'0039

[ Terrestrial Subordinate 0'0078

r « f -D f Principal 0'0156

- ° . , 4 Subordinate 0'0312

[Subordinate 0'0625

* j Q . [Principal 0*125

I J^A, , 4 Subordinate 0'25

^-Terrestrial[ Subordinate 0'5

f Hydrogen weighed

" \ in the cold . . 1

Spectrum.
Line spectrum

Fluted spectrum
Continuous spectrum

Such a conclusion as this, and therefore the reasoning which has
led up to it, must stand or fall according as science knows anything of
such masses.

I shall show subsequently that, thanks to the investigations of
Prof. J. J. Thomson, science is beginning to know a great deal of such
masses, arid the result of this work may therefore favour the view that
polymerisation is a vera causa for molecular complexity, at all events in
the cases of elements of low atomic weight ; if we accept the ordinary
chemical view.

Let us then consider the case of those elements the atomic weight
of which is greater. In the first stages of evolution, in which we deal
with substances of relatively low atomic weight, the stellar evidence
supplies us with definite landmarks, and these are definite because the
spectra of the hottest stars are not overcrowded with lines. After we
haVe passed the gaseous and proto-metallic stages, however, we find

XXIII.] FROM A CHEMICAL STANDPOINT. 183

the spectra full of lines which we see at the temperature of the arc,
and metals of relatively high atomic weight and melting point are
involved ; the exact sequences are naturally more difficult to follow,
and therefore the method of evolution may escape us.

Kayser and Runge have shown that the melting point has a pro-
found influence on the " series " conditions. Those with the highest
melting points, such as barium and gold, present us with no series.
There is generally such a flood of lines that it has been so far impos-
sible to disentangle them ; we have the " structure spectrum " of
hydrogen repeated in these metals at arc temperatures in the so-called
" arc spectrum."

I have already said that I think most chemists would consider
that the formation of larger masses by polymerisation is more probable*
than by the coming together of dissimilar atoms ; but if we consider
chemical compounds, certainly the analogy is all in favour of the
latter view if the principle of continuity be taken into account, for we1
are ignorant of the point at which one evolutionary process resigns in
favour of another. The present separation of compound from simple
bodies is, indeed, simply a measure of our ignorance arising from the
feebleness of our laboratory resources in relation to the temperature
required to produce more and more simplifications.

I discussed the question in my Chemistry of the Sun in 1887,
and showed that the analogy of the completely studied hydrocarbon
series beginning with CH2 suggested1 a hypothetical elemental
sequence.

a b, separate.

a + b, combined.

a + (b + b), written by chemists ab.2.

«+(*2)c*2) » » «C

and so on.

In the concrete hydrocarbon series we have continuous additions
of CH2 to CH4 until we reach a molecule defined by C^H^, and as
the building up of this molecule can be traced without difficulty, so
we can imagine it simplified by successive shedding 's of its constituent
CH2 ; we pass from a simplification which we can bring about by
simple halving to one which provides us with relatively large and
small masses.

184

CHAP. XXIV. — INORGANIC EVOLUTION FROM A PHYSICAL STAND-
POINT.

THE next question which arises is whether there is any way open to
us of getting still more light on this matter beyond that furnished by
orthodox chemistry.

With the progress of science the idea of " atoms " has considerably
changed.

Formerly they were regarded as merely chemically different from
element to element ; the recent investigations have introduced a new
conception. It is now no longer chemically different matter merely,
but matter, whether chemically different or not, carrying an electric
charge. In the first work along this new line, physicists, in order to
grapple with the phenomena of electrolysis and solutions, imagined
sub-molecules or sub-atoms carrying an electric charge in an electrolyte
from the anode to the cathode ; this was called an ion (Gr. a goer).
This conception has been more recently used to explain those move-
ments of particles of matter which produce light, and therefore spectral
lines. The sub-particle, this ion, with its electric charge e and its
mass ra, is supposed to move in an elliptic orbit under the attraction
of a centre. At first the theory supposed the ions to be electrified par-
ticles, but a recent extension considers them to be complex dynamical
systems, the motions of which are registered by spectral phenomena.

It will be gathered from what I have already said relating to the
various questions connected with the study of " series " of spectral
lines how the idea of " complex dynamical systems " is also demanded
to explain the phenomena presented by them.

Thus I have shown it to be probable that the hydrogen atom which
the chemist weighs may be built up of hundreds of the things, call
them what you will, a few of which in the hottest stars produce the
vibrations which we take as demonstrating the existence of hydrogen
in the celestial spaces.

Both these lines of modern evidence tend to justify the view that
the different spectra are not produced by different material, but by
different conditionings of the same material.

These different conditionings may refer either to the electric charge
or to the mass of the ion, or of the molecule round which the ion cir-
culates. The units of matter present in the ion or in the central
molecule may vary in number, or their arrangement may vary.

CHAP. XXIV.] FROM A PHYSICAL STANDPOINT. 185

Imagine a series of substances " chemically " different, the intrinsic
difference of which, from A the simplest to Z the most complex, really
consists merely of their being built up of different numbers of units.
When Z is simplified by heat, its complex system of centre of force
and ion with their electric charges will undergo changes which we may
expect to result in the formation of less complex systems doubtless
built on a like pattern, and therefore capable of producing spectra ;
hence we are bound to see the spectra of some of the intermediate
forms which, when they are stable and go about in company, it may
well be that physicists have already recognised. These we may call
B or C, or R or S, or X or Y, as representatives of various com-
plexities.

The more complex the form experimented on and the higher the
temperature employed m the laboratory, the more spectral lines
indicating different chemical " elements " in intermediate stages may
we see.

I say in the laboratory, because in the stars the result will be dif-
ferent. There, in consequence of the long continued action of heat
and the shielding of the reversing layer from the effects of lower tem-
perature, we may only see at the highest temperature the spectra of
the forms A and near A. We now know what these are.

To take another case ; let us assume that the electric charges or
arrangement, as well as the number of the units of matter, may vary.
Under these conditions, when we dissociate Z, not all, but only some,
of possible intermediate forms may be expected to afford spectral
evidence. Say, to take an example, those in the vertical columns of
Mendeleef 's table ; and I am led to make this suggestion, because
Kayser has shown that in " series " the duplicity or triplicity of lines
is associated with the position of the elements producing them in these
columns. A concrete case would be afforded by contrasting the be-
haviour of sodium and caesium, representing relatively simple and
complex substances. We might observe the lines of sodium when
caesium is dissociated ; we should not expect to see the lines of caesium
when sodium is dissociated.

The two cases taken it is possible may illustrate the difference
between related and not related groups of " elements."

The apparently constant appearance of representative lines of the
spectrum of one substance of a group in that of the other members of
the same group may be thus explained, although it has generally been
attributed to the presence of impurities, as in the case of all common
long lines seen in spectra ; and this -in spite of the protest that if the
purest specimens known (I have worked on beads of Stas' silver which
had never been touched) were so impure, some of the decimals used to

186 INORGANIC EVOLUTION. [CHAP,

express their atomic weight might be well spared. But it is not a
question 'of apparent impurities only.

It is possible that some of the gases of lower atomic weight which1
exist in • the Hottest stars may be represented by A in opposition to
heavy metals represented by Z, the existence of which is -known in the
cooler stars only.

The giving off of gases from metals when high tension electricity
is employed is well known. This has been explained by assuming
them to be " furnace gases," that is gases "occluded" by the metals
during their reduction. But this does not seem to be a sufficient
explanation, for the same gases are given off by meteorites. We now
see why something like this may happen if there is any foundation for
the modern conception of the structure of the "atom"; and do not
these facts explain the chemistry of the hottest stars 1

It is too early yet to attempt to discuss the effects of the electric-
charge in this connection, but it must be pointed out that so soon as
the ions, however associated their units may be, which are supposed
always to have an electric charge upon them, are subjected to the
action of a voltaic or induced current, the spectral phenomena ob-
served when they are heated are liable to great changes in seme
cases, and especially when high atomic weights are in question;
Doubtless we have here a field of research which will ultimately
supply us with the most precious knowledge. I have already shown
that with the gases, such as hydrogen and oxygen, heat alone gives
rise to no spectral phenomena, while in the case of such metals as
sodium, heat is so effective in its dissociating power that the subse-
quent application of electricity produces no further change.

We have, in fact, to consider that the effects produced on different
substances under the same conditions may be different, and that the
stars carry us further than our laboratories; that is, there are staged
of spectral change within and beyond our experimental powers reveal-
ing a shedding of ions or some rearrangement of material at different
temperatures. Of course it is possible that the rearrangement of
material may take place in the central molecule itself ; the point to be
remembered is, that whatever may happen, whether in the central
molecule or the ion, a higher temperature will be associated with a
simplification of the total mechanism.

Dr. Preston's Researches.

Quite recently the study -of magnetic perturbations of spectral
lines has brought a fresh array of evidence on this question.

It has now been proved that spectral phenomena are different

XXIV.] FROM A PHYSICAL STANDPOINT. ISY

when the light source under examination is subjected to the "action of
a strong magnetic field which, among other things, causes a proces-
sional movement of the orbits of the ions to which I have' already
referred.

In order to consider the bearing of this, let us deal with the
spectrum of zinc which contains triplets. It has been shown that
denoting these in ascending order of refrangibility by AI, Bb Ci, A2,
B_>, Co, &c., the lines AI, A2, &c., show the same magnetic effect in
character, and have the same value of 0/w. The lines BI, B-2, BS, &c.,
and Ci, C-2, Cs, &c., form other series, and possess a common value for
the quantity e/m in each case.

Dr. Preston, one of the most successful workers in this new field;
states : —

" The value of e/m for the A series differs from that possessed hy
the B series, or the C series, and this leads us to infer that the atom
of zinc is built up of ions which differ from each other in the value of
the quantity e/m, that each of these different ions is effective in pro-
ducing a certain series of lines in the spectrum, of the metal."

But this is by no means all that is to be learned from Dr. Preston's
researches. He writes —

" When we examine the spectrum of cadmium or of magnesium—
that is, when we examine the spectra of other metals of the same
chemical group — we find that not only are the spectra homologous,
not only do the lines group themselves in similar groups, but we find
in addition that the corresponding lines of the different spectra are
similarly affected by the magnetic field. And further, not only is the
character of the magnetic effect the same for the corresponding lines
of the different metals of the same chemical group, but the actual
magnitude of the resolution, as measured by the quantity e/m, is the
same for the corresponding series of linos in the different spectra.
This is illustrated in the following table, and leads us to believe, or at
least to suspect, that the ion which produces the lines AI, A2, AS, &c.,
in the spectrum of zinc is the same as that which produces the corre-
sponding series AI, A2, AS, &c., in cadmium, and the same for the
corresponding sets in the other metals of this chemical group. Iri
other words, we are led to suspect that, not only is the atom a com-
plex composed of an association of different ions, but that the atoms of
those substances which lie in the same chemical group are perhaps
built up from the same kind of ions, or at least from ions which
possess the same e/m, and that the differences which exist in the
materials thus constituted arise more from the manner of association
of the ions in the atom than from differences in the fundamental cha-
racter of the ions which build up the atoms/'

188

INORGANIC EVOLUTION.

[CHAP.

Nonets

Magnetic effect.

or complex

Sextet*.

Triplets.

triplets.

Cadmium . . . . A =

5086

4800

4678

Zinc A =

4811

4722

4680

Magnesium . . . . A =

5184

5178

5167

Processional spin ..

e/m = 55

ejm = 87

ejm = 100

[This table shows the effect for the three lines which form the first natural
triplet in the spectrum of cadmium compared with the corresponding lines in the
spectrum of zinc and magnesium. It will be seen that the corresponding lines in
the different spectra suffer the same magnetic effect both in character and magni-
tude. Thus the corresponding lines 4800, 4722, 5173 are each resolved into
sextets, and the rate at which the ionic orbit is caused to precess is the same for
each (denoted by ejm = 87 in the table). Similarly for the other corresponding
lines.]

This is a result of the first order of importance. I previously dis-
cussed what might be expected to happ3n if the complex system
giving the spectrum of an element were broken up, and showed that if
less complex systems of the same pattern — that is, consisting of centre
of force and ion with its electric charge — were thus produced, these
systems would be just as capable of giving spectra as the one the
breaking up of which produced them. We should get new ions free to
move 'and vibrate, and new spectra which may reveal the constituents,
that is, the mariner in which the complex system breaks up. But
Dr. Preston goes further that this. He shows that the same ion
associated with different centres of force gives us lines at different
wave-lengths. That a certain ion which in the spectrum of mag-
nesium gives rise to b is also present in zinc and cadmium, though
there is no trace of b in their spectra.

Now, if the views held by those who have worked along any of
these lines be confirmed, we shall be compelled not only to give up
polymerisation as the only cause of greater complexity of the mole-
cules of the elements, but to acknowledge a great strengthening of the
view that all chemical atoms have a common basis, and build new
mental images on this basis. I now pass from the spectroscopic
evidence to work in a new field.

Professor J. J. Thomson's Researches.

I have before referred to the fact that science now has to consider
masses much smaller than the atom of hydrogen. This we owe not
only to a discussion of the phenomena of series, but also to some

XXIV.] FROM A PHYSICAL STANDPOINT. 189

recent researches of Professor J. J. Thomson, made in connection with
his work on the cathode rays.

Since the cathode rays produce -luminous effects their path can be
traced, hence it is known that they are deflected in a magnetic field.
This deflection depends upon the mass of each particle and the electric
charge it carries, that is, upon their ratio, m/e. This ratio Professor J.
J. Thomson finds to be about yj^th of the corresponding value for the
hydrogen ion in ordinary electrolysis.

At the same time it has been found by Professor J. J. Thompson
and Mr. Townsend that the electric charge e is the same for cathode
rays and a hydrogen ion. The m/e in fact may be regarded as inde-
pendent of the nature of the gas. Since then the m/e of the hydrogen
ion is 700 times greater than in the case of cathode particles, the m, the
'smallest mass whose existence Professor J. J. Thomson has glimpsed,
•can only be about TJ^th of the hydrogen ion.

Professor J. J. Thomson writes : — l

" The explanation which seems to me to account in the most simple
.and straightforward manner for the facts is fc*unded on a view of the
constitution of the chemical elements which has been favourably enter-
tained by many chemists ; this view is that the atoms of the different
chemical elements are different aggregations of atoms of the same

kind.

* * * * * *

" Thus on this view we have in the cathode rays matter in a new
•state, a state in which the subdivision of matter is carried very much
further than in the ordinary gaseous state : a state in which all matter
—that is, matter derived from different sources, such as hydrogen,
oxygen, &c. — is of one and the same kind, this matter being the sub-
stance from which all the chemical elements are built up.

* * * * * *

" The smallness of the value m/e is, I think, due to the largeness of
<? as well as the smallness of m. There seems to me to be some evi-
dence that the charges carried by the corpuscles in the atom are large
compared with those carried by the ions of an electrolyte."

Thus the whole question of dissociation has been advanced
because while on the chemical view we have to deal with intrinsically
different kinds of matter from element to element, on the view of Pro-
fessor J. J. Thomson m is a constant for every element, reminding one
•of Rydberg's general formula for series in which N0 is practically a
constant for every element, although Eydberg acknowledges slight
variations which may be due to errors of observation.

1 Phil. Mag., vol. xliv, p. 311, October, 1897.

190 . INORGANIC EVOLUTION. ['CHAP.

Professor J. J. Thomson is thus led to the following view of the
differences of construction of a simple " atom " and a compound
" molecule " : —

" In the molecule of HC1, for example, I picture the components of
the hydrogen atoms as held together by a great number of tubes of
electrostatic force ; the components of the chlorine atom are similarly
held together, while only one stray tube binds the hydrogen atom to
the chlorine atom."

Dr. Preston's results on the magnetic perturbation of lines, to
which I have already referred, leads him to the same general conclu-
sions as those arrived at by Professor J. J. Thomson in favour of the
view of dissociation. He says : — •

" It may be, indeed, that all ions are fundamentally the same, and
that differences in the value of e/m, or in the character of the vibra-
tions emitted by them, or in the spectral lines produced by them, may
really arise from the manner in which they are associated together in
building up the atom."

The Three JVays of Inorganic Evolution.

At the present time, then, we have before us three suggested ways
of inorganic evolution.

Taking the chemical view, this may depend on

(1} Polymerisation, or the combination of similar chemical mole-
cules; or

(2) The combination of dissimilar chemical molecules.
In the new physical view all this is changed into

(3) The gradual building up of physical complexes from similar
particles associated with the presence of electricity.

In this last conception we have the material world, up to the
highest complex, built up of the same matter under the same laws ; as
in spectrum analysis there is no special abrupt change between the
phenomena presented by the simple and compound bodies of the
chemist, so also in the new view there is no break in the order of
material evolution from end to end. I have already, on p. 167,
referred to the opinions expressed by Professor J. J. Thomson and
Dr. Preston, as to the manner in which the new work supports my
view expressed many years ago.

Certainly the new view seems competent to throw light on many
facts which lacked explanation on the old one, by whatever method of
evolution the higher complexes were assumed to be brought about,
because on the ionic theory we can imagine several first forms, so
that the question of descent comes later" with the introduction of more

XXIV.] FROM A PHYSICAL STANDPOINT. 191

complex systems. These various first forms bring about the possi-
bility of evolution along several parallel lines, as well as of the possi-
bility of an infinite number of intererossings. In this connection we
must not forget that the constituents of the reversing layer of Bellatrix
and of protoplasm are nearly identical, while the particular forms of
matter of which they are composed make so little show in the sun.

A consideration of the central congeries of material units and the
ion revolving round it, suggests that the ion may be the more con-
stant in its structure, and that it is to a large extent to the varying
mass and charge representing the centre of force that spectral changes
are due. It may be that the subordinate " series " indicate that very
small variations of complexity are possible, as well as greater ones.

The ions visible in the simple spectra of the hottest stars may be
those associated with the smallest centres of force. These are, so far as
we know at present, hydrogen, helium, asterium, oxygen and nitrogen
among the gases ; carbon and silicium, and calcium, magnesium, and
sodium among the metals in the forms we study by their spectra at
the highest temperatures we can employ in our laboratories.

As the stars cool larger aggregates of material units in the centres
of force round which these ions revolve become possible, and hence
the complexity of the spectrum of uranium and of the sun, repre-
senting a cool star, are both explained by the same process, the various
stages of which can be reproduced in the reverse direction by various
degrees of dissociation.

INDEX.

Absorption, 13, 14.

in space, 131.

phenomena, Jewell, 104.

solar lines, 15.

Acetylene, spectrum of, 20.
Achernian stars, 70.
Alchemy, 164.
Aldebarian stars, 70.
Algrolian stars, 70.
Alnitamian stars, 70.
Aluminium, series of, Cornu, 87.
Ames, Zeeman effect, 112.
Antarian stars, 70.
Arcturian stars, 70.
Arcturus, see Bootis a.
Argonian stars, 70.
Asterium, series, 95.
Atmospheres, stellar, 55.
Atomic weight and periodic law, 176.

series, 93.
Atoms, chemical, 18.

conception of, 185.

dissociation of, 73.

of oxygen, 96.

simple, 190.

vibration period of, 103.
Auriga, nova in, 138.
Aurig-se a, absorbing layer, 52.

Balmer's law, 86.
Bauschinger, 128.
Berthelot, on dissociation, 28.
Boisbaudran, de, 86, 117.
Bootis o, absorbing layer, 52, 150.
Brodie, chemical elements, 166.
Budg-e, Dr., 67.

C.

Calcium, behaviour of, 106.

dissociation, temperature of, 106.

enhanced lines, 37.

pressure shift, 106, 147.

proto-, behaviour of, 64.

series in, 97.
Calculus, chemical, 166.

Camera, prismatic, 43.
Campbell, bright line stars. 133.
Canis majoris a and a Cygni, compari-
son, 68.

and 7 Cygni, comparison, 68.
Capella, see Aurigse a.
Carbon, distribution in space, 142.

flutings of, 75, 83.

in stars, 61, 129.

stars and Milky Way, 129.
Chancourtois, 165.
Chromosphere and enhanced lines, 39.

and Fraunhofer spectrum, 43.

eclipse, 1871, 43.

eclipse, 1878, 43.

position of, 50.

spectrum, 39, 41, 42.
Classification of stars, 66, 70, 72, 129,

130.

Cleve, yttria, 117.
Cleveite gas, and series, 11, 83.

composition of, 96.

Clifford, magnetic perturbations of
lines, 109.

path of molecules, 18, 29.
Clusters, star, and Milky Way, 128.
Compounds, spectra of, 20.
Cornu, aluminium series, 87.

magnetism and wave-length, 112.

reversed lines, 101.

series, constants of, 87.
Corona, 1871, 43.

1878, 43.

1898,43.

variation of, 43.
Crookes, dissociation, 76.

fractionation, 116.

periodic law, 165.

victorium, 119.
Crucian stars, 70.
Crucis, £, 60.
Cygni a, absorbing layer, 52, 80.

and Sirius, comparison, 68.

spectrum of, 49.

7 and Procyon, comparison, 68.
Cyg-nian stars, 70.

D.

Depolymerisation, 176.
Deslandres, series, 87.

194

INDEX.

Deville and Troost, iodine, 74.
Dewar and Liveing, carbon flutings,
75.

series, law of, 86.
Dissociation and eTolution, 158.

and spectral series, 83.

and temperature, 36, 79.

Berthelot, 23.

calcium, temperature of, 105, 106.

celestial, 25.

Dumas, 27.

early idea of, 22, 73.

examples of, 158.

objections, 144.

of metals, 81.

Ritter, 149.

Sir William Crookes, 76.

Sir William Roberts-Austen, 70.
Distribution of carbon stars, 129.

stellar, 126.
Dbbereiner, 165.
Dumas, on dissociation, 27.
Duner, stellar classification, 129.

Earhart, Zeeman effect, 112.
Eclipse, 1871, 43.

1878, 43.
Eder, calcium, behaviour of, 106.

and Valenta, mercury spectrum,
76.

molecular combination, 74.
Eg-eroff, oxygen, 179.
Elements, chemical distribution in

space, 142.
and temperature, 165.

in hottest stars, 159.

initial, 165, 166.

series of, 87, 90.

spectra of, 18, 19.

without series, 93.
Embryology, 156.
Enhanced lines, 31.

and chromosphere, 39.

calcium, 37,

in a Cygni, 48.

in stars, 33.

in sun, 35.

iron, 34.

lithium, 32.

magnesium, 35.

temperature ranges, 57.
Evolution, 152.

and dissociation, 158.

organic and inorganic, 172, 190.

stellar evidence, 62.

F.

Faraday, magnetic perturbation of
lines, 109.

Fievez, magnetic perturbation of lines,

110.
Flutings, 10, 83.

carbon, 75, 83.

magnesium, 83.
Formulae for series, 87, 88.
Fractionation, 116.
Frankland and Kirchhoff's theory, 41 .
Fraunhofer lines, 15, 39, 42. 43.

Gr.

Gas, cleveite, composition of, 96.
G-ases, series in, 91.

spectra of, 30.

Gaug-es, star, Herschel, 126.
Genera, stellar, 70.
Geology, stratigraphical, 154.
Gould, Milky Way, 125.
Gravity, specific, of iodine, 74.
Groups, chemical, and series, 91.

H.

Haemoglobin, formula for, 172.
Hale, calcium, behaviour of, 105, 106.
Hasselberg-, hydrogen spectrum, ]82.
Helium and Fraunhofer lines, 39, 48.

distribution in space, 142.

series, number of, 95.
Hemsalech, artificial shifts, 107.
Herschel, Sir John, star-way, 126.
Herschel, Sir William, and nebulae, 143.

milky way, 125.

Hittorf and Pliicker, spectra of ele-
ments, 19.
Howes, Prof., 170.
Hug-gins, spectrum of Arcturus, 150.

spectra of nebula;, 121.
Humphreys and Mohler, pressure and

wave-length, 101, 105.
Humphreys, behaviour of strontium,

106.

Hydrocarbons, series of, 175, 183.
Hydrogren, behaviour of, in stars, 64.

complexity of, 95.

distribution in space, 142.

ion, 189.

primary element, 165.

proto-, 58, 81.

series of, 31, 59, 86, 87, 95, 98, 179,
182.

structure, spectrum of, 31.

I.

Iodine, specific gravity, 74.

Ions, electric charge, 110, 114, 185, 190.

hydrogen, 189.

theory of, 190.

Zeeman effect, 188.

INDEX.

195

Iron, behaviour in sun, 24.
distribution in epace, 142.
line of, 1474, 35.
spectra of, 32, 34.
Zeeman effect, 112.

J.

Jewell, absorption phenomena. 103,

104.

atoms, vibration of, 103.
dissociation hypothesis, 103, 147.
pressure and wave-length, 101.

Kapteyn, proper motion, 139.
Kayser and Runge, series, 87, 97, 183.
Kayser, hydrogen, new series, 59, 98.

molecules, complexity of, 100.
Kirchhoff's theory, 41.

Lane, Homer, temperature of sun, 151.
Larmor, Zeeman etfect, 110.
Law, periodic, 165, 176.
Layer, absorbing, Arcturus, 52.

Capella, 52.

sun, 52.

a-Cygni, 53, 80.

Lea, Sheridan, protoplasm, 172.
Lines, basic, 77.

enhanced, 57, 177.

long and short, 22.

proto-metallic, 57.

Liquids and metals, differences be-
tween, 160.

path of molecules in, 19.
Lithium, enhanced lines, 32.
Liveing and Dewar, carbon flutings,
75.

series, law of, 86.
Lorentz, Zeeman effect, 110.

Magnesium and periodic law, 176.

flutings of, 83.

in stars, 75.

proto-, behaviour of, 35, 64.

series in, 97.
Magnetism and spectral lines, 109, 112.

field of, ions, 110.

perturbations of lines, 109, 110.
Marignac, de, yttria, 117.
Markabian stars, 70.
Mascart, 12.

McClean, classification, stellar, 60, 67,
141.

oxygen in stars, 60.

star distribution and Milky Way,

131.

Melting-point and series, 93, 97.
Mendeleef, periodic law, 94, 165.
Mercury, spectrum of, 76.
Metals and liquids, differences be-
tween, 160.

dissociation of, 81.

proto-, 57.

without series, 97.
Meyer, atomic volumes, 165.
Milky Way, 124.

and nebulae, 127, 136.

and novse, 137.

and star clusters, 128.

distribution, 125, 1^9, 131, 132, 141.

Gould, 125.

split in, 124.
Mitscherlich, spectra of compounds,

20.
Mohler and Humphreys, pressure and

wave-length, 101, 105.
Molecule, yttria, 118.
Molecules, combination of, 74.

complexity of, 99, 190.

composition of, 171.

diatomic, 18.

dissociation of, 73.

in gases, 19.

masses of, 182.

path of, 18, 19, 29.
Monck, 139.

Motions, proper, and stellar spectra,
140.

proper, and temperature, 139.
Murray, Dr., stellar type names, 70.

N.

Nebulse and Milky Way, 127, 136, 143.

and stars, 132.

distribution of, 136.

origin of, 46.

planetary, 127, 128, 136, 143.

spectra of', 121.
Neovius, oxygen, 180.
Newlands, periodic law, 165.
Nitrogen, spectra of, 31, 61, 86.
Nova Aurigae, 138.

duplicity of, 138.
Novse and Milky Way, 137.

distribution of, 137.

genesis of, 137.

number of, 137.

spectra of, 138.

O.

Orion, stars in, temperature of, 80.
O 2

196

INDEX.

Orionis, f, 67.
Oxyg-en, atom of, 96.

complexity of, 98.

in hot stars, 60, 163.

series in, 95, 179, 180.

spectra of, 31, 180.

P.

Paschen and Runge, see Runge.
Periodic Jaw and atomic weight, 176.
Perry, atmospheres, stellar, 150.
Persei £, mass of, 149.
Perturbations, magnetic, Preston, 167,

190.

Pettenkofer, 165.
Phosphorescence and gaseous stars,

151.
Pickering:, bright line stars, 132.

new hydrogen series, 58, 59, 95,
98, 178.

stellar classification, 67, 130.
Piscian stars, 70.
Plant forms, 153, 171.
Pliicker and Hittorf, spectra of ele-
ments, 19.
Polarian stars, 70.
Pole, galactic, 124.
Polymerisation, 175, 190.
Poulton, Prof., 153.
Pressure and wave-length, 101.

shift, direction of, 105.
Preston, Dr., magnetic perturbations,

110, 113, 167, 187, 190.
Prism, action of, 2.
Procyon, see Canis minoris a, 68.
Procyonian stars, 70.
Proto-hydrogren, see Hydrogen.
Proto-metals, see Metals.
Prout, evolution, inorganic, 165.
Puppis, £, spectrum of, 58.

R.

Ramsay, helium, 48.
Reese, Zeeman eflect, 112.
Rig-elian stars, 70.
Hitter, dissociation hypothesis, 149.
Roberts -Austen, dissociation, 76.
Rocks, fossiliferous, 153.
Runge and Kayser, series and melting
points, 97.

series and temperature, 183.

series formula, 87.
Rung-e and Paschen, oxj-gen series, 179.

series, 11, 85.

Rutherford, on stellar spectra, 25.
Rydberg- and initial element, 89, 166.

hydrogen, series of, 95, 178.

on doubles, triplets, &c., 89.

series formula, 87, 189.

S.

Secchi. 25.

Selenium, series in, 95.

Schemer, Dr., magnesium in stars,

75.
Schuster, molecular combination, 74.

nebula?, 121.

on dissociation, 144.

series, law of, 86.
shifls, artificial, 107.
Series, aluminium, 87.

and atomic weights, 93.
chemical groups, 91.
dissociation, 83.
melting points, 93, 97.
temperature, 183.

calcium, 97.

constants of, 87, 88, 189.

definition of, 11.

doubles in, 89.

early notions of, 12.

formula? for, 11, 87.

hydrocarbons, 183.

hydrogen, 178, 179.
Balmer, 86.
new, 59, 98.
Stoney, 86.

irregularities in, 91.

laws of, 86.

magnesium, 97.

metals, 97.

nebulous, 89.

nomenclature of, 89.

oxygen, 95, 163.

principal, 89.

Runge and Paschen, 85.

sharp, 89.

single lines, 89.

sodium, 179.

thallium, 87.

triplets, 89.
Shifts, artificial, 107.

pressure, 102, 105.
Silicium, in stars, 61.
Sirian stars, 70.
Sirius, see Canis Majoris o, 68.
Sodium, longs and short?, 24.

series of, 178, 179.
Space, absorption in, 131.

chemical elements in, 142.

chemistry of, 120.

distribution of helium in, 142.
Spectroscope, grating, 9.

laboratory, 8.

simple, 2.

stellar, 16.
Spectrum, bright line, 5.

continuous, 4.

dark line, 12.

definition of, 2.

discontinuous, 5, 12.

length of, and temperature, 5.

INDEX.

197

Spottiswoode, induction coil, 32.
Star-way, Gould's, 126.
Herschel's, 126.
inclination of, 127.
Stars and nebula;, relation between,

132.

atmospheres of, 150.
behaviour of hydrogen in, 64.
bright line, 132.

distribution, 134.
brighter, McClean, 60.
carbon, 61.

and Milky Way, 129, 141.
and relative temperature, 56.
chief lines in, 56.
D liner's observations, 129.
length of spectrum, 55.
classification, chemical, 66.

general, 72, 129.
composition and distance, 142.
containing new hydrogen, 60.
cooler, chemistry of, 65.
distribution of, and Milky Way,

131, 141.

gaseous, and phosphorescence, 151.
and proper motion, 139.
and relative temperature, 56.
chief lines, in, 56.
length of spectrum, 55.
genera of, definitions, 70.
grouping of, chemical, 141.
hottest, 67, 159.
magnesium in, 75.
masses of, 150.
metallic, and length of spectrum,

55.

and Milky Way, 131, 141.
proper motion, 139.
relative temperature, 56.
chief lines in, 56.
flutings in, 141.
new, see Novae,
nitrogen in, 61.
Orion, temperature, 80.
oxygen in, 60.

proto-metallic, and proper motion,
139.

and Milky Way, 131, 141.
silicium in, 61.

similar type, localization of, 120.
spectra of, and proper motion,

140.

temperature of, 67, 68, 174.
white, and Milky Way, 131.
Stoney, hydrogen series, 86.
Strata, stellar and geological, 159.
Strontium, behaviour of, Humphreys,

106.

Struve, star distribution, 126.
Sulphur, series in, 95.
Sun, absorbing layer, 52.

convection currents, 146.
enhanced lines in spectrum, 35.

Sun (continued) —

spectrum of, and chromosphere,

39.

and Arcturus, 50.
temperature of, 80, 147, 151, 174.

T.

Tait, magnetic perturbation of lines,

109.

Taurian stars, 70.
Temperature, and chemical elements,

165.
and dissociation, 79.

length of spectrum, 5.
mass, 149.
proper motion, 139.
series, 183.
curve, 46, 68, 173.
high, effect of, 157.
intermediate stellar, 67.
of stars, 60, 68,70.
of sun, 80, 174.
stages of, 32.
Thallium, series, 87.
Thomson, Prof. J. J., chemical ele-
ments, 167, 188.
mass of molecules, 182.
Triplets, in series, 89.
Troost and Deville, specific gravity of
iodine, 74.

U.

Units, chemical, complexity of, 98.

V.

Valenta and Eder, mercury spectrum,

76.

molecular combination, 74.
Victorium, discovery of, 119.
Vog-el, H., Algol, mass of, 149.
Vog-el, W.,| wave-length, change of,

1C2.
Volumes, atomic, 165.

W.

Water, sea, composition of, 169.
Waters, S., nebulae and star clusters,

128.

Wave-length and magnetism, Cornu,
112.

and pressure, 101.
quantity, 103.
hydrogen series, 59.

198

Weismann, 170.

Wilson, temperature of sun, 174.

Wolf, Dr., spectra of nebulre, 121.

Y.

Young-, eclipse 1870, 41.
Yttria, fractionation of,

molecule of, 118.

" old," 117.

spectrum of, 118.

INDEX.

Z.

Zeeman, effect, Ames, 112.
Earhart, 112.
Larmor, 110.
Lorentz, 110.
on iron, 112.
Preston, 113, 114.
Reese, 112.

magnetic perturbations of lines,
110, 188.

HARBISON AND SONS, Printers in Ordinary to Her Majesty, St. Martin's Lane.

RETURN

TO— ^

MAIN CIRCULATION

ALL BOOKS ARE SUBJECT TO RECALL
RENEW BOOKS BY CALLING 642-3405

DUE AS STAMPED BELOW

AirmTwrror

\UG02'94

FORM NO. DD6

UNIVERSITY OF CALIFORNIA, BERKELEY
BERKELEY, CA 94720

m

m

m

m®.

TO*