Full text of "The sun"

See other formats

Jffpliif

SiBi'i? 1

: > ' ' ' • I ; , : ; ; - •

III!

THE SUN

CHARLES G. ABBOT, S.M.

DIRECTOR SMITHSONIAN ASTROPHtSICAL OBSERVATORY

WITH NUMEROUS ILLUSTRATIONS

NEW YORK AND LONDON

D. APPLE TON AND COMPANY

1911

, 1911

Printed in the United States of America

PREFACE

WITHIN the last fifteen years we have seen the publi-
cation of Rowland's great table of solar spectrum
wave lengths, the establishment of the Yerkes, Ko-
daikanal, Mount Wilson and other observatories
largely devoted to solar researches, the photography
of the spectrum of the corona and of the chromosphere
at total solar eclipses, Hale's brilliant discovery of
magnetic fields in sun spots, the determination of the
rotation periods of the sun at different levels, as well as
at all solar latitudes, Langley's bolometric investigation
of the sun's infra-red spectrum, and the recent Smith-
sonian determinations of the absolute intensity of the
solar radiation outside our atmosphere. The great
interest in such researches has been marked by the
establishment of the International Solar Union, and its
enthusiastic gatherings of the foremost investigators
from all lands.

The time seems ripe for collecting the splendid array
of new solar knowledge which such unprecedented
activity has produced, and for discussing the probable
nature of the sun in the light gained.

In the following pages the professional astronomer
will find hitherto unpublished results of researches,
and new explanatory hypotheses, illustrated by many
new text figures and engravings.

PREFACE

Chapter II has been devoted to a description of the
methods and principles employed in modern solar
research, and Chapters VII to X on the relations of the
sun to life upon the earth, and to the starry universe in
general. Thus the book, while primarily devoted to
the sun, may, I hope, serve as an introduction to the
study of astrophysics for school and college use, as
well as for the general reader.

In Chapters VI to IX are given many facts likely
to prove of interest to the meteorologist, geologist,
botanist and engineer.

Professor Young's "The Sun" is now out of print,
and as it is hoped that the present work may to some
extent take its place, I have been permitted to use
some of his illustrations, notably in Chapter IV, and
to make several quotations from his text, the longest
in Chapters IV and VI. I desire also to acknowledge
my obligations to many who have given suggestions,
information and illustrations. Especially I offer thanks
to F. E. Fowle, S. A. Mitchell, W. W. Campbell, G. E.
Hale and the staff of the Mount Wilson Solar Observa-
tory, E. B. Frost, H. M. Chase, A. G. Eneas, Henry
Holt and Company, the Superintendent of the United
States Naval Observatory, M. J. Moore of the United
States Patent Office, and Messrs. Briggs and Shantz of
the United States Department of Agriculture.

C. G. ABBOT.

WASHINGTON,
July 10, 1911.

CONTENTS

CHAPTER I

PAGES

THE SOLAR SYSTEM. THE SUN'S DISTANCE. ITS DIMENSIONS 1-30

CHAPTER II

THE INSTRUMENTS AND METHODS USED IN SOLAR INVESTI-
GATION . 31-84

The Telescope. — The Coelostat. — The Spectrum and
What It Indicates. — Spectroscopes. — The Spectro-
heliograph. — The Heliomicrometer. — The Comparator.
—Nature of Radiation — Laws of Radiation. — Spectra
of Different Sources. — Pyrheliometry. — Bolometry.

CHAPTER III

THE PHOTOSPHERE 85-127

Telescopic View. — The Photospheric Spectrum. — Row-
land's Spectrum Tables. — Chemical Elements Found
and not Found. — Corrections to Rowland's Wave-
lengths. — Levels . — Pressures . — Convection Currents . —
Limb Spectra. — The Variation of the Sun's Brightness. —
Solar Temperatures. — Spectro-heliography. — The Solar
Rotation.

CHAPTER IV

ECLIPSES AND THE OUTER SOLAR ENVELOPES . . 128-182

The Saros. — Eclipse Expeditions. — The Corona. — The

Chromosphere. — Jansen's and Lockyer's Discovery. —

The Spectra of the Chromosphere and Prominences. —

vii

CONTENTS

PAGES

Prominences and the Spectroheliograph.— Recent Flash
Spectrum Observations. — The Heights of Different
Metals in the Chromosphere. — Mitchell's Observations
of 1905.— Campbell's Observations.— Chromospheric
Spectra in Full Daylight.

CHAPTER V

SUN-SPOTS, FACULAE AND GRANULATION . . . 183-214
Sun-spot Periodicity —Drift —The Distribution of Sun-
spots. — Their Formation and Life History. — The Sun-
spot Level. — Langley's Typical Sun-spot. — Faculae. —
Granulation. — Sun-spot Spectra. — Sun-spots and Mag-
netism.— Radial Motion in Spot Penumbras.

CHAPTER VI

WHAT is THE SUN? 215-279

Young's Views. — Halm's Views. — Schmidt's Hypothesis.
— Julius' Views. — The Author's Views.

CHAPTER VII

THE SUN AS THE EARTH'S SOURCE OF HEAT . . 280-330
Causes of Low Temperatures at High Altitudes. — Meas-
urement of the Intensity of Sun Rays. — Dependence of
Solar Radiation on Air Mass — The Transmission of the
Atmosphere. — The "Solar Constant of Radiation." —
The Light of the Sky.— The Dependence of the Earth's
Temperature on Radiation. — The Fluctuation of Solar
Emission. — Geological Temperatures.

CHAPTER VIII

THE SUN'S INFLUENCE ON PLANT LIFE .... 331-361
Plant Requirements. — The Assimilation of Carbon by
Autotrophic Plants. — Etiolation. — Plant Geography. —
Light Requirements of Plants. — Heliotropism. — Plants
as Energy Accumulators.

viii

CONTENTS
CHAPTER IX

PAGES

UTILIZING SOLAR ENERGY 362-390

Experiments with Burning Mirrors. — The "Hot-box"
Principle. — Mouchot, Pifre and Ericsson. — Eneas' Solar
Engines. — Properties of Glass. — Solar Heaters and Res-
ervoirs.—Low Temperature Solar Engines.— Solar
Cooking Appliances.— Solar Metallurgy.— Resume".—
Quantity of Solar Energy Available. — Thermo-dynamic
Efficiency. — Reflecting Powers of Mirror Surfaces.

CHAPTER X

THE SUN AMONG THE STARS 391-434

Stellar Distances. — Magnitudes. — The Sun's Magnitude
and Light Emission. — The Solar Motion. — Star Groups.
— Double Stars. — Stellar Masses and Distances. — Mira
Ceti and the Sun. — Stellar Spectra. — The Classification
of Stellar Spectra. — Radiation Distribution. — Evolution
of the Solar System. — Stellar Evolution

CONCLUSION 435-437

INDEX 439-448

LIST OF PLATES AND ILLUSTRATIONS
IN TEXT

LIST OF PLATES FACING

PAGE

Langley's typical sun-spot Frontispiece

PLATE I. — Smithsonian observing shelter and coelostat

(Mount Wilson) 36

PLATE II.— The Zeeman effect (King) 44

PLATE III.— Solar photograph (Ellerman) .... 86
PLATE IV. — Fig. 1. Spectrum of procyon (Adams); Fig. 2.

Spectrum of east and west limbs of the sun (St. John) . 88
PLATE V. — Calcium spectroheliogram, H2 (Ellerman) . .114

PLATE VI. — Hydrogen spectroheliogram, Ha (Ellerman) . 116

PLATE VII. — Calcium spectroheliogram, Hx (Ellerman) . 118

PLATE VIII.— Calcium spectroheliogram, Ha (Ellerman) . 120

PLATE IX.— Calcium spectroheliogram, H2 (Ellerman) . 122

PLATE X.— Hydrogen spectroheliogram, Hy (Ellerman) . 124

PLATE XL— Hydrogen spectroheliogram, Ha (Ellerman) . 126

PLATE XII.— Solar corona, 1900, May 28 .... 132

PLATE XIIL— Solar corona, 1905, August 30 ... 134
PLATE XIV. — Spectroheliograms of solar prominences (Slo-

cum) 166

PLATE XV.— Flash spectra, 1905, August 30 (S. A. Mitchell) 174

PLATE XVI. — Photographs of a portion of the sun (Jansen) 202
PLATE XVII. — Spectra of photosphere and sun-spot (Mount

Wilson Solar Observatory) 210

PLATE XVIIL— Effect of pressure on gaseous spectra (Gale) 248

PLATE XIX.— The Pleiades (Ritchey) 398

PLATE XX A.— Stellar spectra (Campbell) . . . .406

PLATE XX B.— Stellar spectra (Campbell) ... 408

PLATE XXL— Stellar spectra (Hale and Ellerman) . . 410

PLATES AND ILLUSTRATIONS IN TEXT

PAGE

PLATE XXIL— The great nebula in Orion (Ritchey) . . 418

PLATE XXIIL— Nebula N. G. C. 6992 Cygni (Ritchey) . 420

PLATE XXIV.— The great nebula in Andromeda (Ritchey) 422
PLATE XXV.— Spiral Nebula M. 51 Canum Venaticorum

(Ritchey) 424

PLATE XXVI.— Fig. 1. Nebula H. V. 24 (Ritchey); Fig. 2.

Ring nebula in Lyra (Ritchey) 426

LIST OF ILLUSTRATIONS IN TEXT

FIQ- PAGE

1.— Relative distances of planets 5

2.— Geometrical parallax method 10

3.— Velocity of light (Fizeau) 19

4.— Velocity of light (Foucault) 20

5. — Determination of constant of gravitation .... 29

6. — Projecting the solar image 32

7. — Herschel's solar eye-piece 33

8. — Shade glasses 34

9. — Polarizing eye-piece 34

10. — Rapid exposing solar shutter 35

11. — Prismatic refraction of light 47

12. — Prismatic spectroscope 50

13.— Diffraction of light 51

14. — Plane grating spectroscope 53

15. — Concave grating spectroscope 54

16.— The 150-foot tower (Mount Wilson) .... 58

17. — Energy-spectra of the sun and the perfect radiator . 69

18. — Pouillet's pyrheliometer 73

19.— Abbot's silver disk pyrheliometer 76

20.— Angstrom's pyrheliometer 77

21. — Angstrom's pyrheliometer. Details 77

22. — Abbot's water flow pyrheliometer 79

23.— The bolometer 81

24. — Bolographs of the solar spectrum 83

25. — Brightness on solar disk 100

26. — Energy spectra on solar disk 109

xii

PLATES AND ILLUSTRATIONS IN TEXT

FIG. PAGE

27. — Huggin's first observation of a prominence in full sun-
shine 143

28. — Double reversal of the D-lines 147

29.— Double reversal of C-line 148

30. — " Arrow-head " spectrum 148

31. — Opened slit of the spectroscope 149

32 . — Chromosphere and prominences as seen in the spectrum . 151

33. — Relative frequency of protuberances and sun-spots . 153

34-39. — Eruptive prominences 157

40-45. — Quiescent prominences .. 159

46-51. — Forms of prominences 163

52. — Motion indicated by prominence spectra . . .165

53. — Sun-spots and terrestrial magnetism 187

54. — Sun-spots and terrestrial temperatures and magnetism . 191

55. — Spoerer's curves of sun-spot latitude 194

56.— Solar diagram (Young) 216

57. — Normal and anomalous dispersion 233

58. — Michelson's hypothesis 260

59. — Displacement diagram 261

60.— March of insolation (Mount Wilson) .... 285

61. — Holographs of the solar spectrum 292

62. — Atmospheric transmission plats 294

63. — Insolation and terrestrial temperatures . . . .316
64. — Apparent variations of the sun's radiation (Smithsonian

observations) 321

65. Hypothetical temperature diagram 325

66.— Stomata (Schwendener) 338

67. — Promotion of carbon assimilation by light, and absorp-
tion of light 343

68. — Eneas' solar engine 367

69. — Boiler of Eneas' solar engine 368

70. — Adams' solar cooker 380

71. — Intensity of sun rays (Mount Wilson and Washington) 384
72. — Intensity of solar radiation (Different latitudes and

elevations) 385

LIST OF TABLES

I. — Principal data of the solar system .... 3

II. — Velocity of light in different media .... 46

III. — Principal solar spectrum lines 90

IV. — Chemical elements found in the sun . . . 91

V. — Chemical elements doubtfully occurring in the sun 92

VI. — Corrections to Rowland's wave-lengths ... 96

VII. — Distribution of radiation over the sun's disk . . 107

VIII. — Energy spectrum relations over the sun's disk . 110

IX. — Daily rotation of the sun's surface .... 126

X. — Measures of ninety-two lines in the flash spectrum

near H§ . . . 178

XI. — Years of sun-spot maxima and minima and maxi-
mum intensities 186

XII. — Hydrogen spectrum in sun-spots .... 206

XIII. — Spectrum lines affected in sun-spots . . . 206

XIV. — Sun-spot, hot arc and cool arc spectra . . . 209
XV. — Moissan's experiments on the vaporization of the

elements of the iron family .... 238
XVI. — Differences between observed and computed values

of atmospheric transmission : 242
XVII. — Observed intensity of solar radiation, Mount Wil-
son, July 6, 1910 237

XVIII. — The intensity of radiation in different parts of the

solar spectrum . . . ... . . 288

XIX. — Transmission for total solar radiation . . . 296

XX. — Atmospheric transmission for homogeneous rays 297

XXL— Sunlight and sky light (Exner) . . . .301

XXII.— Sunlight and sky light (Roscoe) . . . .302

LIST OF TABLES

TABLE

XXIII.— Sky light on a vertical surface (Exner, Schramm) 303
XXIV.— Sun and sky light. Relative brightness for dif-
ferent wave-lengths (Mount Wilson) . . 305
XXV.— Average brightness of sky zones. Flint Island

and Mount Wilson . . . 395

XXVI. — Ratio of total radiations: Sky to sun . . 307

XXVII.— Yearly means and mean daily temperature de-
partures 311
XXVIII. — Composition of food products

XXIX. — Economy of Helianthus Annuus . . 361

XXX.— Percentage reflecting powers of various reflectors 3 88
XXXI.— Classification of stellar spectra 410

XXXII.— Intensities in stellar spectra (Wilsing and Schei-

ner) . . . .411

XXXIII.— Spectroscopic binaries. Spectral types, periods and

eccentricities (Campbell) . . . 422

XXXIV.— Spectral types and velocities in space (Kapteyn) 430

INTRODUCTION

WE depend on the sun for life, warmth, light, and
ajl mechanical and electrical powers. Its constant
supply of heat is necessary to prevent the oceans and
even the air from freezing. The supply of coal which
we are now using is but an evidence of the sun's light
in former ages. To pass to the enumeration of the
comforts and luxuries and beauties we owe to the
solar rays would lead us far astray, and is indeed
wholly unnecessary because all men acknowledge
and many worship the sun as the source of these bene-
fits. It would be a gross neglect to omit the closer
investigation of such unique relations as those the
sun maintains to life. Yet the study of the means
of increasing the usefulness of the sun has been ne-
glected, and it is rather in the investigation of its
curious features that solar researches have gone
farthest.

The enormous brilliance and heat of the solar rays
suggestive of temperatures far above any which can
be produced on the earth; the marked dimness and
brown shade of the edge or limb of the sun relative to
the center; the fluctuating march of spots across the
disk; the variable rates of rotation of the sun's sur-
face in different latitudes; the brilliant markings
2 xvii

INTRODUCTION

called faculse which accompany the spots; the weird
and highly beautiful phenomena of total solar eclip-
ses; all these have long been the objects of minute
study. In the last half century the development of
the spectroscope has led to great progress in the more
intimate and satisfactory knowledge of the sun; so
that we now know many of the chemical elements of
which it is composed; the approximate temperature
of its surface ; the motion of the vapors at and near
the surface; the approximate pressure under which
they lie ; the magnetic character and cyclonic struc-
ture of sun spots ; their relative coolness as compared
with their surroundings; besides many other details
hardly to be credited as known of a body situated
nearly ninety-three millions of miles away.

By bonds unseen yet altogether stronger than a
bar of steel thousands of miles in diameter the sun
holds to itself the moving family called the solar
system, comprising the earth and moon, the seven
other great planets with their satellites, half a thou-
sand asteroids, or minor planets, besides numerous
comets and meteorites. It has required the lifelong
labors of many men of exceptional genius, like New-
ton, coupled with centuries of no less praiseworthy if
less brilliant accumulations of accurate observations
to have given us the full knowledge which we now
enjoy of the distances, dimensions, masses and orbits
of the solar system.

Although the distance from the sun to the orbit of
Neptune is 2,800,000,000 miles, the solar system is

xviii

INTRODUCTION

but a speck in the vast universe of the stars. In the
year 1901 there flared up in the constellation Perseus
a new star which for a few d&ys rivaled the brightest
stars of the heavens in its brilliancy, and then slowly
faded away into insignificance. To dwellers on the
earth this sight was new in 1901, but in reality that
new star was so far away that its sudden burst of
light, traveling toward us 186,000 miles a second,
had been on the journey since the days of Cromwell,
and the star had faded away nearly three centuries
ago. At such enormous distances are the stars that,
although some of them are believed to be millions of
miles in diameter, they present no real disks even in
the largest telescopes, so that the details of their sur-
faces cannot be examined. Nevertheless, by the
powerful aid of the spectroscope, much is known of
the chemical constitutions of the stars; and many of
them have been shown to belong to revolving sys-
tems, the components of which, though separated in
some instances by greater distances than is Jupiter
from the sun, are separately indistinguishable to the
telescope. Still we might despair of knowing much
more of the physics of the stars if it were not that the
spectroscope shows also that the sun is but one of
them close by, and that a large class among the stars
is probably in a similar condition as to temperature,
and made up of the same chemical substances as the
sun.

Both telescopic and spectroscopic observations
have shown that the solar system is moving at a rapid

xix

INTRODUCTION

rate towards the constellation Hercules, although no
change in the aspect of the heavens sensible to the
naked eye would accumulate for an immense period
of years. By means of the displacement of the earth
in its orbit around the sun, amounting to over 180
millions of miles, semi-annually, it has been possible
to obtain with fair accuracy the distances of many of
the individual stars, and from these by statistical
methods to go further and estimate the average dis-
tances of all the stars of a given brightness. These
various examples indicate the important place of the
sun in stellar investigations; and indeed the study
of the sun as a typical star, though quite recently
developed, seems bound to throw much more light
on the subject of the nature of the universe.

Considering the sun as the fountain of light and
heat upon the earth, perhaps the first question which
suggests itself is this: How much radiant energy
reaches the earth from the sun in a given time? This
utilitarian branch of solar investigation has been
comparatively neglected. No more striking proof of
the neglect need be cited than to say that the fore-
most text-book on meteorology, published since 1900,
states various determinations of the intensity of
solar radiation at the earth's mean distance which
range from 1.76 to 4.06 calories per square centi-
meter per minute. Of these the author of the text-
book prefers one which is confessedly the mean of
such divergent numbers as 2.63 and 3.50, one of
which numbers was thought by its originator to be

INTRODUCTION

too low, and the other too high! As will be shown
later, there can be little question now (1910) that
the true value is about 1.95 calories; but how remark-
able it is that one of the fundamental constants of
Nature should have been uncertain within such wide
limits, so late as the beginning of the twentieth cen-
tury. Imagine, for analogy, that it had been stated
in a standard work on astronomy, published in 1905,
that the sun's distance (which is of no greater im-
portance than the constant of radiation) might be
anything between 80 millions and 200 millions of
miles so far as known, and that it was generally sup-
posed to be 140 millions!

Some of the more important questions connected
with the sun's action as the fountain of light and heat
are the following: Is the solar radiation uniform or
variable? What losses does it suffer in the earth's
atmosphere? Are there changes of transparency in
the sun's outer layers sufficient to alter the earth's
supply of radiation appreciably? How much solar
radiation does the earth reflect, unused, to space?
How does the earth's temperature depend on solar
radiation and on the transparency of the air? If
there should be variations of solar radiation, how
great changes of temperature of different stations
on the surface of the earth ought to follow, and how
long would such responses be delayed? In short,
are solar studies applicable to weather prediction?
What methods, if any, can be economically used to
store and employ the sun's energy for power or heat-

xxi

INTRODUCTION

ing? What influences do changes in the intensity
or color of the light falling on different plants produce
on their growth and fruitage? May advantageous
variations of plants be promoted by the control of
their radiation supply? What can be done with
solar rays for the promotion of health?

In the pages which follow the sun will be considered
in these three aspects : First as the controlling mem-
ber of the solar system; second, as an object of in-
quiry, interesting in itself, but still more so as the
nearest star, and typical of a large class of stars;
third, as the fountain of light and- heat, and through
them of life on the earth. It is indispensable to any
satisfactory understanding of the second and third
branches to be familiar with the methods and princi-
ples which are now being employed in solar investi-
gations. For the convenience of the reader a general
account of these is given in Chapter II, which there-
fore has to do directly with physics, and only sec-
ondarily with the sun, but which forms the ground-
work of the chapters relating directly to solar phe-
nomena. Illustrated descriptions of some of the
instruments used in solar research will also be found
in Chapter II, and appropriate references to these
descriptions and to the statements of the general
relations will be found in the text of subsequent
chapters.

To avoid premature discussion, the various solar
phenomena will be described first without much
attention to their explanation, except as seems nec-

xxii

INTRODUCTION

essary to fix attention upon significant facts. Solar
theories are dealt with in Chapter VI. One principal
exception to this course is in the frequent noting of
applications of KirchhofTs discoveries on the rela-
tions of temperature, radiation, and absorption. It
may be that future writers on the sun will attribute
to anomalous dispersion many of the phenomena
here set down as due to absorption and motion. But
although anomalous dispersion hypotheses are so
strongly advocated by Julius, the writer feels confi-
dent that his own preference for the older views is
still shared by most students of solar physics.

What, after all, is the sun, and how can we best
explain the principal solar phenomena? No doubt
many will find the views here advanced heretical,
but for the writer the existence of the cloudy photo-
sphere, so firmly believed in by most solar observers,
seems so highly improbable that he has ventured to
advocate the view of a purely gaseous sun. But in
doing so it is not Schmidt's refraction theory to
which he turns to explain the sharp solar boundary.
According to Lord Rayleigh our own atmosphere,
if freed from dust, would still scatter light by the
action of the gases themselves. Schuster and Natan-
son have computed this effect, independently, and
both find that the purely gaseous scattering goes far
to explain in full the observed weakening of the direct
beam of the sun above Mount Wilson for rays
which are not selectively absorbed. This weakening
amounts to several per cent. If then the gases of

xxiii

INTRODUCTION

the atmosphere of the earth, which extend, in density
sufficient to scatter light appreciably, perhaps only
fifty miles in altitude, suffice to scatter several per
cent, of a beam of light, it seems probable that we
can see at the most not more than a very few thou-
sand miles into the gaseous body of the sun, which, at
the layer producing the Fraunhofer lines, seems to
be under several atmospheres of pressure. Admit-
ting this, how deep, measured radially, can one see
near the sun's edge, where the few thousand miles
above mentioned will lie along a line of sight nearly
tangent to the sun? It would seem that at the sun's
edge a shell of gas of only a few hundred miles in
thickness must suffice to fully veil all that lies below.
Viewed from the earth this would correspond to a
fraction of a second of arc, so that a gaseous sun at
93,000,000 miles away would present an apparently
sharp boundary. From these considerations depend
various consequences adapted to the explanation of
solar phenomena. To be sure there are several
apparently powerful objections to this view of a
purely gaseous sun, but they seem not to be in-
superable.

My good friend, Prof. J. C. Kapteyn, has en-
couraged me to set down several hypotheses which
can be regarded as only slenderly founded. Among
these are the hypotheses of the causes of some strange
phenomena of geological climates, touched upon in
Chapter VI, and more fully discussed in Chapter
VII ; the hypotheses of the causes of some peculiar-

xxiv

INTRODUCTION

ities of stellar evolution given in Chapter X; and
as some readers may be disposed to think, even the
explanation of solar phenomena already mentioned
which occupies a large part of Chapter VI. Pro-
fessor Kapteyn is of the opinion that a bushel of chaff
is worth searching by a Crusoe if it contains some
grains of corn that will sprout, and so my defense
for my temerity in including such speculations is that
they may interest some readers to begin some more
fruitful researches.

THE SUN

CHAPTER I

THE SOLAR SYSTEM. THE SUN'S DISTANCE. ITS
DIMENSIONS

THE objects which appear to move among the stars,
namely the sun, planets, minor planets, moons, me-
teors and comets,1 compose the solar system. For-
merly it was believed that all the heavenly bodies
revolve about the earth. But now the theory of
Copernicus is fully verified, and the earth is known to
be only a planet, of much smaller size than Jupiter,
Saturn, Uranus or Neptune, though larger than Mars,
Venus or Mercury, and like the other planets it re-
volves about the sun. Galileo was threatened with
torture and forced to perjure himself because he be-
lieved this, which shows how fortunate we are to live
in the present age.

The moon has actually only — — as great a diameter

4UU

as the sun, although they appear to be about equal.
It is the immense distance by which the sun and
planets are separated from the earth in comparison
with the distances we are accustomed to travel over

1 Not all the comets remain permanently attached to the solar
system, but many of them do.

THE SUN

or even as compared with the distance of the moon,
which prevents us from immediately realizing the
great bulks of these distant bodies. In the following
table is a summary of the approximate dimensions
and principal characteristics of the larger members
of the solar system. The means of determining the
sun's distance, dimensions and rotation will be given
later.

GRAVITATION

Accustomed as we are to regard inches and feet
as ordinary, miles as considerable, and thousands of
miles as very great distances, it may seem almost in-
credible that there should be any bond between the
the sun and Neptune, situated as they are 2,800,000,-
000 miles apart. There is, however, a bond between
them so strong that it would require the strength of a
bar of steel 500 miles in diameter to take its place
in preventing the escape of Neptune from the sun.
This bond we call gravitation. Every body in the
universe is believed to attract every other body in
the universe with a force proportional to the mass or
quantity of matter the body contains, and inversely
proportional to the square of the distance between
their centers of gravity. On the one hand, this law
of gravitation applies between all bodies on the earth
as well as between the earth itself and any one of
them; and, on the other hand, there is evidence that
it holds also among the fixed stars. The weight of a
stone is the measure of the attraction between it and

THE SOLAR SYSTEM

Dens
Wate

^
1-1 O O rH

COff-

CO O t^ O

^-' i"H O O O CO

<NOOOi-iOOOOcOO»O

Axi

rotati

Orbital
revolutio

Inclination
of orbit
to ecliptic.

^i °°

a a a a 3

r~ e- CO b- O -* Jl

i^-d »o co toi-H t^

(M CO
(M (M (

O t^

T-I s^. e^- <N

o O O
t-( CO CO

an d
n mi

nt an
meter

§O 00
O i— i
l^» O5

rco"t>^t^

O O"J

t^ CO -^ »O >O rH 00 t^ OC' «

' O *O
' CO O

i>» O^ ^O
CO C4 O

111

§°9

CT> >O 00
00 •

i-l 't1 O5 • CO O

THE SUN

the earth, and with a sufficiently sensitive apparatus
the attraction of two stones for each other may be
clearly shown. If weighed with a sufficiently deli-
cate spring balance, a weight will be found lighter at
the top of a mountain than at its base, in the ratio
of the square of the distances of the earth's center at
the two places of observation. There are chemical
balances so delicate that an object appears to weigh
differently accordingly as the weights are placed side
by side or one on the top of the other, and this is
because of the difference in distance from the weights
to the earth's center in the two cases.

The attraction of gravitation between the sun
and Neptune amounts to 8 x 1016 (8 followed by
sixteen ciphers) tons. If there was nothing opposed
to this force, Neptune obviously would fall into the
sun. It is the motion of the planet in its orbit, at
right angles to the line leading towards the sun, which
maintains the distance between them.

Kepler's laws of planetary motion are as follows:

I. The orbit of each planet is an ellipse, with the
sun in one of its foci.

II. The radius vector of each planet describes equal
areas in equal times.

III. The squares of the periods of revolution of the
planets are proportional to the cubes of their mean
distances from the sun.

There is but little difference between the major and
minor axes of the elliptical orbits of most of the plan-
ets, or, in other words, their orbits are nearly circular.

THE SOLAR SYSTEM

But this is not true of the orbits of Mercury and Mars,
as is shown in Table 1.

It may be surprising to some readers that Kepler
could have possessed so much knowledge of the dis-
tances of the planets from the sun as to enable him to
verify the third law, while as yet the actual distances
in miles were not even roughly known. But only
the ratios of the distances of the planets were re-
quired by Kepler, and these could be fixed inde-
pendently of the actual distances by the following
method, known since the days of Hipparchus, which
I quote from Young's "The Sun."

"First, observe the date when the planet comes
to its opposition, i.e., when sun, earth, and planet
are in line, as in the fig-
ure, where the planet
and earth are repre-
sented by M and E.
Next, after a known
number of days, say
one hundred, when the

planet has advanced to M'and the earth to E', observe
the planet's elongation from the sun, i.e., the angle
M'E'S. Now, since we know the periodic times of
both the earth and planet, we shall know both the
angle MSM' moved over by the planet in one hun-
dred days, and also ESE', described in the same time
by the earth. The difference is M'SE, often called
the synodic angle. We have, therefore, in the tri-
angle M'SE', the angle at E' measured, and the angle

THE SUN

M'SE' known as stated above, and hence by the
ordinary processes of trigonometry we can find the
relative values of its three sides. "

Thus, by means of comparatively simple astronom-
ical observations, all the relative distances in the
solar system can be fixed with high accuracy. It is a
work of far greater difficulty, as we shall see, to meas-
ure any of these distances absolutely.

Kepler's three laws were known before the year
1620, but without explanation. Sir Isaac Newton
discovered, about the year 1679, that all three laws
are direct consequences of the laws of motion, pro-
viding it is assumed that all bodies attract one an-
other with a force varying inversely as the square of
the distance. This latter principle is Newton's law
of gravitation.

At the present day no well-informed person ques-
tions either the Copernican system or the universal
sway of gravitation ; for, while not every such person
has the mathematical knowledge requisite to examine
all the proofs of these fundamental facts, he yet feels
entire faith in the conclusions unanimously agreed
upon by such masters as Kepler, Newton, Laplace,
and many others of scarcely less renown, who have
overcome the tremendous mathematical difficulties
in which the knowledge of the motions of the solar
system is involved. Every planet and satellite at-
tracts every other, and perturbs its motion from the
simple orbit which would exist if there were only two
bodies concerned. In a large astronomical library

THE SOLAR SYSTEM

one may find printed in a quarto or folio volume the
final equation representing the motion, for instance,
of the moon. Such an equation comprises line upon
line and page after page, including thousands of
terms required to account for all the disturbing fac-
tors. None but a master can handle such a problem.

Prof. E. W. Brown, writing in 1904 1 of his inves-
tigation of the theory of the moon's motion, says:

" A few brief details about the amount of time and
labour expended may not be uninteresting. From
1890 to 1895 certain classes of inequalities were cal-
culated, but the work was only begun on a systematic
plan, which involved a fresh computation of all the
inequalities previously found, at the beginning of
1896. Mr. Sterner began work for me in the autumn
of 1897 and finished it in the spring of the present
year, though neither of us was by any means contin-
uously engaged in calculation during that period.
He spent on it, according to a carefully kept record,
nearly three thousand hours, and I estimate my share
as some five or six thousand hours, so that the calcu-
lations have probably occupied altogether about eight
or nine thousand hours. There were about 13,000
multiplications of series made, containing some 400,-
000 separate products; the whole of the work re-
quired the writing of between four and five millions
of digits and plus and minus signs. Although the
problem now completed constitutes by far the longer

1 Royal Ast. Soc. Monthly Notices, vol. Ixv, p. 107, 1904.
3 7

THE SUN

part of the whole, much remains to be done before
it is advisable to proceed to the construction of
tables."

Every large scientific institution or observatory
has almost daily communications from persons of
very moderate attainments who presume to question,
nay rather to spurn, the most well-attested facts of
human knowledge. Such persons seem to prefer
especially to direct their attacks on the following
facts: the Copernican system; the law of universal
gravitation; the first and second laws of energy;
and, finally, the high temperature of the sun. No
argument can refute them, because they have not
the requisite learning to comprehend it, which is no
disgrace, but which should make men modest enough
to have faith in those who excel them immeasurably.
Hence it is the policy of most scientific institutions to
avoid entirely discussions of these subjects with such
correspondents.

Professor Newcomb tells, in his " Reminiscences
of an Astronomer," of such a critic who called upon
him and announced his disbelief in Sir Isaac Newton's
theory of gravitation. Professor Newcomb proposed
to the skeptic that he jump out of the window and
convince himself of the existence of gravitation.
Being thus pressed, the visitor stated that he be-
lieved that gravitation extended no further than the
air, and did not go up to the moon. Professor New-
comb asked him if he had ever been there to see, and
when his caller answered "No," replied that, until

THE SUN'S DISTANCE

one of them could go to the moon and try the experi-
ment, he doubted if they could ever agree!

THE SUN'S DISTANCE
(1) Geometrical Methods.

Since the ratios of the distances between all the
principal members of the solar system can be fixed
with great accuracy by ordinary astronomical obser-
vations, it suffices to measure accurately in miles or
kilometers the distance from the earth to the sun or
to any one of the planets, and this fixes the scale of
the whole system. The great astronomical unit is
the mean distance from the earth to the sun, and
many determinations of it have been made in the last
250 years. Still astronomers are not quite content,
although there is no doubt that we know the distance

to within part of itself. To avoid using many

J-UUU

figures, it is customary to speak of the sun's " paral-
lax" instead of its distance. The parallax is the
angle which the earth's equatorial radius would sub-
tend if viewed from the center of the sun at the mean
solar distance. This angle is nearly 8.80 seconds of
arc, or about 0.000044 in circular measure. In other
words, the earth's mean radius is 0.000044 of the
sun's mean distance, and the latter is about 92,900,-
000 miles.

Since the solar parallax is so small, the usual
method of surveyors for finding the distance of an
inaccessible object is not applicable here. For, if

THE SUN

the earth's radius of 4,000 miles subtends only 8.8",
no two stations on the earth's surface which can
be seen from each other could possibly be far enough
apart to serve as a suitable base line for a solar
triangulation. Fortunately the observer can avail
himself of the fixed stars in the investigation. These
may be regarded as an infinitely great distance away,
and an apparent displacement of objects in the solar

FIG. 2.

system among the stars may be observed from two
stations at opposite sides of the earth, or at the
same station by two observations several hours
apart. The following explanation of this parallax
method is quoted from Young's "The Sun."

"Fig. 2 illustrates the method of observation.
Suppose two observers, situated, one near the north
pole of the earth, the other near the south. Looking

THE SUN'S DISTANCE

at the planet, the northern observer will see it at N
(in the upper figure), while the other will see it at S,
farther north in the sky. If the northern observer
sees it as at A (in the lower part of the figure), the
southern will at the same time see it as at B; and,
by measuring carefully at each station the apparent
distance of the planet from several of the little stars
(a, 6, c) which appear in the field of view, the amount
of the displacement can be accurately ascertained.
The figure is drawn to scale. The circle E being
taken to represent the size of the earth as seen from
Mars when nearest us, the black disk represents the
apparent size of the planet on the same scale, and the
distance between the points N and S, in either figure
A or B, represents, on the same scale, also, the dis-
placement which would be produced in the planet's
position by the transference of the observer from
Washington to Santiago, or vice versa."

Dr. Gill, lately the Astronomer Royal at the Cape
of Good Hope, has made many highly accurate meas-
urements by this method. He observed the oppo-
sition of Mars in 1877 at Ascension Island, employing
for his measurements a heliometer loaned by Lord
Lindsay. This instrument is a telescope having its
lens cut in halves, and having a micrometer screw for
sliding the parts with reference to each other, thus
enabling the observer to make the images of two
stars formed by the two halves to coincide. This
device is the most accurate one known for measuring
small angular displacements between stars. Dr. Gill

THE SUN

determined the displacement of Mars among the
stars as measured by evening and morning observa-
tions, and continued the work for several weeks. From
these measurements he obtained 8.780"±0.020" for
the sun's parallax.

Several of the minor planets or asteroids, though at
greater distances from the earth, have proved more
favorable objects for these measurements than Mars.
Being smaller and not so highly colored, it appears
that more accurate measurements of their projections
among the stars can be made. In 1889 and 1890 a
concerted system of observations was made upon the
asteroids Victoria, Iris, and Sappho by Dr. Gill, Dr.
Elkin of the Yale College Observatory, and several
German observers. Their results range from 8.796"
to 8.825", and their mean is 8.807" ±0.006".

The discovery of the minor planet Eros, in 1898,
furnished an object so much more favorable than any
other for parallax determinations by this method that
a great parallax campaign was lately carried through
upon this asteroid by many of the leading observa-
tories. Undoubtedly a still more thorough one will
be undertaken in 1931. Eros has a very elliptical
orbit; so much so that when nearest the earth in the
most favorable oppositions its distance is only 13,-
500,000 miles, and its parallax then becomes as great
as 60"; while at its most unfavorable oppositions its
nearest distance is 74,000,000 miles and its parallax
only 11". In the opposition of 1900-1, its nearest
approach was 30,000,000 miles, but in 1931 it will be

THE SUN'S DISTANCE

within half that distance. Prof. Arthur Hinks
has lately completed and published the reduction of
the photographic measurements of the 1900-1 in-
ternational Eros campaign, and he obtains the solar
parallax as 8.807" ±0.0027".

The method just described for fixing the scale of
the solar system is generally believed to be the best of
all at present. But there are several other methods
which deserve mention, and first, on account of, its
historical interest, the method of the transit of Venus.
This planet as a dark spot passes occasionally be-
tween us and the sun's disk. The transits occur in
pairs, about eight years apart, and the pairs occur
only about once a century. Those of June, 1761 and
1769, and of December, 1874 and 1882, were all ob-
served with great attention by astronomers of sev-
eral nations, for they were regarded, until very
recently, as yielding by far the best means of fixing
the solar parallax. All the methods used depend, of
course, on the displacement of the planet on the sun's
disk, as viewed from opposite sides of the earth.

In Weld's " History of the Royal Society of Great
Britain," may be found several quaint items relating
to the transits of 1761 and 1769. We, of America,
sometimes get the impression in our early school days
that King George III was only a crazy old despot,
and it will be a satisfaction to many of us to know
that he was a very liberal patron of the best scientific
enterprises. At the request of the Royal Society he
ordered £1,800 to be appropriated for the observa-

THE SUN

tion of the transit of 1761, and in addition, the Ad-
miralty directed a ship of war, the Sea-Horse, to
convey the observers to Bencoolen in India. The
Rev. Nevil Maskelyne, afterwards Astronomer Royal,
was sent to observe at Saint Helena.

The two observers sent to Bencoolen were Mason
and Dixon, the names so famous in American history
on account of their survey of " Mason and Dixon's
Line," which afterwards led to the popular name of
" Dixie" for the South-land. Their ship, the Sea-
Horse, engaged a French frigate almost at the shores
of England. The stands for instruments were dam-
aged by shots, and the observers could hardly be in-
duced to reundertake the journey. On account of
the delay thus occasioned, they observed at the Cape
of Good Hope, instead of proceeding to India.

For the transit of 1769 the King, at the Memorial
of the Royal Society, provided even irfore liberally.
He ordered £4,000 clear of fees to be paid over, and
that any balance which might be unexpended should
be for the use of the Society. In addition, the ship
Endeavor, under the command of Lieutenant, after-
wards the famous Captain, James Cook, was ordered
to the Pacific Ocean to take part in the observations.
Cook observed the transit successfully at what is now
called Venus Point, on the island of Tahiti. The
Royal Society sent observers also to Hudson's Bay
and to India on this occasion.

In 1874 and 1882 very elaborate preparations were
made by the governments and private astronomers of

THE SUN'S DISTANCE

many countries, including our own. Observations
were made all over the world and with many kinds of
apparatus, including heliometers, micrometers, and
photographic apparatus. Many thousands of ob-
servations were made.

The results of the several transits of Venus are on
the whole disappointing. A general discussion of the
observations of 1761 and 1769 was made by Encke in
1822, and he found the solar parallax 8.5776". More
recent recomputations have shown that the transit of
1769 may be said to indicate a parallax of between
8.7" and 8.9". From the transits of 1874 and 1882
different astronomers have computed widely different
results, ranging from 8.89" down to 8.75". Newcomb
adopts 8.794"0 ± .022".

(2) The Gravitation Methods.

Thus far we have considered only geometrical
methods for determining the parallax, and now we
may notice another class of quite a different charac-
ter— the gravitational methods, so called, which de-
pend on noting the perturbing effects of the different
planets and satellites on one another. One of the
best of them depends on observations of the motion
of the moon. It was Hansen's studies of the moon's
parallactic inequality which led him to announce, in
1854, the inadmissibility of Encke's value, 8.5776",
for the solar parallax. The perturbation of the
moon's orbit by the sun causes the interval from new
moon to first quarter to be about eight minutes longer

THE SUN

than that from the quarter to full moon. This in-
equality depends on the ratio of the radii of the orbits
of the moon and the earth. Hence, as the moon's dis-
tance is known, the solar parallax could be determined
if the inequality could be exactly measured. New-
comb gives 8.794" as the most probable result thus
obtained, but Prof. E. W. Brown's recent able investi-
gation of the moon's motion leads to the value 8.778".

Another gravitational method, proposed by Lev-
erier, has the advantage of cumulative increase of
accuracy. It depends on the secular perturbations
of the orbits of the planets, especially of Venus and
the minor planets, by the earth, which cause motions
of their nodes and periphelia. As time goes on, the
displacements thus caused are continually additive,
and will eventually be so large as to be determined
with very high accuracy. Leverier, indeed, thought
it not worth while to observe the solar parallax by
other methods, since this must eventually lead them
all. Newcomb gives, as the most probable mean re-
sult obtained by Leverier's method, 8.768". A fav-
orable application of Leverier's method may be made
in the case of Eros. G. Witt has found, thus, the
ratio of the sun's mass to that of the earth and moon
combined as 328,882 ±982. From this he computed
a solar parallax of 8.794" ±0.009". Great improve-
ment in the accuracy of this Eros result will come
after the close opposition of 1931.

It is noticeable that the mean of the results
obtained by the various gravitational parallax

THE SUN'S DISTANCE

methods falls below that obtained by the purely
geometrical methods used in the minor planet cam-
paigns. Most astronomers would attribute this to the
lesser accuracy of the gravitational methods at pres-
ent. It is certain, however, that the geometrical
minor planet method tends to give too high results,
owing to the difference of atmospheric refraction be-
tween a minor planet, Eros, for instance, and the
comparison stars; for the minor planets shine by
reflected sunlight, and their light cannot be as rich
relatively in the blue end of the spectrum as sunlight
itself is, because of the smaller reflecting power of
nearly all solid substances for blue than for red light.
On the other hand, the light of most stars is rela-
tively richer in the blue end of the spectrum than is
that of the sun. It is probable, therefore, that, in
the mean, the comparison stars for Eros, for instance,
are bluer than the sun, while Eros itself is redder than
the sun. Now the point of the method lies in deter-
mining the apparent displacement of Eros at two
stations far apart on the earth's surface, or, still
better, by morning and evening observations at the
same station. In the last-mentioned method, when
Eros is low in the east its altitude above the horizon
will be increased by atmospheric refraction, but the
comparison stars, being bluer, will be raised more
than Eros. Similarly in the west. The effect is to
make the parallax of Eros, and hence that of the sun,
too large, no matter whether the observations are
made simultaneously by two distant observatories,

THE SUN

or by one observatory in the morning and evening.
It is not yet determined how considerable this error
is, but it should be investigated with care.

(3) Dependence of the Sun's Distance on Geodesy.

In all the methods thus far referred to, the solar
parallax is obtained before the sun's distance, and the
astronomer requires of the geodetic surveyor to tell
him the dimensions of the earth, if he wishes to pass
from the parallax to the actual solar distance. Accu-
rate measurements of the earth require pendulum
observations at many stations to fix the earth's shape,
and, besides this, they require the actual measure-
ments by triangulation of very long arcs of the earth's
surface, and these depend finally on measurements
of a base line of a few miles in length. Base lines are
measured by repeatedly setting end to end, under
microscopic observation and on specially leveled
supports, short bars or tapes, whose temperature
must be observed throughout the process. Such
measurements of base lines are now made with an
error of less than one part in a million. The final
result from the whole net of triangulation recently
completed across the United States by the Coast and
Geodetic Survey is thought to be accurate to within
eighty-five feet in a total distance of nearly 3,000
miles. Several determinations of the earth's dimen-
sions have been made within fifty years. They give
the earth's mean equatorial radius as 3963.1 miles,
with a probable error less than one part in 20,000.

THE SUN'S DISTANCE

(4) Velocity of Light Method.

We now come to an important class of observations
depending on the velocity of light, and called the
" physical parallax methods," by which we may find
the sun's distance directly. Several ways have been
employed for measuring the velocity of light, but the
two best are the toothed-wheel method of Fizeau and
the revolving mirror method of Foucault. In Fiz-
eau's method (see Fig. 3) a beam of light starts from
a source at L and, after passing through the lens A
and being reflected by the thinly silvered plane glass

FIG. 3.

plate B, comes to focus and passes between two teeth
of a wheel at F. Thence the ray goes on to the lens C
and, after traveling a great distance, is focused by
the lens D upon the mirror E, which returns it on its
course, so that at length it passes again between two
teeth at F, and a part of it comes to the observer at H.
Now imagine the wheel in rapid rotation. The light
will then be cut off by every tooth which passes F,
and will thus consist of a series of flashes. But,
owing to the persistence of vision, the beam will still
seem continuous if observed at H, though it will
be weakened because it shines only intermittently.

THE SUN

Time is required, however, for the light to pass from
F to E and back to F, and meanwhile the tooth next
F has advanced, and may be in such a position as
exactly to cut off the returning beam, so that the eye
at H will see nothing. By gradually increasing the
speed of the wheel the light is alternately cut off and
transmitted. By counting these changes from light
to darkness, and knowing the number of teeth and the
speed of the wheel and the distance FE, the velocity
of light is measured.

The method of Foucault is illustrated by Fig. 4.
Light from the slit S passes through the thinly sil-

B A P

FIG. 4.

vered glass plate P, thence through the lens A, and is
reflected by a plane mirror B to the concave mirror C.
The radius of curvature of the mirror C is equal to
BC, so that the light is returned to B in the same path
that it traversed in going, and thus it again passes
through the lens A, and a part is reflected by the sil-
vered glass P and is observed at 0. If the mirror B
is revolved slowly, the light comes out at 0 as a
series of flashes. These become sensibly continu-
ous to the eye as the speed increases, but when the
speed becomes high the image at 0 becomes dis-

THE SUN'S DISTANCE

placed owing to the motion of rotation of the mirror
B while the light is passing from B to C and back to
B. From the amount of displacement, the speed of
the mirror, and the distance BC, the velocity of light
is computed.

According to the electromagnetic theory of light,
the ratio between the electrostatic and the electro-
magnetic systems of electrical units should also give
the velocity of light. Furthermore, the electric waves
used in wireless telegraphy should proceed with
the velocity of light. Results depending on these
last two considerations, though interesting, cannot
rival in accuracy the velocity directly determined.
The following are among the best results:

OBSERVER.

Method.

VELOCITY OF LIGHT IN

VACUO

Kilometers
per second.

Miles
per second.

Mean of Michelson, New- )
comb and others \
Mean of Cornu and Per- [
rotin f
Various observers
Rosa and Dorsey

Foucault

Fizeau

Hertz waves . .
Ratio of units .

299,860

299,890
299,130
299,710

186,330
186,350

Accepted velocity of light . .

299,860

186,330

The velocity of light just given is probably correct
within one part in ten thousand.

There are three ways of employing this quantity
to fix the distance of the sun. The first we will men-
tion is through the aberration of light. Though light
proceeds by wave motions, and not by particles, the

THE SUN

idea of aberration may be understood by the analogy
of raindrops. If rain is falling vertically, and a man
stands still, his hat screens his face. But if he move
rapidly forward in any direction, the rain strikes his
face, thus appearing not to come vertically but at an
angle thereto. So with the light of the stars: Owing
to the motion of the earth in its orbit the stars are
apparently displaced when the earth is moving at
right angles to the line of sight; the displacement
being in one direction at one time of the year and in
the opposite direction when the earth's motion is re-
versed six months afterwards. Owing to aberration,
stars at the poles of the ecliptic describe little circles
about 41" in diameter, and those in the plane of the
ecliptic merely oscillate in a straight line about 41 "
long. There is an uncertainty of a few hundredths
of a second as to the " constant of aberration," as
astronomers call the radius of the circle of aberration.
The Paris conference of 1896 adopted the value
20.47". But there is now much evidence tending to
recommend a higher value. The long-continued ob-
servations- of Doolittle made with instruments of
different kinds seem to require us to set the constant
of aberration as high as 20.51", perhaps even 20.53".
The following table shows the relation of aberration
and solar parallax values :

Aberration
Constant.

20.46"

20.47"

20.48"

20.49"

20.50"

20.51"

20.52"

20.53"

Solar
Parallax.

8.807"

8.803"

8.799"

8.794"

8.790"

8.786"

8.781"

8.777"

THE SUN'S DISTANCE

Another way to use the velocity of light for fixing
the sun's distance is through the observations of
Jupiter's satellites. Olaf Romer called attention to
this method by taking the problem the other way
about and computing the velocity of light from the
supposed known distance of the sun. The satellites
pass behind the planet and are eclipsed frequently.
These eclipses occur nearly 1,000 seconds later in
time when Jupiter is in conjunction than when in op-
position, for there is a difference of distance amount-
ing to the whole diameter of the earth's orbit for
light to traverse in the two cases. Unfortunately, the
eclipses are not sudden phenomena, so that it requires
careful photometric work to fix the " light equation,"
as the time required for light to travel the radius
of the earth's orbit is called. According to many
years of observation at the Harvard College Obserr
vatory, as reduced by Professor Samson of Durham,
the light equation is 498.64 seconds. From this the
sun's parallax comes out 8.799".

A third way of employing the velocity of light
is through what is known . as the Doppler effect.
Just as a locomotive whistle is higher in pitch when
the train approaches, so the light of a star is bluer
in hue when we are approaching the star in the earth's
orbit. The velocity of the earth is so small compared
to the velocity of light that the magnitude of the
change can be measured only with a powerful spec-
troscope of special design. Nevertheless, it seems
possible that solar parallax determinations by this
4 23

THE SUN

method may before long compare favorably in
accuracy with any others. For parallax purposes,
Kiistner, and lately Halm, have photographed the
spectra of bright stars in comparison with spark
spectra photographed above and below on the same
plate. They repeated these comparisons at inter-
vals of about six months for several years, arid, after
applying necessary corrections, they have determined
the velocity of the earth in its orbit, and from this
the sun's parallax. Their values are not far from
those obtained from other methods, but are not
quite accurate enough to compete with them.

It would seem more promising to determine the
relative velocities of the planets and the earth, by
photographing simultaneously the spectra of Venus
and Mars, or of Mars and the moon at a favorable
time. Suppose, for instance, that two large cce-
lostats (see Chapter II) were arranged one beside the
other to reflect the light of Mars and Venus simul-
taneously upon a single long-focus concave mirror,
and the two images were reflected together by suit-
able devices so as to fall at once one above the other
on the slit of a powerful spectroscope. By the use
of a rotating sector one image could be made equal
to the other in brightness, and the spectra of both
could then be photographed absolutely simultane-
ously, one above the other on the same plate.
Errors such as displacements by change of tempera-
ture of the spectroscope would effect both spectra
alike. The ccelostats should be used alternately

THE SUN'S DISTANCE

on the two objects, so as to follow out Professor
Turner's excellent motto of "reversing everything
that can be reversed." From preliminary trials
made at Mount Wilson it seemed to Mr. Adams and
the writer that it would be practicable to photo-
graph the two spectra on a scale ^ as large as that
which Mr. Adams employed to determine the solar
rotation spectroscopically. As the spectra of the
planets are similar, being solar spectra slightly
altered by selective reflection, there would be
numerous good lines to measure. There would
evidently be no necessity of introducing a terres-
trial comparison spectrum. It seems probable that
by this method the solar parallax could be determi-
nable to about one part in 2000. However, it has
not yet been tried.

(5) Summary.

Excluding parallax values not of the highest
weight, we have the following mean results:

From heliometer work on minor planets 8 . 807'

From the Eros campaign 8 . 807'

From all gravitational methods 8 . 780'

From eclipses of Jupiter's satellites1 8 . 799'

From the constant of aberration of light (assumed 20 . 51") ... 8 . 786'

If we take the mean of all these results as they stand,
we, in effect, give double weight to the geometrical
and velocity of light methods as compared < with the

1 This determination is included, notwithstanding the large probable error
Professor Samson assigns to it, because in the author's opinon, as indicated above,
the constant of aberration and the minor planet parallaxes are uncertain to nearly
the same degree, and the value of an independent method is very great.

THE SUN

gravitational method. This seems justified, and
by doing so we reach the probable value of the
solar parallax as

8.796".

This corresponds to a solar distance of

92,930,000 miles or 149,560,000 kilometers.

DIAMETER OF THE SUN

From the heliometer measurements of Schurr and
Ambronn the sun's angular diameter, as seen from
the earth at mean distance, is 1920.0//±0.03// Other
determinations agree very closely with this. Hence,
the sun's diameter is

865,000 miles or 1,392,000 kilometers.
Poor has lately maintained that observations indi-
cate that the sun's equatorial and polar diameters
vary relatively as much as 0.1" during a sun-spot
cycle of eleven years. According to him, the
equatorial diameter is the larger at sun-spot maxi-
mum, and the polar diameter the larger at minimum.
Ambronn, however, denies that this is supported
by the observations, and Moulton opposes so large a
variation on theoretical grounds.

THE SUN'S MASS

The mass of the sun relatively to a planet which has
a satellite may be obtained in several ways. One
of them, as applied to the earth, is as follows:
Let M be the mass of the sun, earth and moon

THE SUN'S DIMENSIONS

combined; and m, that of the earth and moon; let
R be the mean distance between the centers of
the sun and the earth; r, the mean distance between
centers of the earth and moon ; let T be the number
of days in a sidereal year, and t the number in a
sidereal month. Then, by Kepler's law:

R3 r3 R3 r3 r3

M:m=— :-; whence M-ra : ra = — --:—.

The mass of the moon compared with the earth is
known from other data to be 1/81.53. Small cor-
rections to the periods T and t due to the perturba-
tions are also known. Applying these corrections,
the ratio of the masses of the sun and earth comes
out. It is, according to Newcomb, for a parallax
of 8.796":

332,800.

THE EARTH'S MASS

Astronomy must be displaced by physics if we
would proceed further and obtain the mass of the sun
in ordinary units; for the mass of the earth is then
required. This is determined by comparing the at-
traction of the earth for a body (i.e., the weight of the
body) with the attraction which a known mass pro-
duces when acting upon the body from a known dis-
tance. During the eighteenth century attempts were
made to compare the attraction of a mountain with
that of the earth. The most celebrated of these was
performed in 1775, under the auspices of the Royal
Society, by the Astronomer Royal Maskelyne at the

THE SUN

mountain Schehallien in Scotland. Owing to the im-
possibility of determining accurately the center of
gravity and density of a mountain, this method,
though suggested by Newton, and one very interest-
ing to contemplate, is of little value. What is known
as the method of Cavendish, though proposed by
Michell, is regarded as best. In this method a pair
of small balls are hung at the ends of a rod which is
supported in the center by a fine wire or fiber. Two
large masses are placed in a position to twist the sus-
pending fiber by attracting the small balls. The
force of attraction is measured by the torsion of the
fiber, and this is determined by the period of vibra-
tion of the system. C. V. Boys performed a notable
piece of work by this method in 1894, and for it he
invented the quartz fiber, without which some of the
most delicate and interesting of modern physical
work in other lines would be impossible. His first
method of making quartz fibers was very picturesque.
One bit of a quartz crystal being fastened to a little
arrow, and another to a bow, the two bits are fused
together by a blowpipe flame; and when the quartz
is properly melted the arrow is shot out, trailing be-
hind it a thread of quartz almost too fine to be seen.
Such fibers are almost perfectly elastic, and as strong
as steel in proportion to their size. Boys obtained
for the mean density of the earth, 5.527. In other
words, the earth has five and a half times as great
mass as it would have if composed entirely of water.
The principle of a third method, more simple to im-

THE SUN'S DIMENSIONS

derstand than Cavendish's, is illustrated in Fig. 5.
Two equal balls, A and B, are suspended from the
beam E of an equal arm balance, and two large equal
balls are arranged so as to be in the positions C and
D, or C' and D', at pleasure. In the first position
they tend by their at-
tractions to make A
overbalance B, and in
the second position, the
opposite. Hence the
mere weighing of B
against A by means of
a rider on the beam E is
the principal require-
ment in addition to knowing the masses of the
balls, C and D, and their distances from A and
B. By this means Richarz and Krigar-Menzel
determined the earth's mean density to be 5.505.
Burgess has discussed the different determina-
tions, and gives the most probable value of the
earth's density as 5.5247 ± 0.0013. Corresponding
to this, the constant of gravitation (see the begin-

FIG. 5.

ning of this chapter) is 666.07 X 10~10

cm.

- (or

gr. sec.

dynes). From these figures the mass of the earth is
1.317 X 1025 pounds, or 5.984 X 1024 kilograms, and
that of the sun is 4.38 X 1030 pounds, or 1.990 X
1030 kilograms (that is 4, or 1, followed by 30 places
of zeros).

THE SUN

THE SUN'S DENSITY

Since the volume of the sun is 1,306,000 times that
of the earth, the density of the sun is only 0.255 as
great as that of the earth, and is 1.41 as compared
with water. A most interesting and important con-
clusion follows from these figures on the density of
the sun. Notwithstanding that its density is so
small, we know from its spectrum that the sun has
many of the heavy metals and other chemical ele-
ments found upon the earth, and we presume that it
includes few elements or compounds which in a liquid
or solid state would be of less density than water.
Water and other common liquids can not exist even
as vapors on the sun, owing to the high temperature.
In view of these facts, it follows that the sun is prob-
ably mainly gaseous. Owing to the enormous mass of
the sun, the attraction of gravitation at its surface is
27.6 times as great as that at the surface of the earth,
so that a body which weighed a hundred pounds here
would weigh over a ton there. Hence, the gases
of the interior of the sun must be tremendously com-
pressed, so that probably in their appearance they
would resemble liquids, though still having the prop-
erty of indefinite expansibility characteristic of gases.

CHAPTER II

THE INSTRUMENTS AND METHODS USED IN SOLAR
INVESTIGATION

The Telescope. — The Coelostat. — The Spectrum and what it In-
dicates.— 'Spectroscopes. — The Spectroheliograph. — The Helio-
micrometer. — The Comparator. — The Nature of Radiation. —
Laws of Radiation. — Spectra of Different Sources. — Pyrheliom-
etry. — Bolometry.

FOR a long time the telescope and the observer's
eye were the principal means of advancing solar in-
vestigation, but in the last half century a number of
other less familiar instruments and physical princi-
ples have been employed, which require some ex-
planation.

THE TELESCOPE

A few words may be said first as to the methods of
employing the telescope. The sun is far too bright
to view for any length of time with the naked eye,
much less with the telescope, unless means are used
for reducing the brightness. It is said that the Bel-
gian physicist, Plateau, having looked steadily at the
sun twenty seconds for the purpose of studying the
after images which would be produced, lost his sight
permanently in consequence.

To get a rough general view of the sun a screen is
often used in the manner shown in Fig. 6. The dis-

THE SUN

tance of the screen from the eyepiece depends on
the size of the image desired and the power of the
eyepiece. By moving the eyepiece to and fro in
the draw tube, a sharp image is readily obtained.
It is well to put a screen on the front of the tele-
scope, as shown in the figure, to cut off undesired
light. For Carrington's method of determining the
exact locations of sun spots on the disk, see Monthly

FIG. 6.

Notices of the Royal Astronomical Society, vol. xiv,
p. 153.

Observations of the finer details of the sun cannot
be made with a screen, and there are several ways
of protecting the eye for direct telescopic vision.
This may be done by reducing the aperture of the
telescope objective with a suitable diaphragm and
placing a dark glass in front of the eyepiece, but at

SOLAR INVESTIGATION

great cost of definition if the diaphragm is too small
or the shade glass not perfect. A reflecting tele-
scope, if its object mirror is left unsilvered, may re-
quire only a shade glass for visual work with the sun;
and, on the other hand, the objective lens of a re-
fractor may be thinly silvered to cut down the light.
But both these expedients unfit the telescope for
other purposes. There are several special solar
eyepieces which have
been devised. Fig. 7
shows Sir John Her-
schel's. The light en-
tering at 0 encounters a
prism of glass whose first
surface is placed at an
angle of 45°. More than
90 per cent of the light
passes through the prism
and goes out through
the open end of the tube, while the reflected light
goes up through the eyepiece AB. A shade glass is
still necessary, but need not be very dark. It is
advantageous to employ a long thin wedge of dark
glass ("London Smoke," for instance) compensated
by a corresponding wedge of ordinary glass, as
shown in Fig. 8.

With this arrangement the image is undistorted,
uncolored, and may be made of exactly the proper
brightness for observing. The polarizing eyepieces
on the general plan shown in Fig. 9 are more con-

FIG. 7.

THE SUN

venient, but also more expensive. The light is re-
duced merely by rotating the upper case in its

^^^^•^•^^••Bm^l^HH bearmg m the

lower. The image
is seen in its
proper color and
without being
either inverted
or reversed from
right to left

FIG. 8. ^,

by the eyepiece.

When very small objects are being examined, it is
sometimes advantageous to use Dawes's device of
limiting the field of view by means of a minute
diaphragm made by
piercing a card or plate
of ivory with a hot
needle.

For photographic
work the extreme
brightness of the sun
is advantageous in-
stead of troublesome,
because it enables the
observer to employ
slow plates which have
much finer grain than
rapid ones, and also to
cut down the exposure
time, which is an im-

SOLAR INVESTIGATION

portant gain, since it is favorable to making com-
plete exposures during the occasional instants of
exceptionally good atmospheric conditions which
favor superior optical definition or "good seeing."
As all solar observers know, the atmospheric effect
called " boiling" is generally much worse in the day-
time than in the night, and undoubtedly because
the powerful heating of the surface of the ground
by the solar rays
causes rising air
currents of un-
equal density,
which drift hither
and thither across
the line of sight.
Photographs of the
sun are usually ex-
posed by means of
a sliding shutter
similar in action to
the form shown in Fig. 10. B is a catch which
may be released by the electromagnet, or by hand,
thus allowing the spring S to draw the slide* con-
taining the slit A swiftly across the opening 0,
through which the rays enter the camera. The ex-
posure is proportional to the width of the slit
A, and is governed by the tension of the spring S.
1 1

FIG. 10.

Exposures of grknrk to 77^; of a second are required

The edges of the sun's

5000 100
according to circumstances.

TIIM SUN

disk HIT not us bright us the center, sn I hat the solar
image cannot be properly exposed to show details
equally well in all parts in the same picture.

It is often desirable to know the orientation of the
solar image. For this purpose a cord or wire may
be stretched close to the plate, in some known
position, as horizontal, or parallel to the sun's
drift, or vertical, and its shadow on the image will
serve as a basis of computation. Sometimes in an
optical system containing reflectors it is desired
to know what parts of the solar image correspond to
east and west in the sky. This may always be
determined with ease and certainty by stopping the
telescope motion and letting the sun's imago drift;
for the advancing edge or "limb" of the image must
always correspond to the west edge or "limb" of the
sun in the sky. The data for computing the position
of the sun's equator are published annually in the
pamphlet called The Companion to the Observa-
tory.

THE COSLOSTAT

Most modern apparatus for solar research is of a
complex and necessarily bulky nature, so that it is
highly inconvenient to move it. Accordingly, a fixed
beam of sunlight is almost a necessity. There are
several kinds of instruments for reflecting the light of
heavenly bodies in a fixed direction, called heliostats
or siderostats; but these all rotate the image of the
heavenly object if an image is formed. This is usu-
ally a great disadvantage, and, fortunately, there is

PLATE I

SMITHSONIAN OBSERVING SHELTER AND COXOSTAT, MT. WILSON.

SOLAR INVESTIGATION

one very simple instrument for the purpose, called the
coelostat, which does not rotate the image. In its
simplest form the ccelostat is a single plane mirror
mounted on an axis parallel to the axis of the earth,
and rotated by clockwork at the rate of one complete
rotation in forty-eight hours. In this form the sun's
beam is reflected in a different direction at different
times of the year, according to the position of the sun
north or south of the celestial equator. Even in a
single day the mirror cannot be used to throw a hori-
zontal beam in any desired direction, but only in two,
nearly east and west respectively, the first favorable
for morning hours, the other for afternoon hours.
These limitations are overcome by the introduction
of a second plane mirror, south of and above the level
of the first, on which the beam is first reflected, and
from which it can be sent in any desired direction,
but preferably (for the northern hemisphere) towards
the north. It is necessary to provide longitudinal
and cross motions for one or the other of the mirrors,
to accommodate the change in declination of the sun
at different times of the year. Plate I shows the
fifteen-inch coelostat of the Smithsonian Astrophysi-
cal Observatory at Mount Wilson, Cal. The first
or rotating mirror is provided with means of moving
it on tracks both east and west, and also north and
south. The second or fixed mirror reflects the beam
horizontally northward to the spectroscope within
the observatory.

THE SUN

THE SPECTRUM AND WHAT IT INDICATES

Since Kirchhoff and Bunsen's great discovery of
spectrum analysis in 1859, the spectroscope has be-
come more and more indispensable to progress in
solar research, so that now the greater portion of our
knowledge of the sun is due to this instrument. White
light is not a simple but a composite thing, containing
potentially all the different colors familiar to the eye,
and still other rays which the eye sees not at all. As
we shall describe, light can be analyzed so as to pre-
sent to the eye the colors, and this presentation is
usually in the form of a long ribbon of color gradation.
When light is thus analyzed to show the colors which
are potentially in it, the spectrum is said to be pro-
duced.

If sunlight is resolved into a spectrum, under good
conditions we see a ribbon of light shading gradually
from dull red through brighter and brighter hues to
orange, then yellow, next green, then blue, indigo and
violet. If we had eyes of unlimited capacity we
should see beyond the violet still other rays, and
beyond the red yet others, also. We can detect
such invisible rays by the heat they produce or by
photography, but just as the ear cannot hear sounds
above a certain pitch, or below a certain other pitch,
the eye is limited as to its recognition of radiation.
Rays lying beyond the violet end of the visible spec-
trum are called " ultra- violet" and those beyond
the red are called " infra-red." In the visible spec-

SOLAR INVESTIGATION

trum the shading is not perfectly continuous, for
there may be seen almost innumerable vacancies of
color, or dark lines crossing the colored ribbon at
right angles. These dark lines are called, after the
name of their discoverer, Fraunhofer lines. It is
their presence, and not the beautiful colors, which
has been the means of teaching us many things about
the sun and stars which would have seemed to the
contemporaries of the Herschels to be beyond the
possibilities of future discovery.

The cause of the dark lines of the spectrum was
unknown until Kirchhoff and Bunsen's researches,
about 1859, showed that they correspond in position
to certain bright lines which form the spectra of me-
tallic vapors. For example, if metallic sodium, or
any of its compounds like common table salt, is
thrust into an alcohol-lamp flame, the spectrum of
the flame shows two brilliant yellow lines which agree
in place with two prominent dark lines in the yellow
part of the solar spectrum. Not only so, but if an
incandescent oxy-hydrogen calcium light, whose nat-
ural spectrum shows neither bright nor dark lines in
the yellow at these places, is caused to shine through
an alcohol-lamp flame charged with sodium vapor and
placed before the spectroscope, the two dark lines
like those in the solar spectrum will appear instead of
the two bright lines of the sodium-charged flame it-
self. Other chemical elements, also, when heated to
vaporization, emit bright spectrum lines, and the
vapors of these elements, if placed in a beam of white
5 39

THE SUN

light, absorb the rays they themselves emit. If their
own emission is more intense than the emission they
absorb from such a transmitted beam, the effect will
be brighter lines in a continuously bright spectrum.
If their own emission is less intense than the emission
they absorb, the resulting spectrum will be crossed
by dark lines. The former effect occurs in the spec-
tra of certain stars, the latter in that of the sun. As
the emission of a vapor falls off rapidly as the temper-
ature falls, it is natural to suppose as the cause of the
sun's characteristic dark spectrum lines that the
metallic vapors of the sun's outer layers, since they
are free to loose heat to space, continue cooler than
the sun's inner layers, and hence cannot, by their own
emission, fully compensate and supply the place of
the rays they absorb.

In Fig. 1 of Plate IV a part of the spectrum of the
star Procyon is shown, with comparison spectra of
iron above and below it. The stellar spectrum shows
numerous dark lines for the reasons indicated above ;
and many of these correspond closely in position and
relative importance (or intensity, as it is called) to
bright lines in the iron spectrum. For a reason to be
explained below, the stellar lines are all shifted a little
towards the violet with respect to the comparison
spectrum, but it is evident that iron has left its sign
in the star's spectrum as well as in that of the electric
spark.

First of all, then, the dark lines of the solar and
stellar spectra show what chemical elements are pres-

SOLAR INVESTIGATION

ent in the sun and stars. By comparing the -solar
spectrum with bright line spectra of pure metals pro-
duced in the laboratory, it has been shown that
nearly all of the elements found in the sun are also
present in the earth. In the second place the lines
of the solar spectrum serve as reference marks to en-
able us to recognize the effect of certain influences in
the sun, such as varying degrees of temperature, of
velocity, of pressure, and of magnetism.

As regards temperature: A dark spectrum line
generally indicates a cooler vapor in front of a hotter
source, and a bright line that no hotter source lies
behind. Furthermore, many elements give in the
laboratory a large number of bright lines whose
relative intensities differ according to the tempera-
ture of the source. Similar differences of intensity
among the lines of a given element, as found in the
solar spectrum, give a basis for estimating differ-
ences of temperature there, as, for instance, between
a sun spot and the photosphere.

As regards velocity: We have noted the Dop-
pler effect already in speaking of methods of meas-
uring the sun's distance. It depends on the fact
that light travels by waves. Those waves which
are visible to the eye, differ in length from 0.0004
millimeter (0.4 ^) to 0.0007 millimeter (0.7 p), cor-
responding to violet and dull red respectively.
The time period of a complete vibration of a wave of
violet light is given by the ratio of its length (.0004
millimeter) to the velocity of light (300,000,000,000

THE SUN

millimeters per second) . Hence, 750,000,000,000,000
waves of violet light are emitted from all parts of
the sun's surface each second.1 Doppler's principle
is as follows: If a star is approaching the earth with
a velocity v, the effect is to shorten the length of each
wave of light reaching the earth by an amount vt,
where t is the period of vibration of the wave. If
V is the velocity of light and X the original wave-
length, X = Vt. Suppose the apparent wave-
length to be X,, then X, = (V- v)t. Hence, (X- X,)

If the lines of a spectrum are displaced towards the
violet by amounts which are less at the violet end
of the spectrum than at the red in the ratio of the
wave lengths, this may be an indication that the
source of light is approaching the earth. By com-
paring the positions of the spectrum lines at different
parts of the edge of the sun's disk, it has become pos-
sible to measure the rate of rotation of the sun for
all solar latitudes. Similar studies of the shifting
of lines in the spectra of the stars in all parts of the
heavens indicates towards which of the stars the
solar system is approaching in its motion through
space. In Fig. 1 of Plate IV the dark stellar iron
lines show a displacement towards the violet as com-
pared with the bright iron companion lines and thus
we find that Procyon was approaching the earth at
the time when the observation was made. In Fig.

1 For "additional remarks on this subject, see Chapter VII.
42

SOLAR INVESTIGATION

2 of Plate IV is shown a pair of superposed solar
spectra from the eastern and western edges of the
sun. The great oxygen band called B gives rise to
most of the lines in this spectral region, and as these
lines are terrestrial, not solar, they are not displaced
in the two spectra. But all the solar lines show a
displacement, due to the fact that one edge of the
sun is approaching, the other receding.

As regards pressure: The experiments of Hum-
phreys, Mohler and Jewell first showed that the
spectrum lines of different elements are shifted
towards the red by varying amounts, if the pressure
of the surroundings of the source is increased. These
pressure shifts are usually very minute, and they
follow a different law from shifts due to velocity.
Thus the examination of the solar spectrum can
indicate the range of pressure under which its
absorption lines were produced.

As regards magnetism: It was first shown by
Zeeman that a powerful magnetic field may split
an ordinary single spectrum line into several com-
ponents, differing in the character of the polar-
ization of their light waves. The polariscopic ex-
amination of double, or triple, or merely widened
spectrum lines may yield evidence as to the mag-
netism in the sun. It is from such study that Hale
has discovered the existence of magnetic fields in
sun spots. In most cases the lines appear separated
into two components when viewed along the lines
of force of the magnetic field, and into three com-

THE SUN

ponents when viewed at right angles to the lines of
force; but sometimes four, or six, or even more lines
are seen. In the simpler cases, first mentioned, the
doublet seen longitudinally consists of two circularly
polarized rays, one polarized left-, the other right-
handed. The triplet seen transversely consists of
three plane-polarized rays, of which the central one
occupies the same position as the line seen in the
absence of a magnetic field, and the two side rays
occupy the same positions as the two lines seen
longitudinally. The plane of polarization of the
central component is at right angles to that of the
two side components. Hence, the central compo-
nent may be extinguished by interposing a Nicol
prism in a certain position, while the two side com-
ponents may be extinguished and the central one
again seen when the Nicol is rotated 90°. In the
case of the doublet seen longitudinally, the two rays
may be transformed into plane-polarized light by
introducing a Fresnel rhomb. After this they may
be extinguished alternately by a Nicol prism rotated
90° at a time. In Plate II, Fig. 1 shows a part of the
ultra-violet spark spectrum of iron viewed trans-
versely to the lines of force of a magnetic field. Figs.
2 and 3 show the effects of interposing the Nicol
prism in two positions. Fig. 1 appears exactly the
same as the spectrum would if seen along the lines
of force without the Nicol prism, although, in fact,
the polarization is not the same. Most of the lines
in this spectral region are of the ordinary type, but

SOLAR INVESTIGATION

one is unaffected, while some are very complex. One
even has twelve components; but this will probably
not be discerned in the spectrum as reproduced.

Like the straws which show the way the wind
blows, or the hieroglyphics which hold the history
of ancient times, the lines of the spectrum yield to
painstaking, minute examination a wealth of knowl-
edge wholly unthought of by the careless, and, at
first glance, unknowable. Hardly anything in
science is more wonderful than the extent of knowl-
edge of the heavenly bodies, situated millions,
billions, and trillions of miles away, which has been
acquired through the spectrum.

THE SPECTROSCOPE

In common practice two very different pieces of
apparatus are employed to produce the spectrum,
namely the prism and the grating; but a device con-
taining either one of these, in suitable combina-
tion with its adjuncts needed to produce a spectrum,
is called a spectroscope. To understand something
of the action of the prism it is needful to know that
light travels with different velocities in different
substances, and that in general the different-colored
lights travel with different velocities in the same
medium. In a vacuum all kinds of light travel with
equal velocity, and in most gases, at ordinary
pressures, there is little difference for the different
colors, but with transparent liquids and solids the
difference is very considerable, as is shown in the

THE SUN

following little table. The numbers give the values
of the ratio

velocity of light in vacuo
velocity of light in medium'

TABLE II. — Velocity of light in vacuo, and its ratios to the velocity of
light in different media.

Colo

Violet

Blue

Green

Yellow

Red

Wav

e length1

4,200

4,800

5,400

5,900

6,500

Velo
un

city in vacu-
i2

299,860

Ratio of velocity in vacuum 1
1 to velocity in:

Air at atmos-
pheric pres-
sure

1.000297

1.000295

1.000294

1.000293

1.000292

Water

1.3420

1.3381

1.3354

1.3336

1.3320

Average flint
glass

1.6366

1.6238

1.6157

1.6108

1.6066

Carbon bi-
sulphide . . .

1.6835

1.6553

1.6374

1.6273

1.6188

Diamond

2.4570

2.4373

2.4242

2.4170

2.4108

In consequence of the difference of velocity of light
in different media, a beam of light of a single color is
bent when it crosses obliquely the boundary between
two media. This is illustrated in the accompanying
diagram, Fig. 11. The reader must assume, what is
shown by the methods of physical optics, that the
direction of propagation of light at each instant is at

'The unit of wave length is the Angstrom, which ia one ten-
billionth of a meter.

2 The unit of velocity is the kilometer per second.
46

SOLAR INVESTIGATION

right angles to the light front. A. beam of light,
which at a certain instant presents the front ac, begins
to enter a denser medium at c. In consequence of the
less velocity in the denser medium the light from c
moves only to cl}
while that from a
moves to a\t so that
the lower portion of
the light front is
thereby turned to the
direction 6^1. Suc-
cessive positions of

the light front are shown at a^Ci, a262C2, et cetera.
Suppose that the cross section of the denser medium
is triangular in shape, so that the light from a4 begins
to proceed more rapidly while that at c4 is still being
retarded. This state of affairs is that of the prism of
the usual form, and the light front finally emerges at
a7c7, proceeding in a different direction from that
which it had at first. The difference of direction de-
pends on the fractional difference of velocity of the
ray in the two media. Since this is greater for violet
light than for red, a beam of light containing both
colors will be split up by such an instrument, and the
violet part will be more bent or deviated from its
original direction than the red. Such action of the
prism is said to be " refraction, " the difference in
direction between the entering and emerging beams
is the "deviation," and the difference of direction
between the different colors as they emerge is called

THE SUN

" dispersion." The angle CcA between the entering
ray and the line perpendicular to the face of the prism
is called the angle of incidence, and the angle c7cB the
angle of refraction. A principal law of refraction is
this: The sine of the angle of incidence divided by
the sine of the angle of refraction is constant for a ray
of a single color entering a given substance, whatever
the angle at which it enters. Calling i the angle of
incidence, r the angle of refraction, and n the con-
stant, or index of refraction:

sin i

n = — — .
sinr

The value of the index of refraction is, therefore, a
matter of much importance for calculation, and it is
still more interesting because it is also the ratio of the
velocities of the light in the two media concerned.
For yellow light arid for ordinary telescope flint glass
and air as the two media, the refractive index is about
1.61 (see Table II).

It follows mathematically from the law just stated,
that for a given prism and a given color of light the
deviation can never fall below a certain minimum
value, whatever the angle of incidence. This small-
est angle of deviation is called the angle of minimum
deviation, and is secured when the angle of incidence
is equal to the angle of emergence. For this position
of the prism the following relation holds, if we desig-
nate the angles of incidence and deviation as i and
D, and the angle at the apex of the prism as A :
D=2i-A.
48

SOLAR INVESTIGATION

In the use of prismatic spectroscopes it is generally
preferable that the rays of light composing the beam
shall be parallel to one another as they enter the
prism, and that the prism shall be set for minimum
deviation. The beam which emerges when white
light passes the prism under these circumstances con-
sists of a mixture of bundles of parallel rays of light of
different colors, with the neighboring shades differing
by almost imperceptible inclinations one from an-
other in their paths of emergence. It is necessary to
bring them to focus by means of a lens or mirror if
they are to be sharply separated. If the light comes
originally from a star the rays will be practically par-
allel without alteration; but if they come from the
sun they converge from opposite sides of the solar
disk with an angle of over 30' of arc. For solar work,
therefore, and often for stellar work, also, two other
adjuncts to the prism are used. The first is a narrow
slit between sharp metal jaws parallel to the line of
intersection of the prism faces, and the second is a
lens or mirror placed at such a distance as to render
the rays which diverge from the slit parallel. A lens
or mirror serving the latter purpose is called a
collimator, and the lens or mirror which focuses the
spectrum is called the objective, or image-forming
piece. This arrangement is indicated in Fig. 12.

With most prisms the violet part of the spectrum is
much more extended than the red, owing to the more
rapid proportional variation of the velocity of light in
glass at the violet end of the spectrum.

THE SUN

The grating, as a means of dispersing light, de-
pends on the phenomenon called interference. Light,
like sound, is propagated by wave motions. If a
tuning fork be set in vibration and slowly rotated
while held in the hand at some distance from the ear,
the sound will be found to wax and wane in loudness,

FIG. 12

although the fork continues to vibrate steadily. The
position of faint sound occurs because the vibration
of air excited by one prong of the fork reaches the ear
so much later than that excited by the other that,
while one wave is in what corresponds to a crest, the
other is in what corresponds to a trough. At all
parts of the waves their effects are similarly opposed,
and the result, if the waves are equal in strength, is
silence.

With light a similar thing may occur. From two
slits, di and a2 (Fig. 13), imagine light of a single
color to proceed in all directions. Then at fcj, b2, etc.,
the waves may be supposed to arrive in opposition,
and thus to produce darkness, while at cl} c2, c3,
etc., there is light. From a great number of other

SOLAR INVESTIGATION

slits, a3 a4, etc., placed in a plane, and equally
spaced, the directions of light and darkness will be
the same, so that if a piece of plane glass is coated
with silver, and the silver coat is scratched off
in a series of parallel and equidistant lines, a bundle
of rays of light passing through the slits will some
of them proceed parallel to the direction AB, while
the others will be deviated, or diffracted, as it is
called, in various definite directions on either side of
the central beam. These directions depend upon the

FIG. 13.

interval between successive rulings, and on the length
of the wave of the given color of light. The devia-
tions are less for violet light than for the red, which
shows that the length of wave is less for violet rays.
Such a grating, as has just been mentioned, is
called a transmission grating, but it is more common
to employ a reflecting grating. A carefully ground
and polished surface of speculum metal is scratched
with a diamond point in parallel rulings very close to-
gether, not uncommonly as many as 20,000 to the

THE SUN

inch. One may suppose that with such close ruling,
the spaces between the ruled scratches are probably
like the rough ridges turned up by a plow, and, as
they would reflect but weakly, they may be assumed
to correspond to the opaque parts of the transmission
grating, while the smooth sides of the scratches act as
bright sources of light. To the late Prof. Henry A.
Rowland, of Baltimore, is due the principal share
of the credit for the great advance in knowledge of the
solar spectrum, and of that of the spectra of vaporized
substances, which has come in the last twenty-five
years; for he it was who designed the perfected screw,
and thereby was enabled to construct hitherto un-
equaled ruling machines. Rowland gratings, having
a total of as many as 60,000 lines or more, each two
or three inches long, are in nearly every large labo-
ratory and observatory of the world. Not only did
he thus promote the work of others, but his own em-
ployment of his gratings has left some branches of
solar spectroscopy at the furthest forward mark they
have yet reached.

Diffraction gratings may be ruled on flat surfaces
and used with a collimator and objective like a prism,
but many of them are ruled on concave surfaces, and
are used, after a design of Rowland, without collima-
tor or objective. Thus, we have the plane grating
and the concave grating spectroscopes. The arrange-
ment of the former is shown in Fig. 14. Frequently,
however, the collimator is used also as the image-
forming lens. Such an arrangement is called the

SOLAR INVESTIGATION

Littrow form of spectroscope. It may be employed
also, for prismatic instruments if a plane mirror is in-
troduced to return the beam through the prism. It
is necessary to tip the grating or mirror a little so
that the spectrum is formed above or below the slit.
Fig. 15 shows the concave grating arrangement.
In this figure, S is the slit, G the grating, and I the
spectrum. C is a rigid bar which carries the grating

FIG. 14.

and the means of observing the spectrum. This bar,
mounted on carriages, k and k' ', slides over the tracks,
R and R'. These tracks are placed at right angles,
with their intersection at S.

From a white light a grating spectroscope produces
a series of spectra more and more diverging on either
side of a single white band in the center. These spec-
tra are called first, second and third order, et cetera,
according to their divergence. Only one spectrum can
be made use of at a time among all this multitude,
and the greater the number of the order, the greater
the dispersion of the spectrum. The higher orders
overlap, so that the red of one order falls on the violet
or some other color of the next higher order. When it

THE SUN

is necessary to separate entirely one color from the
other, it is customary to interpose somewhere in the
beam an absorbing screen which is opaque to the
color not desired, but transparent to the other. Spec-

FIG. 15.

tra of very high orders, however, are hopelessly mixed
and are practically white light. It is seldom that
spectra above the fourth order are employed. The
relative brightness of grating spectra depends on the

SOLAR INVESTIGATION

form of the grooves ruled. Some diamond points
produce gratings very bright in one or two particular
spectra, and are preferred for this reason. The selec-
tion of a good diamond point is the result usually
of trial rather than of microscopic examination. A
spectrum may be very bright for some colors and not
for others. At best, a grating seldom throws as
much as one-tenth of the light into one spectrum,
and, therefore, in researches where loss of light is very
serious a prism is often preferred, since it may trans-
mit as much as eighty-five per cent. In a prismatic
spectrum the violet is greatly extended as compared
with the red, while in a concave grating spectrum the
dispersion is a linear function of the wave length.
That is to say, equal distances along the concave-
grating spectrum correspond to equal differences of
wave length. Such spectra are said to be "normal."
A plane-grating spectrum is nearly normal for short
distances.

The wave lengths in the spectrum, as visible to the
eye, range from about 0.39/*1 to 0.80/*. Beyond the
violet the solar spectrum runs to a wave length of
0.29/A, where it is practically cut off, partly by the
nontransparency of our own atmosphere (particu-
larly the nontransparency of ozone) and perhaps
imperatively by the opaqueness of the solar envelope.
Beyond the red the solar spectrum extends to a wave
length of about 20/*, though with several long inter-

irThe micron, or thousandth of a millimeter, is denoted by the
Greek letter j*.

6 55

THE SUN

missions due to the nontransparency of the atmos-
phere (especially of water vapor, carbonic acid and
ozone), from which cause it practically ceases at
20fju. Ordinary glass apparatus ceases to be trans-
parent at about wave length 0.35/* in the ultra-violet,
and at about 2.5/* in the infra-red, but the limits
differ with different kinds of glass. Quartz apparatus
is transparent to rays of all wave lengths from less
than 0.20/I,1 to more than 4.0u. Fluorite is transparent
in the ultra-violet, and in the infra-red its transpar-
ency extends to about 7.0/i. Rock-salt is also trans-
parent in the ultra-violet, and as far as 17/* in the
infra-red. Silvered glass mirrors reflect almost totally
for all rays of the infra-red and visible spectrum, and
their reflecting power remains high as far as wave
length 0.33/z in the ultra-violet. Between wave
lengths 0.33/x and 0.29/*, the reflecting power of silver
does not reach fifteen per cent. Speculum metal,
which is used for gratings, reflects much less strongly
than silver in the visible spectrum, but continues to
reflect forty per cent or more to beyond wave length
0.30/i.

As stated above, it is the minute study of the lines
found in spectra which yields many of the most inter-
esting results, and in the solar spectrum these lines
become increasingly numerous towards the violet,
and in the ultra-violet. Fortunately, the ordinary

1 Although solar rays of less wave length than 0.29^ are not found,
terrestrial sources give rays of much shorter wave lengths, even to
O.lOn

SOLAR INVESTIGATION

photographic plate is highly sensitive in this thickly
lined violet and ultra-violet part of the spectrum, and
at present most spectrum investigations are made
photographically. There are special photographic
plates which are sensitive in other parts of the spec-
trum. By staining ordinary plates with certain dyes,
they may be employed for red rays, and even a little
beyond the visible limit of the red spectrum. For
spectrum investigations far beyond the red, it is
necessary to use sensitive heat measuring apparatus,
such as will soon be described.

For some purposes it is sufficient to allow the rays
of the sun to shine directly into the spectroscope,
but ordinarily it is necessary to confine the obser-
vations to selected areas of the sun such as a sun
spot, or to the sun's edge or "limb" as distinguished
from the center. To do this the slit of the spec-
troscope must be placed in the focus of a lens or con-
cave mirror which forms a solar image of suitable
dimensions for the investigation. When the spec-
troscope is large, and the work requires it to be main-
tained at perfectly constant temperature for long
photographic exposures, it becomes highly desirable
to keep the spectroscope fixed and to employ a
ccelostat to reflect light to the lens or mirror. Fig.
16 shows the new 150-foot tower telescope, with a pit
75 feet deep beneath for the spectroscope, as just
being completed at the Mount Wilson Solar Obser-
vatory. A smaller tower telescope has been doing
good work there for a considerable time. The

THE SUN

coelostat is on the top of the tower 60 feet high, and
reflects a beam of sunlight vertically downwards

through a lens which forms
a solar image over 7 inches
in diameter upon the slit of
the spectroscope near the
surface of the ground. The
slit is in the center of a turn-
table which supports, by
rigid steel construction, the
collimator and plane grating
30 feet below ground. Thus,
the whole spectroscope can
be rotated about the axis of
the beam of light. The col-
limating lens acts, also, as
an image-forming lens (the
Littrow type of spectro-
scope), and the spectrum
falls on the photographic
plate fixed upon the surface
of the turntable near the slit.
Below ground the tempera-
ture is very constant. At
the top of the tower, the air
is nearly free from the trem-
ors which cause " boiling" of
the image. As the beam de-
scends vertically from the
top of the tower it is less
58

FIG. 16.

SOLAR INVESTIGATION

likely to be distorted by " boiling" than it would be if
coming obliquely, as from the sun directly. Hence,
altogether, the tower plan of solar observatory is
highly favorable for carrying on exact investigations
with powerful apparatus. The new tower telescope
of over 150 feet focus just being erected for the Mount
Wilson Solar Observatory will doubtless yield very
remarkable results.

THE SPECTROHELIOGRAPH

The spectroheliograph, invented by Dr. G. E.
Hale, is a device for photographing the sun in the
light of a single wave length. Let us suppose that
the solar image is brought to focus on the slit of a
spectroscope, and that the slit is longer than the
diameter of the image. The spectroscope may be
adjusted so that a certain Fraunhofer line, perhaps
the line called C, or otherwise Ha (due to hydrogen),
falls in the center of the field of view. Then, if the
solar image is allowed to drift across the slit, the
observer will see the masses of hydrogen on the sun
which emit the light in question as their images pass
in succession over the slit. But it would be practically
impossible to note and remember or sketch these
details. If the photographic plate is substituted
for the eye, and a slit placed just in front of it, so
narrow as to permit only the Ha line to pass, a
photographic record would be made, but this would
be a mixture of all the successive views of the hydro-
gen masses, and would be useless. But by moving

THE SUN

the plate along at the same rate that the image of
the sun drifts, there would be a new part exposed for
every succeeding impression, and the result would be
a photograph of the hydrogen masses which emit
Ha light, as they exist over the whole sun's disk.
This is one form of the spectroheliograph. In
another form, which is employed for the five-foot
spectroheliograph of the Snow telescope of the
Mount Wilson Solar Observatory, the whole spec-
troscope is floated on mercury, and moved slowly, at
right angles to the beam, across the sun's image and
the photographic plate, both of which remain station-
ary. The solar image of the Snow telescope is about
7 inches in diameter, and, if the correspondingly long
slit of the spectroscope were straight, the spectrum
lines would be greatly curved, and the sun's image
taken with the spectroheliograph would be distorted.
This defect is avoided by using curved slits, dividing
the necessary curvature between that for the spectro-
scope and that in front of the plate. The curvature
of these slits differs for different spectrum lines, so
that as many pairs of slits are required as there are
spectrum lines in which spectroheliograms are de-
sired. Thus far Ha, H/3, H7, HS of hydrogen, H and
K of calcium, and a few preliminary tests of other
lines have been tried.

THE HELIOMICROMETER

It ordinarily requires considerable measurement
and calculation to determine the positions of objects

SOLAR INVESTIGATION

with reference to the solar equator, seen on the solar
photographs, whether direct or spectroheliographic.
This labor is largely avoided by the use of a device
of Mr. Hale's called the heliomicrometer. It con-
sists of a sphere marked with circles of latitude and
longitude, and adjusted so that its poles correspond
in position with those of the sun for the date in
question. A long focus concave mirror throws an
image of this sphere, and of the photographic plate
to be examined, simultaneously into a double-field
eyepiece. Thereby, the two images are superposed,
and the observer sees the solar photograph appar-
ently marked with lines of latitude and longitude
corresponding with those of the sun. A micrometer
is provided for accurately measuring the distances
of the images of the solar objects from the nearest
reference lines.

THE COMPARATOR

In all photographic spectrum work, the main
thing is accurate measurements of the positions of
the spectrum lines with reference to each other,
or with reference to certain standards of position.
In many cases the slit of the spectroscope is partly
covered by a diaphragm of peculiar shape, which
can be moved so as to uncover different portions of
the slit. Thus, successive exposures may be made
to different sources of light, as, for instance, the
center and limb of the sun, or the sun and the
iron arc light. In the resulting photograph there

THE SUN

are several spectra corresponding to these different
sources, all accurately aligned one above another.
For measurement, the photograph is placed on the
table of a measuring machine, or comparator, and
this table is moved to and fro by an accurate screw
with graduated head, thus bringing chosen spectrum
lines to the cross hair of the observing microscope.
Measurements of position to the ten-thousandth

part of a millimeter (to ^,7^ inch) are some-

times made in this manner.

The wave lengths of the solar spectrum lines and
of the bright spectrum lines of the chemical elements
are the fundamental data of spectroscopy. In Row-
land's great table of the solar spectrum the wave
lengths are given to seven places of significant figures,
that is, to thousandths of an " Angstrom unit." It
has lately been found that there are certain syste-
matic errors of the table due to various causes,
chiefly to an obscure source of error in the use of the
grating for determining wave lengths, so that there
are corrections of the order of one or two hundredths
of an Angstrom to be applied to make Rowland's
table homogeneous. To reduce to the absolute
scale of the international metric system, a some-
what larger correction is needed. By means of the
interferometer these corrections are gradually being
determined, and it is probable that within a few
years we shall have a standard table of solar and
terrestrial spectrum places accurate to within two or

SOLAR INVESTIGATION

three units in the seventh place of significant figures.
It seems extraordinary enough that so small a quan-
tity as the wave length of light should be measurable
to such extreme precision, and still more extraor-
dinary that such a decree of accuracy is at all
necessary for promoting investigation. But so it
is, and much of the remarkable progress of solar
knowledge in recent years depends on differences of
wave lengths, as in the case of pressure and velocity
shifts, not larger than 0.005 of an Angstrom, or less
than one-millionth part of the wave length of yellow
light.

THE NATURE OF RADIATION

Not less important, perhaps, than these questions
of exact wave lengths, is the measurement of the
intensity of light, or rather, speaking more broadly,
of radiation. All solar rays, whether visible or
photographically active or not, produce heat when
absorbed upon a blackened surface. Sometimes the
infra-red rays are called "heat rays/' the light rays,
"visible rays," and the blue, violet and ultra-violet,
" actinic," or " photographic rays." But there is no
distinction of kind between these things. All are
regarded as transverse vibrations of the luminifer-
ous ether, differing only in wave length. Just as
there are sound waves too high or too low in pitch
to be heard, so radiation may be too long or too short
in wave length to be seen, but this implies no dif-
ference in kind of vibration.

The intensity of radiation can be quantitatively
63

THE SUN

estimated only very imperfectly by the eye, or by
the aid of the photographic plate, although both the
eye and the plate are excessively sensitive to radia-
tion of certain wave lengths. But waves of all
wave lengths produce their just effects when trans-
formed into heat. Though both are forms of energy,
radiation is not heat, but may be transformed com-
pletely into heat. We regard radiation as wave mo-
tion in the ether, heat as irregular motion of the
molecules of material substances. All heated sub-
stances give off radiation; but the amount and quality
of radiation given off at a given temperature are
different for different substances. Substances at
any temperature above the absolute zero ( — 273° C.)
are supposed to consist of molecules in rapid mo-
tion. These moving molecules may be supposed
to communicate some of their energy to the unseen
ether which is assumed to permeate all space, even
the interstices between the molecules of solid bodies.
Thereby the ether may be assumed to be set in con-
fused vibration, and from this confusion is extri-
cated by the prism, or grating, the orderly succession
of wave lengths which we term the spectrum. The
relative intensity of the several parts of such a
spectrum depends on the temperature of the exciting
body.

Kirchhoff introduced the notion of the perfect
radiator. This is sometimes called "the absolutely
black body," because a perfect radiator is a per-
fect absorber of radiation, and most black substances

SOLAR INVESTIGATION

are also nearly perfect absorbers. The perfect radi-
ator emits for a given temperature the maximum
possible amount of radiation of each and every
wave length; so that no other body at the same
temperature can excel its emission for any wave
length.1

LAWS OF RADIATION

Kirchhoff proved the following important rela-
tion, now known as Kirchhoff 's law: For any given
temperature and wave length the ratio of the emis-
sion of a body to its absorption is a constant, and
equal to the emission of a perfect radiator for the
same temperature and wave length. In order to
understand this law, the force of the expressions
emission and absorption must be clearly grasped.
By emission is meant the rate of escape of energy by
radiation, and to fix ideas it may be regarded as the
amount radiated from each square centimeter of sur-
face in a minute of time. By absorption is meant
the fraction which would be absorbed in the body
if shined upon by radiation from another source.
For instance, if thus shined upon, and three-fourths
of the rays received are absorbed and go to warm
the body, while the other fourth is reflected away,
or transmitted, the absorption is said to be three-
fourths. Such a body, by KirchhofTs law, would
emit only three-fourths as copiously, for the wave

1 An exception must be made, perhaps, of a certain class of bodies
excited to radiation by other causes than temperature, as, for in-
stance, chemical action. The remarks above concern the relations
of temperature and radiation alone.

THE SUN

length and temperature in question, as would the
perfect radiator.

The importance of the conception of the perfect
radiator will appear as we go on. No substance in
the world answers to its requirements, but lamp-
black is very nearly a perfect radiator at low tem-
peratures. However, if a closed hollow chamber is
formed of any substance whatever, and its walls
maintained at uniform temperature, the radiation in-
side the chamber will be that of the perfect radiator.
If a small hole be made in the wall the radiation
which escapes through the hole will be practically
perfect radiation. Instruments of this form have
been constructed within the last fifteen years, and
careful measurements have been made of the inten-
sity of their emission for a great range of wave
lengths, and for temperatures from that of liquid air
up to that of melting platinum. These results have
been compared with the theoretical radiation formu-
lae connecting temperature, wave length and radia-
tion which have been proposed.

The formula of Wien, as modified by Planck, is
found to express the observed results. Let e be the
emission of wave length X by the perfect radiator of
the temperature T, and let e be the base of the Napier-
ian system of logarithms, and let c± and c2 be two con-
stants determined by experiment. Then:

e = cA~5(e AT - 1 )-! (The Wieri-Planck formula) . I

As stated above, no body emitting rays by virtue of

(5(1

SOLAR INVESTIGATION

temperature can exceed the radiation determined by
this formula for any wave length or temperature.

Another formula of nearly equal importance, due
to Stefan, gives the measure of the sum total of radi-
ation, E, of all wave lengths, for a perfect radiator of
the absolute temperature, T. It is this:

E = o-T4 (Stefan's formula). II

The quantity, cr, is a constant determined by experi-
ment.

A third formula, called Wien's displacement law,
connecting the wave length of maximum emission,
^max. (expressed in thousandths of a millimeter, or /*) ,
with the absolute temperature T is as follows:

Vax. T ~ 293° (Wien's displacement formula). Ill
It is from these three formulae that we are able to
obtain some definite ideas of the minimum tempera-
ture of the sun. Many bodies appear to approach
the state of being perfect radiators at high tempera-
tures, although departing greatly from it at low tem-
peratures. But no body radiating by virtue of its
temperature can excel, either in the sum total of its
radiation, or in that of any wave length, the emission
of the perfect radiator of the same temperature.
Hence, if we can determine by Formula II the temper-
ature which the perfect radiator would have in order
that its radiation should approximate in quantity
the emission of the sun, then it is sure that the solar
temperature must be as high or higher.

Before giving the values of the constants in these
67

THE SUN

formulae, we must consider how energy of radiation
can be measured. There are no accurate means of
measuring radiant energy while it remains such. It
must first be transformed into heat. The unit of
measurement of heat is the calorie, or that amount of
heat which is required to warm one gram of water at
15° C. through one degree. With this unit we must
combine the notion of intensity. We then define
the unit intensity of radiant energy as that which,
if completely absorbed by a surface at right angles
to the beam, will produce one calorie of heat per
square centimeter per minute. We therefore meas-
ure radiation in calories per square centimeter per
minute.

To suit this definition, and to correspond with
wave lengths expressed in microns (/*), and tem-
peratures in absolute degrees of the Centigrade scale,
the values of the constants of formulae I and II are as
follows :

c,= 5.29 X 105; C, = 14,550; a- = 76.8 X 1Q-12.

SPECTRA OF DIFFERENT SOURCES

In Fig. 17 the curves A and B give the distribution
of radiation in the spectrum of a perfect radiator at
7000° and 6200° of the absolute Centigrade tempera-
ture scale, as computed from the Formula I. The
curve C gives the distribution of radiation as it would
be found in the average spectrum of the sun's entire
disk, if it could be observed outside of our atmos-
phere, according to determinations made by Smith-

SOLAR INVESTIGATION

a: AH3Sso

rfg'

at/W'XVx*

sonian expeditions on the summits of Mount Wilson
and Mount Whitney. The wave lengths are given
by the horizontal distances (abscissae) and are in
thousandths of a millimeter, or microns, usually de-

THE SUN

noted by the Greek letter /*. The visible spectrum
practically extends from 0.4/* to 0.7/*1, so that much
of the solar radiation is invisible. The vertical heights
of the curves (ordinates) are proportional to the energy
of the rays of corresponding wave lengths as meas-
ured by their heating effects. It will be noted that
the forms of the computed and observed curves differ
most in the ultra-violet, where the observed solar
radiation falls off more rapidly than the computed
radiation of the perfect radiator. Further remarks
on the subject of the sun's temperature will be given
in the next chapter.

The reader will note that the maximum ordinate of
curve A occurs at a less wave length than that for
curve B, and that curve A is at all points the higher
of the two. The perfect radiator is supposed to emit
rays of all wave lengths at all temperatures, whether
high or low; but when the temperature is low the
shorter wave lengths, including those which would be
visible, are too weak to be detected, even by such a
highly sensitive organ as the eye. As the tempera-
ture increases the intensities of rays of all wave
lengths increase, but the intensities of rays of shorter
wave length increase most rapidly. Hence, as ex-
pressed in Formula III, the wave length of the
maximum emission grows less and shifts towards the
violet end of the spectrum as the temperature in-

1 By special devices, the spectrum can be observed visually from
0.37 /u to 0.83 /A, but as ordinarily observed it falls within the limits
above stated.

SOLAR INVESTIGATION

creases. Most common solids and liquids emit a
continuous spectrum, which, as the temperature in-
creases, grows in intensity more rapidly for short
wave lengths than for long. But there are usually
special regions, or bands of the spectra of solids and
liquids where the radiation is stronger than that of
the adjacent wave lengths. These are called regions
of " selective emission," and, as follows from Kirch-
hoff's law, they are also regions of " selective absorp-
tion."

When gases or vapors are examined under ordinary
conditions of low pressure, and with small quantity
present, as when the electric arc is caused to play
between metallic poles, the spectrum appears to be
made up chiefly of narrow lines or bands of selective
emission, without a prominent accompanying con-
tinuous spectrum. Some authors hold that the con-
tinuous background is totally absent in gaseous spec-
tra, but it seems more likely that there is, in fact, a
very slight vestige of it present, which, if the quan-
tity of gas was increased, so that the observer
could look towards immense thicknesses, would be
increased until the emission for -all wave lengths
would finally approach the intensity of a perfect radi-
ator. This view is supported by the circumstance
that, if the pressure upon the emitting gas is in-
creased to several atmospheres, the spectrum lines
widen out, till at length there is, for some distance
from the lines, a perceptible continuous background.
Whether or not, then, it be true that gases under
7 71

THE SUN

less than atmospheric pressure would give continu-
ous spectra if in great depth, it is certainly highly
probable that such gases would do so if more and
more compressed with the increasing thickness. Re-
gions of strong emission are regions of strong absorp-
tion by Kirchhoff's law, so that in the case of a thick
gas, as just proposed, it would be only the front
layers which would give rise to the lines or bands of
high selective absorption, while the deep-lying layers
would be those which would produce the continuous
spectrum. If the gas is not of uniform temperature,
but grows hotter with increasing thickness, it is easy
to see that the continuous spectrum might exceed
the line spectrum in its intensity, so that the really
bright lines would appear dark by contrast with the
background. As is well known, the solar spectrum
has the character of a continuous bright ground
crossed by darker lines, and evidence will be pre-
sented later which indicates that it is indeed to be
regarded as a gaseous spectrum of the kind just
described.

PYRHELIOMETRY

In the year 1838 Pouillet devised the instru-
ment which he called the pyrheliometer, shown in
Fig. 18, and used it for measuring the intensity of
the sun's radiation. A flat silver-plated vessel a&,
blackened with lampblack on its upper surface, is
filled with water, and contains also the bulb of the
thermometer d. The instrument is held in the

SOLAR INVESTIGATION

clamp c, and pointed towards the sun as indicated
when the shadow of the box ab falls centrally on the
plate ee. By rotating the whole apparatus in the
clamp c, the water can, in effect, be stirred to equalize
its temperature. To observe the
intensity of the solar radiation
the instrument is first shaded,
and the change of temperature
occurring in a certain time, as,
for instance, five minutes, is
noted. Then the screen is re-
moved, and the observer notes
the change -of temperature due
to the sun's heating in the same
time. Finally the shade obser-
vation is repeated. Correcting
the average rate of rise of tem-
perature per minute during the
sun exposure by the average
rate of cooling, shown by the
shade readings, the result gives
the rise of temperature per min-
ute of a mass of water and cop-
per, of known heat capacity, due to the sun's rays
shining at right angles and absorbed on the known
area of the top of the box. A correction of about
2.5 per cent must be added on account of loss by
reflection from the lampblack.

Pouillet observed the intensity of the sun's rays
with this instrument at different hours of the day.

FIG. 18.

THE SUN

The atmosphere weakens the sun rays by the diffuse
reflection of its molecules and dust particles. This
effect is more and more apparent as the sun nears
the horizon. The atmosphere extends upwards for
a great distance, but becomes less and less dense,
so that at one hundred miles elevation what re-
mains above is negligible, so far as cutting off the
sun's rays is concerned. Hence, we may regard the
effective part of the atmosphere as a layer whose
thickness is very small compared 'to the earth's ra-
dius; and so, whenever the sun is 15° or more above
the horizon, the length of path of its rays in air is in
proportion to the length of the path when the sun is
in the zenith simply as the secant of the zenith dis-
tance at the time of the observation.

Bouguer and Lambert had shown independently,
in the year 1760, that when a ray traverses a homo-
geneous transparent medium, the intensity, E, after
traversing any given thickness, t, of the medium is
given by the following formula, in which E0 is the
original intensity, and a is a constant which repre-
sents the proportion transmitted by unit thickness:

E == E0a<.

Pouillet applied Bouguer 's formula to his observa-
tions, taking unit thickness as that traversed by
rays when the sun is in the zenith, so that if z is the
zenith distance, the formula becomes:

E = E0asecantz.

He computed the value E0, which is the intensity of
the sun's radiation outside the atmosphere, and, re-

SOLAR INVESTIGATION

ducing to mean distance of the sun1, obtained E0
=. 1.76 calories per square centimeter per minute.
This value Radau, and, also, Langley afterwards
showed must be below the true value of the " solar
constant of radiation" because Pouillet made no
spectrum observations, and it is necessary to do so on
account of the unequal losses suffered by rays of dif-
ferent wave lengths in passing through the air.

Pouillet's pyrheliometer was improved by Tyndall,
who substituted an iron box containing mercury in
place of the copper box containing water. In re-
cent years Tyndall' s design has been improved at
the Smithsonian Institution. First, a copper box
rilled with mercury was employed; then a copper
disk with a hole drilled radially to contain the cylin-
drical bulb of a thermometer with, also, a little mer-
cury surrounding -it to make good heat connection;
now (1910), the Institution uses a blackened silver
disk (shown in section at a in Fig. 19) with a radial
hole lined by a thin steel thimble. In this is inserted
in mercury a cylindrical-bulb thermometer, 6, bent
at right angles so as to point towards the sun when in
use. The disk is enclosed in a brass-walled, black-
ened chamber, c, and this is protected from changes
of temperature by a wooden wall, d, outside. The
sun's rays are admitted through a tube, e (shown
partly in section) , which contains diaphragms, / / /,

1 The sun's radiation varies in its intensity inversely as the
square of the sun's distance. Hence the earth receives on this
account nearly 7 per cent more solar radiation in January than in
July.

THE SUN

FIQ. 19. — SILVER DISK PYRHELIOMETER.

SOLAR INVESTIGATION

to prevent air currents from reaching the silver disk.
An equatorial mounting enables the observer to
point the instrument towards the sun. Several in-
struments of the type shown in Fig. 19 have been

constructed and com-
pared with those at the
Institution, and sent to

FIG. 20 — ANGSTROM'S PYRHE-
LIOMETER.

FIG. 21.

different solar observers abroad to convey to them
exactly the scale of measurements employed here.
In 1896 K. Angstrom devised his electrical com-
pensation pyrheliometer, which has been used very
extensively. Fig. 20 gives a general view of the
instrument and Fig. 21 an enlarged detail view of
the interior. It consists of two thin strips of man-
ganin, U U, of measured area, which are blackened on
the front surfaces, and have fixed to the rear of each
a thermoelectric junction for determining their tem-
peratures. The binding posts, Kx K2, communicate
respectively to the strips and the thermal junc-
tions. A measured current of electricity is passed
through one strip, while the other is exposed to the
sun, and when a galvanometer connected with the

THE SUN

thermal junctions indicates equality of temperature
it is assumed that the known amount of heat in-
troduced by the electrical current is equal to that
absorbed from the sun's rays. By reversing the
screen, W, and the commutator, C, the two strips
are heated alternately by the sun and by electricity,
and the mean result is employed. After applying a
correction for loss by reflection, the results are com-
puted in terms of calories per square centimeter
per minute. The instrument is inclosed in a dia-
phragmed tube R, and is mounted on an alt-azimuth
stand provided with the screws, Si S2, for following
the sun. A thermometer, T, indicates the tempera-
ture of the strips.

In both forms of pyrheliometers described above,
if used as standard instruments, a correction
must be determined and applied to allow for the
radiation reflected. Besides this, there is another
source of loss, arising from the fact that part of the
heat produced by the absorption of solar radiation
in lampblack is carried off by the air, and by re-
radiation of great wrave length, and this part does
not produce any effect on the thermometer or ther-
moelectric junction.

To avoid these sources of error other forms of
pyrheliometers have been devised in which the rays
are absorbed within a hollow cylindrical blackened
chamber. Such a chamber, as stated a few pages
above, is practically a perfect radiator, and hence
is a perfect absorber, so that no correction for rays

SOLAR INVESTIGATION

^reflected is needed. The
rays are principally ab-
sorbed at the rear end,
and, as the tube is deep,
the heat tending to es-
cape will be absorbed
somewhere on the side
walls. Two means of
using the hollow cham-
ber have been employed,
the first about 1894, by
W. A. Michelson, the
second, 1905 to 1910, by
the writer. Michelson
surrounds the chamber
by melting ice and water,
and determines the heat
introduced by measuring
the contraction of the
ice as it melts.

In the form devised by
the writer, as shown in
Fig. 22, a measured
stream of water, enter-
ing at E and emerging
at F, flows continually
in a spiral channel round
the walls of the blackened
chamber, A A, carrying
off the heat as fast as

THE SUN

formed. The rise of temperature of the stream of
water due to the solar heating (admitted through the
vestibule, B B, and the measured diaphragm, C) is de-
termined by a differential electrical thermometer com-
posed of four fine platinum wires wound longitudi-
nally on ivory spirals. These wires are bathed by the
stream of water which follows the spiral channels of the
ivory. Two coils are situated at Db in the entering
stream of water, and two at D2, after its passage
through the walls of the chamber. The four are joined
to form a Wheatstone's bridge, and their indications
are read by a sensitive galvanometer. The pyrheliom-
eter is protected from outside temperature changes
by the Dewar vacuum flask, K K. In order to test
the accuracy of the instrument two coils of man-
ganin wire, G and H, are placed within the chamber
near its rear, and a known quantity of heat may be
produced there in either coil by the passage of
a measured current of electricity. This heat is then
measured just as if it were from the sun, and if all
that is introduced is found, it may be supposed that
the instrument is a correct recorder of solar radia-
tion, especially as the coil G is very unfavorably sit-
uated for giving up its heat to the walls.

Two such water-flow pyrheliometers of different
dimensions were tested at Washington in 1910 and
gave closely agreeing results on solar radiation, be-
sides recovering almost completely the electrically
developed heat used as a test. These water-flow
pyrheliometers are used as standards, and the read-

SOLAR INVESTIGATION

ings of the silver-disk pyrheliometers are reduced
to the scale they give. The water-flow pyrheliometer,
when in use, is mounted equatorially and driven by
clockwork to follow the sun. It is alternately
shaded and exposed to solar radiation.

BOLOMETRY

For measuring the intensity of the rays in the
solar spectrum, the instrument most used is the
bolometer, a delicate electrical thermometer, in-
vented by Langley about 1880. As now construct-
ed, it comprises two exactly
similar, narrow, blackened,
platinum strips hardly as
wide as hairs, ten times
thinner than they are wide,
and about half an inch
long. Referring to Fig. 23,
such strips, a, b, having an
electrical resistance of about
four ohms each, are joined, as shown, to two coils, c, d,
of manganin wire, each of about 20 ohms resistance,
forming with the -two strips a Wheatstone's bridge.
A variable resistance, e, of several thousand ohms is
shunted around one coil and serves to bring the
whole to an electrical balance. Sometimes a small
resistance of copper, /, is included in one arm of the
Wheatstone's bridge to prevent its unbalancement
as the surrounding teniperature changes. A cur-
rent of about 0.1 ampere from a storage battery of

— L|J O-vWV-O

FIG. 23.

THE SUN

several cells, in parallel, flows constantly through
the bridge, and the adjustment is observed by a
highly sensitive galvanometer, g. If the radiation
is caused to fall on one of the bolometer strips, its re-
sistance increases, and there results a deflection -of
the galvanometer proportional to the heat produced
by the radiation. The record of the galvanometer is
kept automatically on the photographic plate which
is moved vertically by the clockwork at the same
time that the spectrum is moved across the bolom-
eter strip, so that rising and falling temperatures
of the strip, due to changes of intensity of the spec-
trum, are indicated by higher and lower parts of the
curve, photographically traced by the little spot of
sunlight reflected by the tiny mirror of the galvanom-
eter needle. Fig. 24 gives a pair of such energy
curves or holographs of the solar spectrum. Some
of the principal Fraunhofer lines give great depres-
sions of the curve, and are indicated on the margin
of the figure. At the points marked * * a shutter
was introduced in front of the slit of the spectro-
scope to give the zero of radiation. At the points
marked f t diaphragms were introduced to diminish
the intensity of the spectrum, so that the photo-
graphic trace would not run off the plate. The scale
of the intensity as thus altered is indicated on the
margin.

In Chapters III and VII are given the applica-
tion of the bolometer for the determination of the
"solar constant of radiation," the transparency of

SOLAR INVESTIGATION

THE SUN

the atmosphere for rays of different wave lengths, the
investigation of the comparative brightness of dif-
ferent parts of the solar image, and the determina-
tion of the temperature of the sun. The astonishing
sensitiveness of the bolometer may be understood
when it is said that, in ordinary use, changes of

temperature of less than 77; of a degree C. are

measured, and by special installation this sensitive-
ness may be increased 1000 fold. The still more
astonishing sensitiveness of the eye is indicated by
the fact that we receive enough light through the
pupil of the eye from a star of the sixth magnitude
to see it, though with the most sensitive bolometer
it would require a mirror perhaps ten feet in diameter
to concentrate enough rays from such a star to make
its heating observable. This is the more striking
because the eye is affected by only a short range of
spectral colors, while the bolometer measures the
total radiation of all wave lengths.

CHAPTER III

THE PHOTOSPHERE

Telescopic View. — The Photospheric Spectrum. — Rowland's Spec-
trum Tables. — Chemical Elements Found and Not Found. —
Corrections to Rowland's Wave Lengths. — Levels. — Pressures. —
Convection Currents. — Limb Spectra. — Variation of the Sun's
Brightness. — Solar Temperatures. — Spectroheliography. — Solar
Rotation.

As viewed through the telescope, or photographed,
the radiating surface of the sun, called the " photo-
sphere, " presents a brilliant disk covered by indis-
tinct mottlings sometimes spoken of as the " rice-
grain-structure." Objects much less than a second of
arc or 400 miles in diameter, cannot be well seen on
the sun, so that these " rice-grains, " which appear
according to different authors from 100 to 500 miles
in diameter, are really large areas. Some authors
speak of the bright areas of this mottled appearance
as " granulations, " and the darker parts as "pores."
Generally a few very dark patches called "sun
spots" may be seen, and around them, if they hap-
pen to be observed near the edge or "limb" of
the sun, are found very bright areas called "faculse. "
The faculse are seldom seen very much more than a
quarter radius within the limb. Photography reveals
at once, what the eye recognizes less easily, that the

THE SUN

photosphere falls off in brightness towards the sun's
limb. A photograph well exposed at the center will
be very weak at the limb. Plate III shows this
clearly, and also exhibits the rice-grain structure,
sun spots, and faculse. Sun spots march nearly reg-
ularly across the sun's disk in about 13.6 l days, and
appear after an equally long absence, which indicates
that the sun rotates upon its axis.

THE PHOTOSPHERIC SPECTRUM

The spectrum of the sun's photosphere is a con-
tinuous bright background of color crossed by dark
lines and bands. Newton recognized seven colors in
the spectrum, comprising violet, indigo, blue, green,
yellow, orange and red, but these blend into one
another by perfectly imperceptible gradations of in-
numerable hues. By photography and by the bolom-
eter, the solar spectrum has been followed beyond
the violet end as seen by the eye (which occurs
about wave length 0.38^), as far as wave length 0.29/*.
Here the rays are almost wholly cut off by losses in
the earth's atmosphere and in the sun's outer envel-
opes. Beyond the red, which may be observed with
the eye to wave length 0.80/*, Abney has photo-
graphed, by the aid of specially dyed plates, to wave
length l.l/*, and with the bolometer the solar spec-
trum has been measured at the Smithsonian Astro-
physical Observatory as far as wave length 5.3/*.

'The earth is meanwhile advancing, so that this is not the half
period of the sun's sidereal rotation.

PLATE III

DIRECT SOLAR PHOTOGRAPH. (EllermarO
1908, April 30. G. M. T. 2 h 30 m. P. S. T. 6 h 30 m A. M.

THE PHOTOSPHERE

Probably sun rays might be recognized with the
bolometer at intervals as far as 20^, but beyond
this they would probably be practically all cut off
by losses in the earth's atmosphere.

The dark lines and bands of the solar spectrum,
named from their discoverer "Fraunhofer lines/'
have two different sources. A considerable number
of lines, notably in the red and infra-red regions of the
spectrum, are caused by the absorption of gases and
vapors in the earth's atmosphere. The chief of
these terrestrial absorbents are oxygen, water vapor,
and carbonic-acid gas. By far the greater number
of the Fraunhofer lines, however, are formed by the
absorption of solar rays by gases in and about the
sun itself; notably by iron, nickel, calcium, titanium,
cobalt, chromium, manganese, carbon, vanadium,
sodium, magnesium, and hydrogen. The existence of
these elements and many others in the sun is proved
by the occurrence in the solar spectrum of dark lines,
occupying the same relative positions as to wave
length, and generally of nearly the same relative in-
tensity, that the characteristic bright lines of these
elements occupy in their spectra as produced in the
laboratory. As shown by Kirchhoff and Bunsen in
1859, dark lines are produced in a bright continuous
spectrum by interposing cooler vapors or gases be-
tween the source of light and the spectroscope, and
these lines occupy the same positions that the bright
lines of the vapors or gases would occupy if the latter
were themselves the sole sources of light. Conform-
8 87

THE SUN

ably to this discovery it will be shown in a later chap-
ter that the spectrum of the outermost solar layer,
called the " chromosphere/' when seen alone at solar
eclipses, is a bright line spectrum »vhich is almost the
exact reversal of the photospheric spectrum. The
layer in which the dark lines have their rise is accord-
ingly called "the reversing layer."

As any gases between the observer and the sun may
produce dark absorption lines in this way, it is not
at first apparent how to distinguish between terres-
trial and solar gases. There are two ways of testing
whether a given Fraunhofer line is solar or atmos-
pheric. The first is by observing its intensity rela-
tive to other lines at high and low elevation of the
sun above the horizon. Atmospheric, or as they are
called, " telluric," lines will generally be strengthened
at low sun, because the layer of air traversed will then
be greater. A second and better method of discrim-
ination consists in forming an image of the sun and
causing rays from its east and west limbs to be re-
flected together simultaneously into the slit of the
spectroscope, so as to give rise to two superposed
solar spectra, one of light from the east limb, the
other from the west. Telluric lines will occupy the
same position in the two spectra, but solar lines will
be shifted with reference to one another owing
to the rotation of the sun, which produces a very
notable Doppler effect. This is shown in Plate
IV, Fig. 2, which includes the oxygen band B and
some solar lines in its vicinity as photographed

3 g

* -a

^ o

02 g
"'I

^ W

£ 5s

I o
I 3

s I

w o

& PQ

THE PHOTOSPHERE

at Mount Wilson by St. John under remarkably
fine conditions.

ROWLAND'S SPECTRUM TABLES

The solar spectrum has been photographed at great
dispersion by numerous observers, but most notably
by Rowland. He published about 1895, in the early
volumes of the Astrophysical Journal, his great " Pre-
liminary Table of Solar Spectrum Wave Lengths,"
which still forms the basis for solar and stellar re-
searches. Rowland states in his introduction that
he photographed the arc spectrum of all the then
known elements except gallium in connection with
the solar spectrum, but that the work of identifica-
tion of lines in the solar spectrum with the arc lines
would be a further labor of years. This work of
identification has never yet been completed, nor has
a correspondingly full comparison of the solar spec-
trum with the spark spectra of the elements been
attempted. In Rowland's " Preliminary Table" there
are about 14,000 lines recorded. Their wave lengths
are given to seven places of figures, that is, to thou-
sandths of an Angstrom. For each line is given its
intensity. The intensities go from 1, a line just
clearly visible on Rowland's spectrum map, up to
1000 for the strong calcium lines H and K. Below
1, the intensities go down to 0000, indicating lines
more and more difficult to see.

The great lines of the solar spectrum, named long
ago for the letters of the alphabet, are as follows :

THE SUN

TABLE III. — Principal solar spectrum

7593.842
Oxygen3

7184.57
Waters

6860.970

Oxygen3

C(Ha)

6562 . 835
Hydrogen

.->,SX!».!)7.~
Sodium

52(i!»..r).r)l
Iron

Corrected
wave length1
Element

5183.620
Magnesium

F(H/3)

4861.350
Hydrogen

G(Hy)

4340.471
Hydrogen

3968.491
Calcium

3933.680
Calcium

Corrected
wave length1 , .
Element

1 According to the table of corrections below.

2 Edge of the head of A.
a Terrestrial lines.

About one-third of the 14,000 solar lines were iden-
tified by Rowland and ascribed by him to various
chemical elements. In a good many cases a line
is attributed to several elements at once. In such
cases the coincidence with them all is probably not
generally exact, but only so close that, even with
Rowland's very high dispersion, the several linos
overlap. Investigations with much higher dispersion
on bright line spectra indicate that in many cases
apparently single lines of single elements are really
resolvable into groups. But perhaps even such very
high resolving powers would generally fail to separate
the blended lines of Rowland's table, because, owing
to pressure or other conditions, the several lines
involved are so much widened as to overlap. For
many years Lockyer maintained plausibly that the
elements had common constituents which gave rise
to common lines in the spectrum, but this so-called
"basic line" hypothesis is not now generally held.
The following summary of Rowland's identifications

THE PHOTOSPHERE

is taken from Young's "The Sun," with slight
changes :

Chemical Elements Found and not Found in the
Sun.— The first columns of the following table give
the chemical elements found by Rowland to exist
in the sun arranged according to the intensity of their
solar lines, and with their atomic weights annexed.
The last columns give them arranged in the order of
the number of their solar lines, and with the numbers
occasionally annexed. The sign f indicates that
the element has not been identified in eclipse chro-
mospheric spectra.

TABLE IV. — Chemical elements found in the sun.

9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.

Calcium (40.09)
Iron (55.85)
Hydrogen (1.008)
Sodium (23.00)
Nickel (58.68)
Magnesium (24.32)
Cobalt (58.97)
Silicon (28.3)
Aluminum (27.1)
Titanium (48.1)
Chromium (52.0)
Strontium (87.62)
Manganese (54.93)
Vanadium (51.2)
Barium (137.37)
Carbon (12.00)
Scandium (44.1)
Yttrium (89.0)
Zirconium (90.6)
fMolybdenum (96.0)
Lanthanum (139.0)
fNiobium (93.5)
fPalladium (106.7)

Iron (2000 or more)
Nickel
Titanium
Manganese
Chromium
Cobalt

Carbon (200 or more)
Vanadium
Zirconium
Cerium

Calcium (75 or more)
Neodymium
Scandium
Lanthanum
Yttrium
Niobium
Molybdenum
Palladium

Magnesium (20 or more)
Sodium (11)
Silicon
Hydrogen
Strontium
91

THE SUN

24. tNeodymium (144.3) Barium

25. tCopper (63.57) Aluminum (4)

26. Zinc (65.37) Cadmium

27. Cadmium (112.40) Rhodium

28. Cerium (140.25) Erbium

29. tGlucinum (9.1) Zinc

30. tGermanium (72.5) Copper (2)

31. tRhodium (102.9) Silver

32. Silver (107.88) Glucinum

33. Tin (119.0) Germanium

34. Lead (207.10) Tin

35. Erbium (167.4) Lead (1)

36. fPotassium (39.10) Potassium

Besides these thirty-six elements, thus arranged, it
has been found that helium (4.0) and gallium (69.9)
certainly show solar lines, although helium lines are
hide-and-seek things, and for some reason only occa-
sionally appear as dark lines in the solar spectrum.
There also appear very faint dark solar lines, nearly
or exactly corresponding in their position to some of
the strongest arc lines of:

TABLE V. — Chemical elements doubtfully occurring in the sun.
Ruthenium (101.7), Indium (114.8), Tantalum (181.0),
Tungsten (184.0), Osmium (190.9), Iridium (193.1),

Platinum (195.0), Mercury (200.0), Thallium (204.0),
Bismuth (208.0), Thorium (232.42), Uranium (238.5).
The mean atomic weight of these elements is 186 . 95.

The lines of the important elements of the halogen
group, Fluorine, Chlorine, Bromine, Iodine; those
of the oxygen group, Oxygen,1 Sulphur, Selenium, and

1 Since this was written St. John has found that a triplet of faint
lines attributed to oxygen occurring beyond A in the extreme1 red,
shows relative displacements at the sun's limbs. Hence, we must
probably admit free oxygen as giving a solar spectrum. Combined
oxygen and combined nitrogen give solar band spectra.

THE PHOTOSPHERE

Tellurium; those of the nitrogen group, Nitrogen,
Phosphorus, Arsenic, and Antimony (Bismuth doubt-
ful) do not appear to have been found in the photo-
spheric spectrum, or in the spectrum of the chromo-
sphere. This singular omission comprises nearly all
of the prominent " negative" elements, and Boron,
another of them, is also absent from the solar spec-
trum. Further remarks on this subject will be made
later.

There is considerable interest attaching to the
relations of atomic weight of the elements and the
intensity of their solar lines. Taking the thirty-six ele-
ments of the intensity table in order, in four groups of
nine each, the average atomic weights are as follows:

Elements 1-9, 35.26; elements 10-18, 64.04;
elements 19-27, 101.27; elements 28-36, 107.25.

In the last group, as thus divided, occur glucinum
(9.1) and potassium (39.10). The former has two,
and the latter one identified line, and, as these lines
are also very weak, it is not impossible that these two
elements may by future investigation fall out of their
strange company.1 If so, the mean atomic weight of
the remaining seven elements would be 131.00. In
Group II of this arrangement appears carbon (12.00),
but, judging from Kayser's " Handbuch," the solar
" carbon " lines belong to carbon compounds of high
molecular weights. Hardly less interesting than the

and Runge question the existence of potassium lines in
the photospheric spectrum.

THE SUN

classification just given is the further fact that most
of the elements of the platinum group, and some other
elements of very high atomic weight found commonly
on the earth, are only doubtfully recognized in the
sun, although they give strong lines in the arc. The
full significance of these relations will be further dis-
cussed in Chapter VI, but it may be said here that
the explanation of the decrease of intensities with in-
creasing atomic weights seems to depend on the depth
of these gases below the sun's surface. We may sup-
pose that the interesting elements radium and ura-
nium might not produce lines in the solar spectrum,
even if these elements exist in the sun, because of
their high atomic weights.

The element oxygen undoubtedly exists in the sun
because the flutings of titanium oxide are very prom-
inent in sun-spot spectra. It might be anticipated
that the well-known oxygen lines themselves would
be found in the photospheric spectrum if it were not
that the earth's atmosphere itself contains so much
oxygen as to produce such intense oxygen lines that
solar effects are unrecognizable. However, photo-
graphs of the spectra of the two opposite limbs of
the sun show the negative, for in these spectra all
solar lines are displaced by Doppler effects, but the
well-known oxygen lines show none. Nitrogen, also
found plentifully in the earth's atmosphere, behaves
similarly. It is a peculiar feature of the solar spec-
trum that very few of the so-called negative, or non-
metallic, elements are recognized from it. Thus, the

THE PHOTOSPHERE

important halogen group of elements, which includes
such common elements as chlorine and bromine, is
unrecognized. So also with the important element
sulphur. These omissions are very remarkable and
not yet, I think, well understood.

However, it is found frequently in the laboratory
that the spectrum of a mixture or compound of two
elements is apt to show one of them predominatingly,
or even alone. Especially does a metal thus often ex-
clude a nonmetal. But yet oxygen and helium, which,
although existing in the sun, are of slight effect in the
solar spectrum, are very prominently in evidence in
the spectra of many of the stars. Since oxygen is
certainly present in sun spots as an oxide, and nitro-
gen as cyanogen, though they do not give their char-
acteristic lines as elements,1 the other elements just
mentioned may also be present in the sun without
giving their spectral lines.

Some of the " unknown" lines have now been as-
signed to their appropriate elements, but more than
half of Rowland's lines are still unidentified. A large
number of these are, however, very weak. It is
probable that within the next decade many of them
will be identified, either with spark or arc spectra.

Corrections to Rowland's Wave Lengths.

It has been shown that the wave lengths assigned
by Rowland must be altered. His system is based

1 Three faint lines attributed to oxygen are, however, now known
to be solar. See note on preceding page.

THE SUN

on measurements by several observers of the wave
length of the yellow sodium lines. Measurements
by the interferometer in the hands of Michelson,
Fabry, Perot, Buisson, and other experimenters
have shown that Rowland's assumed wave-length at

D should be reduced by about . This change,

though considerable as wave lengths go, would be
of little consequence, if Rowland's system was self-
consistent. But it is further shown that the differ-
ence from the true scale differs for different parts of
the spectrum about as follows:

TABLE VI. Corrections to wave lengths in Rowland's Preliminary
Table of Solar-Spectrum Wave Lengths.

Wave lengths

3000

3200

3400

3600

3700

3900

4100

4300

4500

4700

Corrections

-.106

-.124

-.148

-.155

-.140

-.144

-.152

-.161

-.172

-.179

Wave lengths

4900

5100

5300

5400

5600

5800

6000

6200

6400

6500

Corrections

-.176

-.170

-.172

-.212

-.218

-.209

-.213

-.212

-.209

-.210

These discrepancies are to be ascribed largely to
certain deficiencies of the grating as a means of
measuring wave lengths, and not to avoidable inac-
curacy of Rowland's work, although he neglected
certain small corrections not strictly negligible. An
effort is now (1910) being made, with international
cooperation, to establish a consistent and highly
accurate system of wave lengths. The results,
while not yet officially announced, can hardly differ

THE PHOTOSPHERE

appreciably from those indicated in the above table
of corrections to Rowland's wave lengths.

An accurate table of solar wave lengths and of the
wave lengths of the lines of all the chemical elements
constitutes the fundamental groundwork of all mod-
ern spectroscopic investigation. What the great
star catalogues are to astronomy, the wave length
tables are to astrophysics. On them are based in-
vestigations of motion and pressure in the sun and
stars, of the elements present, the magnetic fields
which exist, the possibility of anomalous dispersion
phenomena, and other solar and stellar conditions.

LEVELS

In the general spectrum of the solar photosphere
we have an index of conditions which exist in a layer
practically at the surface of the sun, for, as shown by
terrestrial experiments, it takes only a little of an
absorbing gas to produce a dark line in the spectrum.
But it is thought that a difference of average level
exists in the positions of the layers which produce
lines of different elements, and even different lines of
the same element. The layer of the sun which gives
rise to the dark Fraunhofer lines, though thin rela-
tively to the solar radius, may yet be thought of as
made up of several layers of differing level. Calcium
lines are thought to represent a higher level than
iron lines, and hydrogen lines one still higher. Yet
further, as the longer wave lengths are often more
readily emitted by an element than the shorter ones,

THE SUN

that is, are emitted at lower temperatures, it may be
that a red line of an element on the whole represents
a higher level than a violet line of the same element.
The continuous background of the solar spectrum
represents a lower average level than any of the spec-
trum lines, as, of course, follows from Kirchhoff and
Bunsen's principle. However, the continuous back-
ground offers less opportunities of investigation
than the lines, so that less can be learned of the
levels it represents than of the so-called " re versing
layer" where the lines are formed. The lines them-
selves are not to be regarded as dark except by con-
trast. If seen against a black ground they would be
dazzlingly bright, but, as they are formed in the
outer and cooler layers of the sun, they are less
bright than the spectrum background against which
they are seen. The light of the deeper solar layers
cannot get out, if it is of a wave length where great
absorption occurs, as is the case in the Fraunhofer
lines.

PRESSURES

The effect of pressure is two-fold. It broadens
lines and shifts them in wave length. Generally the
effect is the same whether a gas is compressed by a
like or a foreign mass of gas. Pressure shifts can be
distinguished from velocity shifts, because, while the
former increase on the whole with increasing wave
length, they affect different lines of the same ele-
ment and of different elements with shifts of quite
arbitrarily differing amounts, and some lines, indeed,

THE PHOTOSPHERE

are practically unaffected; velocity causes shifts
which differ, it is true, in different parts of the
spectrum, but which are directly proportional to the
wave lengths. Several investigations have been
made to determine the pressures prevailing in the
reversing layer. Jewell, in 1896, by examination of
grating spectra, found that for most solar lines the
wave lengths are greater by a few thousands of an
Angstrom than the corresponding lines in the arc
spectrum at atmospheric pressure. He found, to be
sure, many anomalies which tended to throw doubt
on the explanation of these shifts as due to pressure,
but the following estimates of the pressure in the
reversing layer are given by Jewell, Mohler, and
Humphreys :l

ELEMENT.

Alumi-
num.

Cobalt.

Silicon.

Cal-
cium.

Chro-
mium.

Man-
ganese.

Iron,
Nickel,
Copper,
each.

Pressure ....

2atm.

6 or 31

1 Depending on what group of lines is observed. The H and K
lines, however, are not included, nor is 4227.

In 1909 Fabry and Buisson examined numerous
iron lines, mostly between wrave lengths 4,000 and
4,500 Angstroms, by interference methods, and dis-
covered small shifts in the same sense as found by
Jewell. They also investigated the behavior of the
anomalous cases and explained them as due to un-
symmetrical broadening under pressure. They con-

1 Astrophysical Journal, vol. iii, p. 139, 1896.
99

THE SUN

eluded that the solar reversing layer for iron lines
lies under a pressure of 5.5 atmospheres. Evershed,
however, criticises their interpretation of the behav-
ior of the anomalous lines, and thinks the evidence
tends to show that the pressure is less than one atmos-
phere.

CONVECTION CURRENTS

Some recent measurements of Adams indicate
velocities of ascent of from 0.1 to 0.3 kilometers per
second in the solar layer where the metallic absorp-
tion lines are formed. This seems at first sight hard
to accept, because what goes up must surely come
down again, so that we might suppose there would be
as much of a Doppler effect of descent as of ascent.
But in this connection we must consider the temper-
atures of the ascending and descending currents.
Adams refers to unpublished experiments of Fox
which indicate the brighter areas or " granulations "
of the sun's surface as yielding a spectrum strong
in " enhanced" or high temperature lines, and the
darker spaces or " pores" between as regions of "arc"
or low temperature lines. Adams finds the "en-
hanced" lines indicate maximum velocities of ascent.
He argues that the spectrum would be predominat-
ingly influenced by the hotter and brighter parts, and
as these are shown to be ascending the whole spec-
trum would hence be indicative of ascent. Evershed
had advanced a similar argument in 1902 to account
for peculiarities of the "flash spectrum."

It is to be supposed that vertical circulation may
100

THE PHOTOSPHERE

be active in the sun because the interior is, of course,
hotter than the exterior; the latter is continually
being cooled by radiation, and, being thereby made
denser, would tend to fall. Velocities of 0.1 to 0.3
kilometers per second are, to be sure, greater than
those of any winds we know of on the earth. On the
earth, moreover, the vertical circulation and the
winds are to a large extent due to the variable tem-
perature conditions depending on the changes from
day to night anU from summer to winter. As the sun
has neither night nor day, summer nor winter, it is
to be regarded rather as having approximately
reached a steady state of affairs; but still, in con-
sideration of the sun's enormous temperature, Mr.
Adams' results give no cause for surprise.

St. John has still more recently published a beauti-
fully accurate study of the displacements of the cal-
cium lines H and K, and of the calcium circulation to
be inferred thereby in the sun. He distinguishes
three parts of each of these broad lines, which he in-
dicates by the subscripts 1 , 2, 3. K3 is the narrow dark
line in the center, K2 the bright lines on either edge of
K3, and KX the dark, broad, diffuse edges on the out-
sides of the K2 regions. Similarly for H, St. John
concludes:1 "The calcium vapor producing the ab-
sorption band K3 in the solar spectrum has a descend-
ing motion over the general surface of the sun of 1.14
kilometers per second in the mean. . . . The cal-
cium vapor to which the bright emission line K2 is

1 Contributions of the Mount Wilson Solar Observatory, No. 48.
101

THE SUN

due has an ascending motion over the general surface
of the sun of 1.97 kilometers per second in the mean.
. . . The wave lengths of K2 (mean of both parts of
K2) and K3 reduced to the limb are 3933.667 and
3933.665 respectively. The corresponding wave
length in the arc at atmospheric pressure is 3933.667.
The mean pressure in the intermediate emitting layer
is, therefore, approximately one atmosphere. . . .
The shorter wave length of the K3 line may be inter-
preted as indicating a somewhat lower pressure in the
upper absorbing layer, though the smallness of the
quantities involved does not permit a positive con-
clusion. ... In the case of the intermediate and
highest levels of calcium vapor [there is indicated an]
absence of currents of appreciable velocity parallel to
the solar surface. ... The widths of the H3 and K3
lines at the center [of the disk], compared with the
corresponding widths in the arc, point to an extremely
small quantity of the calcium vapor in the upper
levels of the solar atmosphere. . . . The average
appreciable height of the atmospheric calcium shown
by a radial slit is about 5,000 kilometers above the
photosphere. The thickness of the upper absorbing
layer is approximately 1,500 kilometers. Allowing
700 kilometers for the reversing layer, the emitting
layer would have a thickness of approximately 3,000
kilometers. The elevation at which the K line is
appreciable is about 500 to 600 kilometers above the
level at which the H line ceases to show. . . . The shift
between limb and center is 0.015 Angstroms for the

102

THE PHOTOSPHERE

H3 line, and in agreement with that obtained for the
K3 line."

An interesting result on the rotation of the sun as
measured by the line K3 will be given below.

It is by no means to be supposed that the fact of
the enormous transfer of heat from within the gase-
ous body of the sun to the exterior, to supply that
which is lost by radiation to space, requires us to
imagine a strong vertical circulation to carry it on.
At low temperatures, as for instance, between a body
at boiling temperature and one at freezing, convec-
tion is rather more important than radiation as a
means of transferring heat ; but this is probably not
the case at the temperatures prevailing within the
sun. For radiation increases with the fourth power
of the temperature, and convection by no means at
such a tremendous rate of increase. Hence, as the
material of the sun is probably transparent, we must
suppose that the heat from within the sun becomes
available at the surface to supply the losses of energy
by radiation to space chiefly by a process of internal
radiation, gradual absorption in a long path outward,
and reradiation nearly counterbalancing the absorp-
tion. This process is repeated as many times as nec-
essary, and except for the very short time occupied by
absorption and reradiation, is performed at velocities
of nearly 186,000 miles a second, and produces quick
communication of energy from within outward.1

'See Schwartz child, "Ueber das Gleichgewicht der Sonnenatmos-
phare," Gottingen Nachr., Mathphys. KL, 1906, pp. 1-13. Prof. T. J. J.
See also takes this view of the function of internal solar radiation.
9 103

THE SUN

If we admit that Adams has shown an effective
velocity of ascent averaging 0.12 kilometers per
second and the shifting, thereby, of average solar
lines of wave length 4200 by 0.0015 Angstroms
toward the violet, then a correction must be applied
to Fabry and Buisson's results tending to increase by
one atmosphere the supposed pressures in the re-
versing layer. Adams has investigated by a purely
differential method the shifting of lines between the
center and limbs of the sun, and finds that, after
correcting his results for this supposed velocity of
ascent, there remain in the spectra of the limbs well-
substantiated displacements towards the red, which
are best explained by ascribing them to effects of
pressure.1 Hydrogen, sodium, calcium, and mag-
nesium lines show almost no displacement. Lines of
titanium, vanadium, and scandium show moderate
displacement, and those of iron and nickel consider-
able shifts, averaging .007 Angstroms. Lines of the
elements of high atomic weight show very small dis-
placements, as do also lines strengthened at the
limb. Enhanced lines, as a class, show maximum
displacements, which apparently grow with the de-
gree of enhancement of the several lines. These, at
first sight highly discrepant, observations harmonize
beautifully under Adams' clever discussion, which we
shall reserve till we come to the chapter on solar
theory. Adams confirms Fabry and Buisson's ob-
servation that the violet edges of lines do not shift.

1 Contributions of the Mount Wilson Solar Observatory, No. 43.
104

THE PHOTOSPHERE

LIMB SPECTRA

The spectrum of the sun's limb is, as would be ex-
pected from the general darkening of the sun towards
the limb, weaker than that at the center. In
violet light eight or ten times as long photographic
exposure is required for the limb as for the center.
This ratio is reduced to four or five for red light.
But, besides this general effect, the Fraunhofer lines
are much altered, especially in the violet. The
stronger lines almost completely lose their side
shadings or "wings" in the limb spectrum, while in
sun-spot spectra, as we shall see in Chapter V, the
wings have increased prominence. As against this
marked difference from spot spectra, the limb spec-
trum is like that of spots in having similar changes
of relative intensity of lines, so that lines strengthened
in spots are strengthened, though in less degree, at
the limb, and vice versa. As in the spots, the so-
called spark or "enhanced" lines are often weakened
at the limb. The H> line of hydrogen, on the con-
trary, is widened and perhaps strengthened at the
limb, although narrowed and weakened in spots.

VARIATION OF THE SUN'S BRIGHTNESS

The variation of the brightness of the sun from
the center to limb is much more readily determined
by the bolometer than by the photographic plate.
Fig. 25 shows the distribution of brightness along a

105

THE SUN

diameter of the sun's disk for rays of different wave
lengths. The reader will notice how great the con-
trast in brightness between center and edge is for
the shorter wave lengths. This fact is also shown

FIG. 25. — BRIGHTNESS ON SOLAR DISK

by the following table, which gives the brightness
at different percentages of a solar radius from the
center of the solar disk, and with which the data of

Figs. 25 and 26 agree.

106

THE PHOTOSPHERE

TABLE VII. — Distribution of radiation over the sun's disk.

Fraction

xvaaius

0.00

0.40

0.55

0.65

0.75

0.825

0.875

0.92

0.95

Wave length

i t

Oft. 323

144

129

120

112

0.386

338

312

289

267

240

214

188

163

141

0.433

456

423

395

368

333

296

266

233

205

0.456

515

486

455

428

390

351

317

277

242

0.481

511

483

456

430

394

358

324

290

255

0.501

489

463

437

414

380

347

323

286

254

0.534

463

440

417

396

366

337

312

281

254

0.604

399

382

365

348

326

304

284

259

237

0.670

333

320

308

295

281

262

247

227

210

0.699

307

295

284

273

258

243

229

212

195

0.866

174

169

163

159

152

145

138

130

122 •

1.031

111

108

105.5

103

94.5

90.5

1.225

77.6

75.7

73.8

72.2

69.8

67.1

64.7

61.6

58.7

1.655

39.5

38.9

38.2

37.6

36.7

35.7

34.7

33.6

32.3

2.097

14.0

13.8

13.6

13.4

13.1

12.8

12'. 5

12.2

11.7

Wave length

of Max.

Oft. 458

Oft. 467

Oft. 471

Oft. 474

Of*. 478

Of*. 483

Of*. 489

Oft. 496

Oft. 505

Following the lines of the table from left to right,
the reader may note the decrease of brightness from
the center of the sun to 95 per cent of the radius out-
ward1. The results are arranged vertically in order
of wave length, and the numbers have been so ad-
justed that, by taking any single vertical column, as
for instance, that for 75 per cent, out on the radius,
the reader may find for a single zone of the sun the
distribution of brightness on a uniform scale of wave
lengths for the normal spectrum outside the earth's
atmosphere. The data as regards distribution along
the radius for wave length 0.323^ are from results

1 There is a tendency of all the data plotted in Fig. 25 to show a
less rapid fall of brightness from 95 to 97 per cent out on the radius,
than would be expected. This may be due to error.

107

THE SUN

of Schwartzchild and Villager, who obtained them by
photographing the solar image formed by a silvered
lens. The remainder of the data are from the bolo-
metric results of Abbot and Fowle.

In the preceding table the maximum number in
each vertical column is indicated by black-faced
type. But the wave length intervals are not small
enough to show accurately in this manner the amount
of shifting of the wave length of maximum radiation
for light coming from greater and greater distances
from the center of the sun's disk. By means of
plotting the values, we find that the true wave
lengths of maximum intensity are as given in the
lower line of the table. This shows a shifting of the
maximum of radiation from 0.458/* at the center of
the sun's disk to 0.505/* at 95 per cent out on the
radius. We shall see that a similar shifting of the
wave length of maximum radiation occurs between
the photosphere and the umbra of a sun spot. The
dotted curve of the accompanying Fig. 26 shows the
distribution of radiation in the spectrum for light
of the whole sun's disk as it would be if viewed
outside the earth's atmosphere. Similar curves
are given also in Fig. 26 for the center of the
sun's disk and for points 55, 82.5 and 95 per cent
of the radius towards the limb. No account is
made in the figures of the Fraunhofer lines sepa-
rately, although collectively they doubtless affect
the forms of the curves, especially for the shorter
wave lengths.

108

THE PHOTOSPHERE

SOLAR TEMPERATURES
First Method.

These five energy curves of Fig. 26 are of inter-
est as they indicate the probable temperatures in

;

OS O6 Q7 Q8 Q9 1.0

_I5 1.6

FIG. 26. — ENERGY SPECTRA ON SOLAR DISK.

the photosphere. From Wien's displacement law
( xmax. T = 2930) given in Chapter II, we may find,

by substituting the values indicated for

109

the

THE SUN

values of the absolute temperatures for which a per-
fect radiator would give the same wave lengths of
maximum radiation. The values are given in Table
VIII.

Furthermore, as the five curves of Fig. 26 are
plotted with ordinates proportional to intensities,
and abscissae proportional to wave length, their in-
cluded areas are proportional to the intensities of
the emission of all wave lengths combined, as emitted
from the selected regions of the sun's disk. If the
total emission is comparable to that of a perfect
radiator, then, by Stefan's law, it is proportional to
the fourth power of the temperature of the emitting
body. Hence, the fourth roots of the areas included
by the five given curves should be in inverse ratio of
the wave lengths of maximum emission. The fol-
lowing table shows in its fourth and sixth lines how
the matter comes out:

TABLE VIII. — Energy spectrum relations over the sun's disk.

POSITION.

Whole
Disk.

Center.

55%

82.5%

95%

Wave lengthof max-
imum

0.^468

O.Ai458

0./i471

0./I483

0.M505

2930

6260°

6400°

6220°

6070°

5800°

Xmax.

Ratios by maximum

1.079

1.104

1.073

1.047

1.000

Ratios of Areas ....

1.407

1.620

1.476

1.249

1.000

Ratios by fourth
roots of areas ...

1.090

1.128

1.102

1.057

1.000

'On the absolute scale of Centigrade degrees water freezes at
273° and boils at 373°.

110

THE PHOTOSPHERE

The greatest disagreement between the ratios
through maximum and through the fourth roots of
the areas is about 2J/£ per cent.

Second Method.

Another method of estimating the probable solar
temperature is by attempting to match, as well as
possible, the distribution of energy in the whole
range of solar spectrum with the distribution compu-
ted by the Wien-Planck formula given in Chapter
II. Referring to Fig. 17, the reader will find in
curves B and A the distribution according to Wien-
Planck in the spectrum of a perfect radiator at
6200° and 7000° C. absolute, and also in curve C the
energy spectrum for the general solar surface. No
account is made in the computations of the relative
values of the constant CA and the solar constant of
radiation. The 6200° curve has been repeated at
B7 on a larger scale of ordinates, and the observed
curve also repeated at C' on a scale nearly matching
that of B'. The observed curve falls below the com-
puted ones in the ultra-violet, but this discrepancy is
to be expected, partly because the ultra-violet solar
spectrum is crowded with lines of selective absorp-
tion.

On the other hand, the observed curve rises above
the computed ones in the infra-red, a feature to which
Professor Bigelow has repeatedly called attention.
It has just been said, and it will be spoken of at
greater length in Chapter VI, that the rays from the

111

THE SUN

center of the sun's disk seem to arise from a source at
higher temperature than those emanating from the
sun's limb. In accordance with the explanation of
this phenomenon which will be advanced in Chapter
VI, it would be expected, also, that solar rays of long
wave lengths would appear to come from sources
of higher temperature than would those of shorter
wave lengths. If so, we shall thereby understand
why the infra-red parts of curves C and C' (Fig. 17)
rise above curves A, B and B', respectively, for
curves C and C' do not represent the spectrum of a
source at constant temperature. Their infra-red
parts correspond to much hotter sources than do
their visible and ultra-violet parts.1 It is evident,
however, that the 7000° curve, except in the ultra-
violet, is a better match for the observations than the
6200° curve. The large discrepancy in the ultra-
violet is probably due in part to the general tendency
toward lower temperature in the sun for short wave
length rays, but far more to the throngs of Fraunhofer
lines in that region of spectrum, which though not
shown separately, very greatly affect the form of the
curve.

Third Method.

Pointing seemingly to a lower solar temperature
than those we have considered are the following

1 The accuracy of the observed curve for wave lengths beyond
2/i is seriously impaired by the effect of terrestrial water vapor, so
that no conclusion should be drawn from the fact of the falling off
of the curve in this region.

THE PHOTOSPHERE

facts. As recently done for a large number of stars
by Wilsing and Scheiner, we may compute the ap-
parent temperature of the sun by the formula:

where E! and E2 are the intensities of energy at two
wave lengths \ and \, c2 a constant for which
Wilsing and Scheiner prefer the value 14200, and T
the absolute Centigrade temperature. Taking a
number of values of the intensity within a given
range of wave lengths, and proceeding according to
the method of least squares, 1 find:

Wave length
range

0.^30-
Q.ftSQ

0./i35-
0.^50

0./*50-
0.^70

0.^80-
l./*50

l./iOO-
l./xSO

l./dO-
l./i50

Temperature . . .

3932°

5142°

6900°

4493°

4006°

3840°

The falling off of computed temperatures for long
wave length rays is due to the fact that the observed
curve of Fig. 17 rises less rapidly from the infra-red
towards shorter wave lengths than does the 6200°
curve, and far less rapidly than the 7000° curve.
But, as we have said, and in accordance with a line
of explanation to be given in Chapter VI, we may
assume that as the wave length decreases the effec-
tive source of radiation approaches the exterior of
the sun, and, therefore, is cooler. Hence, although
the effective temperatures of emission for the infra-
red rays are probably exceeding 7000°, the observed
energy curve does not rise towards its maximum
from the infra-red side as fast as does the 7000°

113

THE SUN

curve, because each successive shorter wave length
is emitted from a lower average temperature than its
next longer neighbor, and is, therefore, less intense
than it would otherwise be. In the ultra-violet,
however, we may consider the temperature of effec-
tive emission not only apparently, but really far be-
low 7000°, on account of the superficial region of its
origin.

On the whole, the preceding review of the form of
the solar energy curve inclines us to set the average
temperature of the photosphere certainly above
6200°, and possibly near 7000°.

Fourth method.

It will be shown in Chapter VII that the intensity
of solar radiation at the earth's mean distance from
the sun is 1.95 calories per square centimeter per
minute. From Stefan's law, with Kurlbaum's con-
stant (see Chapter II), a perfect radiator emits radi-
ant energy from each square centimeter of its surface
at the rate of 76.8 X 10~12T4 calories per minute.
The radius of the sun being 696,000 kilometers, and
the mean radius of the earth's orbit 149,560,000 kilo-
meters, we would have the following equation for a
perfect radiator of uniform absolute temperature T
in the sun's place :

(696,000)2 X76.8 X 10'12T4 = (149,560,000)2 X 1.95

From this, T = 5860° absolute C. As this value falls
below those obtained previously, we may suppose the

1 1 I

CALCIUM SPECTROHELIOGRAM, H2. (Ellerman.)
1908, April 30. G. M. T. 12h 53m. P. S. T. 4h 43m P.M.

THE PHOTOSPHERE

sun's constant of emission is a little less than that of a
perfect radiator.

An observation which may be regarded as con-
firmatory of the view that the photosphere falls some-
what short of perfect radiating power is stated by
Jewell as follows : 1

" When some of the very best negatives of the solar
spectrum are carefully examined, it is found that
some of the sharp-edged, clean-cut, and unshaded
lines of iron, chromium, manganese, titanium, etc.,
have a faint, dark shading just outside the edge of the
line. It is very faint and difficult to observe (only
slightly darker2 than the general background of the
solar spectrum), but it is not due to contrast, as it is
not always present. It is a difficult observation to
make, but was observed sometime before the explan-
ation forced itself upon me. The correct explanation
undoubtedly is that this faint, dark shading (dark in
the negative [overbright in the spectrum]) is the re-
mains of an emission line, either produced at the
photosphere or lower down in the solar atmosphere
than the absorption line. "

This interesting observation, which has been con-
firmed by Evershed, appears to indicate that the
photospheric radiation in general, though undoubt-
edly coming from hotter, because deeper, layers than
the rays within the influence of the Fraunhofer lines,
yet lacks something of the full intensity of perfect or

1 Astrophysical Journal, vol. Ill, p. 99, 1896.

2 Darker in the negative, brighter in the spectrum.

115

THE srx

" black-body" radiation. For thus it might occur
that deep-lying (yet not the deepest lying) metallic
vapors would give in the immediate proximity of
their lines of powerful selective emission a more in-
tense radiation than the deeper lying and hotter, but
intrinsically less strongly emissive, layers of the
photosphere.

Summary.

In all of these ways discussed of estimating the
solar temperature, we have to go on the hypothesis
that the sun is a perfect radiator. This is, of course,
very unlikely, but if the sun's radiating power is
not perfect, then its temperature must, at any rate,
exceed that (5860° abs.) calculated by the fourth
method from Stefan's law of radiation. It is scarcely
less probable that the solar temperature exceeds that
(6260° abs.) calculated by the first method through
Wien's displacement law. For the influences tend-
ing to distort the form of the solar spectrum energy
curve seem to be of a kind to diminish the violet
most, and thereby to shift the maximum of energy
towards the red. Hence, we conclude that there is a
high probability that the average temperature of the
apparent photosphere exceeds 5860° or even 6260°
of the absolute Centigrade scale, and may be as high
as 7000° absolute Centigrade.

The reader may be disposed to question whether
a difference of temperature probably exists between
the center and edge of the apparent photospheric disk,

116

HYDROGEN (Ha) SPECTROHELIOGRAM. (Ellerman.)
1908, April 30. G. M. T. 13 h 6 m. P. S. T. 5 h 6 m P. M.

THE PHOTOSPHERE

as brought out in Table VIII, but this matter will
be further discussed in Chapter VI. One highly in-
teresting conclusion seems to follow from the fact of
the enormously high temperature of the photosphere,
taken in connection with the spectroscopic proof of
moderate pressures in the reversing layer. This con-
clusion is that no known substances can exist in the
photosphere except as gases.1 It has generally been
held that the photosphere is a cloudy layer. If so,
the materials composing the clouds are not known to
exist on the earth.

THE SPECTROHELIOGRAPH

When we examine the sun visually or by direct
photography, the source of the light is highly com-
plex. Many chemical elements, existing in a layer
many hundreds, or perhaps thousands, of miles deep
take part in sending the light. After tentative trials
in the early days of the spectroscope and of photog-
raphy, the matter of obtaining a view of the sun in the
light of one element, and substantially at one level,
was taken up about 1890 by Hale and by Deslandres
independently, and in 1891 Hale first employed his
spectroheliograph. Deslandres has long used a simi-
lar principle, but with intermittent instead of con-
tinuous displacement of the view over the solar sur-
face, in his "spectroscope a vitesse." He has lately
employed the spectroheliograph itself with great
success. The spectroheliograph, as explained in

1 See also Chapter VI.
117

THE BUN

Chapter II, is in effect, a screen which cuts off all
light except that of a single spectral line, and enables
the observer to see how the vapor of a single element
lies on the sun's surface.

We shall now examine some beautiful spectroheli-
ographic results obtained on Mount Wilson by Mr.
Ellerman, which Mr. Hale has kindly allowed me to
reproduce here. Plate V is taken with the spectro-
heliograph in the H2 line of calcium.1 Comparing it
with the direct photograph of the sun taken on the
same day, shown in Plate III, at the beginning of
this chapter, there is seen a greater distinctness and
prominence of detail. Hale has called the mo tt lings
shown by the spectroheliograph "flocculi, " and dis-
tinguishes between bright and dark flocculi. A pho-
tograph through the Ha (C) line of hydrogen, made
within a few minutes of Plate V, is given in Plate VI.
The hydrogen flocculi are generally of more well-
defined shapes than the calcium flocculi, and usually
dark where these are bright. Bright hydrogen floc-
culi, however, often appear in sun spot and active
regions, and such bright flocculi frequently change in
form with eruptive rapidity.

In a broad line, like the H or K lines of calcium, the
slit of the spectroheliograph may be set in several
positions. Hale distinguishes three such, which he
terms, H!, H2, H3 or Kb K2, and K3. In an eclipse

1 The faint structure of parallel lines seen on all spectroheliographic
plates is not a solar feature, but is caused by very slight irregu-
larities of the motion of the instrument.

118

THE PHOTOSPHERE

photograph of the chromosphere with radial slit (see
Chapter IV) the H and K lines have frequently an
" arrow head" appearance. That is: The light of
the center of H or K is found at a high level above the
sun, and the matter which produces the light of the
edges, or wings, does not extend out so far. H3 or
K3 corresponds to the center of H or K (seen as a
dark line in the solar spectrum). Thus, when we
look at a K3 spectroheliographic plate, there is a
deep layer of calcium vapor behind the regions shown,
and, as it takes but a small portion of this thickness
to cut off by absorption the light of this wave
length, our view is of the highest levels where cal-
cium occurs. The K2 and K! positions on the sides
and extreme wings of K, respectively, correspond to
moderate and low level calcium distribution. In
the spectrum of hydrogen a similar difference of
effective level in spectroheliograph observations
is attained by employing lines of different wave
lengths. In eclipse observations high hydrogen
prominences are red, owing to the predominance in
their light of rays of the Ha (C) line. Hence, pho-
tographs taken through the Ha (C) line give high
level phenomena, and, as might be plausibly inferred
from a consideration of Wien's displacement law,
the hydrogen lines of successively shorter wave-
lengths would be most copiously emitted at hotter,
and hence lower levels. We then regard an Ha or
K3 photograph as a high-level, an H£ or K2 as a
medium, and an H7 or Kj as a low-level phenomenon
10 119

THE SUN

for hydrogen and calcium respectively. However,
these gases are both high-level gases on the sun, and
the photographs of the sun through their lines are
above the levels where most Fraunhofer lines are
produced. It is to be expected that when, with in-
creasingly powerful instrumental appliances, the
spectroheliograph can be employed in the narrower,
and, therefore, more difficult, lines of the heavier
and less easily vaporized elements, the conditions at
lower levels will be shown.

The following illustrations bring out the differences
due to level in a striking manner. Unfortunately,
it was not possible for Mr. Ellerman to furnish me a
series showing all the different kinds of spectrohelio-
grams above mentioned for a single day, and, indeed,
it was found necessary to omit altogether an example
of H3 in the calcium series. Plates VII and VIII
show a spotted area of the sun's surface as it appeared
July 16, 1907, in Hx and H2 calcium spectrohelio-
graphic exposures. Plates IX, X, and XI illustrate
a spotted region of the solar surface as it appeared
September 10, 1909. They are taken in H2 of cal-
cium, HY and Ha of hydrogen, respectively. In
this latter series the first plate gives no hint of the
pronounced vortical structure revealed by the high
level hydrogen in the last plate. One is struck by
the similarity of these curved structural forms to the
lines-of-force diagrams given by the familiar experi-
ment of shaking fine iron filings on a glass plate held
horizontally over a couple of magnets. In Chapter

120

THE PHOTOSPHERE

V we shall have occasion to refer again to Plate XI
when we come to deal with the magnetic character
of the sun spots.

The spectroheliograph results will receive further
attention in Chapter IV in connection with the
study of solar prominences. These objects are great
flamelike protuberances which extend for thousands,
sometimes hundreds of thousands of miles above the
photosphere. First observed at eclipses, the fact
that they shine principally by the bright spectrum
lines of calcium and hydrogen made it possible to see
them at the sun's limbs at all times with the spec-
troscope, and now the spectroheliograph has enabled
us to recognize them frequently as dark hydrogen
flocculi on the disk itself. A view of the sun through
the Ha (C) line is best adapted for this purpose, and,
indeed, it may well be said to reveal the sun in quite a
new aspect. Direct photographs and spectrohelio-
graphic results through Hx (C) and H and K all
show a mottling of the solar surfaces, but in Plate
XI the mottling, especially in the neighborhood of
sun spots, shows a marked tendency toward curved
and spiral forms, as if the hydrogen at this high
solar level were definitely arranged by cyclonic mo-
tions. Still there are not usually found observable
motions along these curved lines, although in ex-
ceptional cases series of Ha spectroheliographic
plates have given evidence of definite and very rapid
motion. Thus St. John observing on Mount Wilson
on June 3, 1908, photographed a hydrogen flocculus,

121

THE SUN

probably a prominence, apparently moving 105,000
kilometers (60,000 miles) in 18 minutes towards a
double sun spot. When near the spot the flocculus
divided, and apparently each branch was sucked
into a sun spot. The apparent motion in this case
was almost exactly radial to the sun spot pair. A
dark flocculus of a similar type, which is also probably
a prominence, is seen in Plate VI.1

THE SOLAR ROTATION

The rotation of the sun has been measured by ob-
serving the march of sun spots, faculae, and, of late,
spectroheliographic flocculi across the disk. The
classical researches of . Carrington and of Spoerer on
the march of sun spots showed :

(1) That the sun rotates about an axis inclined
about 7° to the plane of the ecliptic, and so that the
sun's axis points midway between the polar star and
Vega to a position in right ascension 18h 44m and
declination 64°.

(2) At the solar equator the rotation occurs in
about 25 days.

(3) The period of one rotation increases on either
side of the equator about equally, and is about 27J/2
days at 45° north or south solar latitude.

(4) Individual sun spots drift in different directions
on the sun's surface, so that it is only the mean re-

1 An interesting conclusion relating to the part played by eruptive
prominences in the life history of sun spots is quoted in Chapter V
from Bpectroheliographio observations of Fox.

122

THE PHOTOSPHERE

suit of the motions of many spots which can give
accurately the solar rotation period.

(5) The daily rate of solar rotation, and the fact of
different rotation periods for different solar latitudes,
were both expressed by Carrington in the following
formula, in which X is the daily rate of rotation, I
the solar latitude :

X = 865' -- 165' sin1 I

Faye assuming on theoretical grounds that the
exponent of sin I should be 2, derived from Carring-
ton's observations of 1853-1861 the expression:

X = 862' - 186' sin2 1.

Spoerer, from observations of his own between
1862 and 1868, combined with those of Secchi and
others, obtained:

X = 1011' - 203' sin (41° 13' + /)
Tisserand from observations of 1874-1875 ob-
tained :

X = 857.6' - 157.3' sin2 1.

Wilsing and later Stratonoff have determined the
solar rotation from observations of faculae. As
these objects can seldom be followed much more than
a quarter way across the solar disk, and as their
appearance is usually altered when they reappear
on the other limb, the results have less weight than
those obtained by sun-spot observations. Wilsing
found no evidence of equatorial acceleration, but
Stratonoff found from the faculae similar results to
those of Carrington and Spoerer on sun spots. Very

123

THE SUN

recently Chevalier has published results of a long and
excellent series of determinations of the solar rota-
tion by measurements of faculae. His work con-
firms that of Stratonoff.

In 1908 Hale published determinations of solar
rotation from spectroheliographic plates of the hydro-
gen and calcium flocculi, taken through the HS and
H2 lines respectively. His results with H2 calcium
flocculi are in close agreement with those obtained
by Fox in 1903-4 for the same line. Their results
agree, also, at all latitudes with the rates of solar
rotation derived by various observers from observa-
tion of sun spots. With H8 hydrogen flocculi, the
rate of equatorial rotation was about the same, but
there was found no retardation at higher latitudes, a
fact of high interest and significance.

According to Doppler's principle the spectral lines
of a source receding must be displaced towards the
red with reference to those of a source approaching
the observer. By forming the solar image with a
telescope, and reflecting light from the two limbs si-
multaneously upon the slit of a spectroscope, two
spectra may be produced, one immediately above the
other, which exhibit at a glance the shifting of all
solar lines owing to the sun's rotation. See Plate
IV. Atmospheric lines are not thus shifted.

In this way the rate of solar rotation has been de-
termined with great accuracy by Duner, Halm, and
lately by Adams. Their results bring out clearly the
fact discovered by Carrington from the study of sun

124:

fij

THE PHOTOSPHERE

spots, namely, that the sun's angular rotation is
slower at high latitudes than at the sun's equator.
Since sun spots do not occur near the sun's pole, this
peculiarity could not be thoroughly studied by Car-
rington, but by the spectroscopic method the solar
rotation period has been determined at high as well
as low latitudes. There might easily be a doubt
whether the spectroscopic results on solar rotation
ought to agree with those obtained from observing
the sun spots, the faculae, and the flocculi exposed
by the spectroheliograph, for the sun, as indicated
by several lines of evidence, is at so high a tempera-
ture as to be probably almost wholly gaseous, and
our vision may penetrate to some distance below its
surface. The several objects which are sources of the
light phenomena employed for the different methods
of studying the solar rotation may lie at different
levels, and may, therefore, move at different rates.
Accordingly it is interesting to compare, in the fol-
lowing table, the rotation periods indicated by the
several visual methods and by the spectroscopic ob-
servations of lines of different chemical elements.
The table is compiled from those given by Hale l
and by Adams.2

According to the results referred to in the following
table, the sidereal rotation of the average solar sur-
face is completed in about 24.6 days at the equator,
26.3 days at ± 30° latitude, 31.2 days at ± 60°,

1 Contributions of the Mount Wilson Solar Observatory, No. 25.

2 Ibid., No. 33.

125

THE SUN

TABLE IX. — Daily rotation of the sun's surface,
Various methods of observing.

object
observed. -»

Sun spots.

Faculse.

Ca. Flocculi
H • line.

H. Flocruli
H5 line.

Many spec-
tral linos.

Observer. -»

Mean of
Carrington,
Spoerer,
Maunder.

Mean of
Stratonoff
and
Chevalier.

Mean of
ll:i!f and
' Fox.

Halo.

Adams, 100S.
(Doppler

effect.)

Latitude. ;

0° to ± 5°

14.40°

14 . 56°

14.54°

14.3°

14.59°

± 5 ±10

14.35

14.52

14.41

14.4

14.48

±10 ±15

14.25

14.33

14.30

14.6

14.33

±15 ±20

14.13

14.21

14.13

14.5

14.15

±20 ±25

13.98

14.19

13.99

14.7

13.95

±25 ±30

13.80

14.04

13.97

14.7

13.74

±30 ±35

13.60

13. 601

13.75

14.9

13 . 50

Stratonoff only.

Adams' spectroscopic results, including high latitudes.

Chemical
elements. -»

Many.

'ttSy?"

Fe, Ti,
TiFe.

Mn, Fe,
Fe.

Ca.

Wave

4196.699

4265.418

4257.815

4226 .91

6563 . 054

lengths. -»

Many.

4197.257

4287 .566

4290.542

(Ha or C).

Latitude. ;

4216.136

4288.310

4291 .6-30

0.3°

14.65°

14.49°

14.65°

14.72°

15.0°

15.2°

14.9

14.28

14.21

14.31

14.34

14.9

15.0

29.7

13.66

13.49

13.65

13.74

14.2

14.6

44.7

12.81

12.74

12.85

12.95

13.6

14.0

60.0

11.52

11.35

11.53

11.62

12.5

13.7

74.9

10.84

10.50

10.93

11.04

13.1

14.3

and 35.3 days at ± 80°. The agreement between
Adams' and Duner's work, done in different years,
is so exact that there seems little reason to suspect
a secular variation of the retardation towards high
latitudes. Adams finds his mean results and those
of Duner and Halm well expressed by the following
formula :

? = 10°.62 + 3°.99cos2 </>.
126

THE PHOTOSPHERE

Where £ is the angular sidereal rotation per day and
<#> the solar latitude.

The highly interesting change in the observed
rotation period for lines of different chemical ele-
ments is regarded as indicating differences of effec-
tive level of the production of the Fraunhofer lines.
The results in this direction gain added value be-
cause they agree with several other lines of evidence
that point to the same conclusions. The whole
subject will be discussed in Chapter VI.

Very recently St. John has determined the rate of
daily rotation spectroscopically in the K3 line of
calcium.1 He says: " The angular velocity of the
high-level calcium producing the absorption line K3
is nearly constant for the latitudes of observation,
being 15°.5 and 15°.4 per day at the latitudes 6°.6 and
38°. 4, respectively. The corresponding values de-
duced from Adams' results are 15°. 1 and 14°.3 for
hydrogen, and 14°.4 and 13°. 2 for the reversing layer.
The high velocity of the calcium vapor producing the
K3 line points to a higher elevation of this layer of
calcium vapor than of the hydrogen effective in the
production of the Ha line." It is a very singular
thing that calcium occurs at such very high levels in
the sun. We shall see the fact confirmed in the next
chapter, but the reason for it is one of those many
puzzles which whet the appetite of the student in
solar research.

1 Contributions of the Mount Wilson Solar Observatory, No. 48.
^ 127

CHAPTER IV

ECLIPSES AND THE OUTER SOLAR ENVELOPES

The Saros. — Eclipse Expeditions. — The Corona. — The Chromo-
sphere.— The Eclipse of 1868 and Jansen's and Lockyer's
Discovery. — Spectrum of the Chromosphere and Prominences. —
Prominences and the Spectroheliograph. — Recent Flash Spec-
trum Observations. — The Heights of Different Metals in the
Chromosphere. — Mitchell's Observations of 1905. — Campbell's
Observations. — Chromospheric Spectra in Full Daylight.

WHEN the moon passes directly between the earth
and the sun it sometimes completely covers the
latter, and there is a total solar eclipse. At such
times the brilliant glare of day ceases for a few mo-
ments to illuminate our atmosphere, and in the semi-
darkness we may see the objects which closely sur-
round the sun. Total solar eclipses occur almost
every year, but as the moon is never much greater in
angular diameter than the sun, the area of the earth's
surface on which the eclipse appears total at a given
instant is rarely greater than 100 miles in average
diameter. The rapid motion of the moon, though
partly offset by the rotation of the earth, hurries the
region of totality along faster than 1,000 miles an
hour, making a belt seldom wider than 100 miles, but
sometimes more than 5,000 miles long, on which the
eclipse is total at sometime between sunrise and sun-

128

ECLIPSES AND SOLAR ENVELOPES

set. Over enormous areas on either side of the line
of totality the sun is partially eclipsed, and appears
for some hours as a crescent figure.

THE SAROS

The ancients discovered a cycle of eclipses called
the Saros, which indicates approximately the times
when solar eclipses will occur. In 223 synodic
months there are almost exactly nineteen " eclipse
years" of 346.62 days, the interval between the times
when the sun in its apparent annual path crosses the
two nodes of the moon's orbit. Hence, if we count
forward 6,585 days, or 18 years 11 days, from one
total eclipse, we are apt to find the occurrence of
another either partial or total. A family of eclipses
thus occurs, separated by intervals of about eighteen
years. Such a family generally numbers about sixty-
five or seventy eclipses, of which perhaps eighteen
will be total, and the rest annular or partial. Many
total eclipses are visible only at regions unfavorable
for observation, such as oceans, the polar regions, or
very cloudy localities. As totality at a given place
never lasts more than eight minutes, and generally
does not exceed three, there have been hardly more
than a couple of hours of time employed in total solar
eclipse observing in the last half century. Yet so well
have the moments been utilized that a large stock of
information has been gathered.

Not infrequently eclipse expeditions have led as-
tronomers to experiences of hardship, disappointment

129

THE SUN

and in one instance to death. Father Perry, of Stony-
hurst, who led the English eclipse expedition to Cay-
enne, in 1889, was taken ill before the awaited day.
He insisted on observing, supported by an attendant,
and called for three cheers when the eclipse had been
successfully observed, saying: "I can't cheer, but I
will wave my helmet!" A few days later he died at
sea.

A total eclipse having been predicted to occur at a
certain place favorable for observing, astronomers
journey there several weeks in advance, equipped
with photographic telescopes, spectroscopes, auxil-
iary apparatus, and supplies. The instruments are
set up and carefully adjusted and are provided with
every possible contrivance to facilitate and shorten
their operation at the critical moment. Rehearsals
of the eclipse begin as soon as possible. Time signals
are counted off, photographic apparatus is manip-
ulated, and the whole program is gone over and over
again, just as if the totality were on. In this way
the observers try to anticipate all possible contin-
gencies, and acquire skill and rapidity in performing
their parts. Mr. Langley used to say that if a pin
were likely to be dropped during the eclipse the ob-
server should practice dropping one and filling its
place at rehearsal. The hour, minute, and second of
the eclipse are predicted long in advance, so that on
the appointed day all is prepared for action at a well
known time. At first contact of the moon a notch
begins to appear in the sun's disk, and this grows

130

ECLIPSES AND SOLAR ENVELOPES

larger and larger during the next hour and a half,
until only a narrow crescent remains. This hour
and a half has always seemed to the writer the saving
element; for during its slow passage the unhurried
march of events tends to calm the nervous agitation
which comes on with the first contact, when one feete
that his opportunity is now or never. As the cres-
cent becomes thin the sun's light becomes noticeably
weak and yellow, for only the limb now remains vis-
ible, and its light, as stated already, is very much
weaker, especially in the violet end of the spectrum,
than the light of the center of the disk. Just before
totality flickering bands, called " shadow bands,"
steal rapidly along the ground, and then, as the last
crescent of the photosphere suddenly vanishes, a thin
ring of rosy light encircles the moon, and beyond this,
for perhaps one or even two diameters of the sun,
blooms forth the pearly hued corona,

THE CORONA

There is a cycle of changes in the form of the
corona having a period of about eleven years, sup-
posed to be identical with that of sun-spot frequency,
which will be noticed in Chapter V. As the corona
can be observed only at total solar eclipses, the march
of the cycle of changes is as yet only imperfectly
known, but for the last half century it has been
observed that there are long equatorial coronal
streamers at the time of sun-spot minimum, while
at maximum of sun spots the corona extends only

131

THE SUN

to moderate distances, but nearly uniformly in all
directions from the sun. The accompanying views,
Plate XII from a drawing of Mr. P. R. Calvert pre-
pared from the Yerkes Observatory photographs of
the 1900 eclipse, Plate XIII from a drawing of Mrs.
C. G. Abbot prepared from U. S. Naval Observa-
tory photographs of the 1905 eclipse, illustrate the
characteristic forms of the corona at sun-spot
minima and maxima respectively.

Many efforts have been made, but thus far with-
out success, to devise a method of observing the
corona without an eclipse. Success is unlikely, for
in its brightest parts, even within TV radius of the
sun's limb, the brightness of the corona is only about
one-tenth as great as that of the daylight sky at
20° from the sun, if viewed from sea level. Close
to the sun the daylight sky is many fold brighter
still, so that the coronal brightness is insignificant
in comparison with it. By ascending a very high
mountain, it is true, a considerable gain might be
made, for the corona wrould be a little brighter and
the sky several fold less bright, but the brightness
of the sky would still be far too great to permit the
corona to be seen, even in its brightest parts, by any
contrivance yet devised.

The corona fades rapidly with increasing distance
from the sun. According to Turner, who has dis-
cussed results of various eclipses, it falls off approxi-
mately as the sixth power of the distance from the
sun's center. L. Becker has discussed photographic

132

ECLIPSES AND SOLAR ENVELOPES

observations made at the eclipse of 1905, and gives
the following formula of distribution of the intensity
of the blue and violet coronal radiation at different
distances, H, from the sun's limb; I is the intensity,
C is a constant, and H is expressed in thousandths
of a solar diameter.

I = C (H + 140) -4.

At the eclipse of January 3, 1908, the present writer,
assisted by A. F. Moore, made bolometric observa-
tions of the intensities of the coronal radiation at
several distances from the sun's limb. These were
made both with and without a screen of asphaltum
varnish on glass. This screen was used to cut off
the visible spectrum while still transmitting the
infra-red. The following is a comparison of these
results with those computed according to the for-
mulae of Turner and of Becker.

H =

45.

121.

364.

Total radiation
Visible radiation . . .

100
100

29.9

29.8

0
0

Infra-red radiation
Computed via Becker
Computed via Turner

100
100
100

30.1
25.2
45.7

0
1.8
6.3

The agreement between the bolometric observations
and the computation by Becker's formula is pretty
good, so that for a sun-spot maximum corona it
seems to represent the distribution for all kinds of
radiation, at least in the inner corona.

The light shows distinct radial polarization in the
133

THE SUN

outer corona, but the percentage of polarization de-
creases, and at length vanishes near the limb of the
sun. Polarization of the coronal light is generally
interpreted as evidence of the presence of reflected
photospheric rays in the coronal brightness, just as
sky-light and its polarization is produced by the dif-
fuse reflection of sunlight in the air. Some writers
have inferred from the absence of polarization near
the sun's limb that the light of that part of the corona
contains almost no reflected photospheric rays. But
a particle near the sun's limb must be shone upon
from every direction within a hemisphere, so that
the light which it reflects, being partially polarized
in every plane, would show polarization in none.
Hence, the absence of reflected photospheric rays
from the inner coronal brightness cannot rightly be
inferred from the absence of polarization.

The spectrum of the corona is more nearly con-
tinuous than that of the photosphere. ,A few bright
lines are found, but these are not conspicuous at most
eclipses. There is a famous bright coronal line in
the green at wave length 5303. This line was dis-
covered by Young, in 1870, and it has been seen with
more or less distinctness at many subsequent eclipses.
It does not correspond in wave length to a line of
any known substance, or to a photospheric line, so
that it is ascribed to a hypothetical element "coro-
nium." As the element helium was found in the
earth after its spectrum had long been known in the
sun and stars, so it may happen with "coronium."

134

PLATE XIII

SOLAR CORONA. 1905, AUGUST 30.

From Drawing by Mrs. C. G. Abbot from Photographs by the United States
Naval Observatory Eclipse Expedition.

ECLIPSES AND SOLAR ENVELOPES

Several bright coronal lines have been discovered
in the ultra-violet by Deslandres, Dyson, Lewis, and
others. In the outer corona the Fraunhofer lines
of the photospheric spectrum have been seen, and
have been repeatedly photographed by Campbell,
Perrine and others. Lewis found them only in the
ultra-violet spectrum in the eclipse of 1908. These
dark lines fade and disappear near the limb of the sun.
Their presence in the outer corona is a proof of the
presence in the outer coronal light of a large pro-
portion of reflected photospheric rays, but Campbell
infers from their absence near the sun's limb that
the inner corona shines almost wholly by light of
incandescence of the material there, due to its being
heated by proximity to the sun. There are, however,
several causes other than a great admixture of coronal
light of incandescence which must contribute to
diminish the distinctness of the Fraunhofer lines of
the inner corona near the sun's limb. Among these
are (1) atmospheric reflection of the strong bright
line spectrum of the chromosphere; (2) over ex-
posure of photographic spectra for the very inner-
most corona, etc.

It was inferred by Bigelow and by Holden, from
studies of eclipse photographs, that the corona par-
ticipates in the rotation of the sun. This view is con-
firmed by spectroscopic observations of Deslandres,
Campbell, and Belopolsky. It has been supposed by
many that, as the polar streamers of the corona ap-
pear much like terrestrial auroras seen in high north-
11 135

THE SUN

ern and southern latitudes, the corona may have, like
them, an electrical origin. They would regard its
light, like that of the aurora, as largely of lumines-
cence similar to that of a glow electrical discharge,
and not true temperature radiation. The writer's
bolometric observations of the inner corona at the
eclipse of 1908 seemed to be incompatible with the
view that it shines mainly by light of ordinary incan-
descence. For by the aid of absorbing screens it was
shown that the ratio of intensity of infra-red to total
radiation is almost the same for the inner corona as
for the photosphere. If the corona shines mainly by
incandescence, and its high temperature is produced
by the absorption of sunlight in its particles, then the
fraction of its radiation occurring in the infra-red
spectrum should be disproportionately greater than
that for the photosphere, because the temperature of
the corona must be much the lower. Lewis, however,
at the same eclipse found the ultra-violet coronal
rays disproportionately weaker than those of the
photosphere, and inferred therefrom a low coronal
temperature. The composition of the inner coronal
light cannot yet be regarded as settled. There
is undoubtedly some reflected light, some light
of incandescence, and perhaps some of lumines-
cence. It may be that it is the latter which is
the key to the perplexing observations above re-
corded. The nature of the corona will be further
discussed in Chapter VI.

136

ECLIPSES AND SOLAR ENVELOPES

THE CHROMOSPHERE

Close to the limb of the sun there is seen at total
solar eclipses, and by special contrivances also in full
sunlight, a thin ring of rosy light called the " chro-
mosphere, " from which project irregularly, some-
times as much as 50,000 or even 100,000 miles, rosy
forms called " prominences. " The spectrum of the
chromosphere consists of bright lines on a faint con-
tinuous background. These bright lines are the
counterparts in position, and generally, also, in rela-
tive intensity, of the dark Fraunhofer lines of the
photospheric spectrum. The prominences appear
to be but higher extensions of the chromosphere,
yet their spectra are usually simpler. Prof. C. A.
Young made a prolonged study of the prominences
and of their spectra, and I cannot introduce the mat-
ter better than to quote his descriptions (pages 197
to 226 of "The Sun") supplementing his story by
mention of most recent work on the subject. Young's
explanations of some of the phenomena differ some-
what from those which the present writer would pre-
fer, as Young was a believer in the cloudy photo-
sphere.

The Eclipse of 1868. Janssen's and Lockyer's Dis-
covery.

"Every one is more or less familiar with the story
of this eclipse. Herschel, Tennant, Pogson, Rayet,
and Janssen, all made substantially the same report.

137

THE SUN

They found the spectrum of the prominences to con-
sist of bright lines, and conspicuous among them were
the lines of hydrogen. There were some serious dis-
crepancies, indeed, among their observations, not
only as to the number of the bright lines seen, which
is not to be wondered at, but as to their position.
Thus, Rayet (who saw more lines than any one else)
identified the red line observed with B instead of C;
and all the observers mistook the yellow line they saw
for that of sodium.

"Still, their observations, taken together, com-
pletely demonstrated the fact that the prominences
are enormous masses of highly heated gaseous matter,
and that hydrogen is a main constituent.

"Janssen went further. The lines he saw during
the eclipse were so brilliant that he felt sure he could
see them again in the full sunlight. He was pre-
vented by clouds from trying the experiment the
same afternoon, after the close of the eclipse; but
the next morning the sun rose unobscured, and, as
soon as he had completed the necessary adjustments,
and directed his instrument to the portion of the
sun's limb where the day before the most brilliant
prominence appeared, the same lines came out again,
clear and bright; and now, of course, there was no
difficulty in determining at leisure, and with almost
absolute accuracy, their position in the spectrum.
He immediately confirmed his first conclusion, that
hydrogen is the most conspicuous component of the
prominences, but found that the yellow line must

138

ECLIPSES AND SOLAR ENVELOPES

be referred to some other element than sodium,1 being
somewhat more refrangible than the D lines.

"He found also that, by slightly moving his tele-
scope and causing the image of the sun's limb to take
different positions with reference to the slit of his
spectroscope, he could even trace out the form
and measure the dimensions of the prominences;
and he remained at his station for several days,
engaged in these novel and exceedingly interesting
observations.

"Of course, he immediately sent home a report of
his eclipse-work, and of his new discovery, but, as his
station at Guntoor, in eastern India, was farther from
mail communication with Europe than those upon
the western coast of the peninsula, his letter did not
reach France until some week or two after the ac-
counts of the other observers; when it did arrive, it
came to Paris, in company with a communication
from Mr. Lockyer, announcing the same discovery,
made independently, and even more creditably, since
with Mr. Lockyer it was not suggested by anything
he had seen, but was thought out from fundamental
principles.'

"Nearly two years previously the idea had oc-
curred to him (and, indeed, to others also, though he
was the first to publish it) that, if the protuberances
are gaseous, so as to give a spectrum of bright lines,
those lines ought to be visible in a spectroscope of

irThis element is helium and was discovered on the earth long
afterwards.

139

THE SUN

sufficient power, even in broad daylight. The prin-
ciple is simply this :

" Under ordinary circumstances the protuberances
are invisible, for the same reason as the stars in the
daytime: they are hidden by the intense light re-
flected from the particles of our own atmosphere near
the sun's place in the sky, and, if we could only suf-
ficiently weaken this aerial illumination, without at
the same time weakening their light, the end would
be gained. And the spectroscope accomplishes pre-
cisely this very thing. Since the air-light is re-
flected sunshine, it of course presents the same spec-
trum as sunlight, a continuous band of color crossed
by dark lines. Now, this sort of spectrum is greatly
weakened by every increase of dispersive power, be-
cause the light is spread out into a longer ribbon and
made to cover a more extended area. On the other
hand, a spectrum of bright- lines undergoes no such
weakening by an increase in the dispersive power of
the spectroscope. The bright lines are only more
widely separated — not in the least diffused or shorn
of their brightness. Moreover, if the gas is one which,
like hydrogen, shows dark lines in the ordinary solar
spectrum (and therefore in that of the air-light),
the case is even better: not only is the continuous
spectrum of the air-light weakened by the high dis-
persion, but it has dark gaps in it just where the
bright lines of the prominence spectrum will fall.

"If, then, the image of the sun, formed by a tele-
scope, be examined with a spectroscope, one might

140

ECLIPSES AND SOLAR ENVELOPES

hope to see at the edge of the disk the bright lines
belonging to the spectrum of the prominences, in case
they are really gaseous.

"Mr. Lockyer and Mr. Huggins both tried the ex-
periment as early as 1867, but without success; partly
because their instruments had not sufficient power to
bring out the lines conspicuously, but more because
they did not know whereabouts in the spectrum to
look for them, and were not even sure of their exis-
tence. At any rate, as soon as the discovery was an-
nounced, Mr. Huggins immediately saw the lines
without difficulty, with the same instrument which
had failed to show them to him before. It is a fact,
too often forgotten, that to perceive a thing known
to exist does not require one half the instrumental
power or acuteness of sense as to discover it.

"Mr. Lockyer, immediately after his suggestion
was published, had set about procuring a suitable
instrument, and was assisted by a grant from the
treasury of the Royal Society. After a long delay,
consequent in part upon the death of the optician
who had first undertaken its construction, and partly
due to other causes, he received the new spectroscope
just as the report of HerscheFs and Tennant's ob-
servations reached England. Hastily adjusting the
instrument, not yet entirely completed, he at once
applied it to his telescope, and without difficulty
found the lines, and verified their position. He imme-
diately also discovered them to be visible around the
whole circumference of the sun, and consequently

141

THE SUN

that the protuberances are mere extensions of a con-
tinuous solar envelope, to which, as mentioned above,
was given the name of Chromosphere. (He does not
seem to have been aware of the earlier and similar
conclusions of Arago, Grant, Secchi, and others.)
He at once communicated his results to the Royal
Society, and also to the French Academy of Sciences,
and, by one of the curious coincidences which so
frequently occur, his letter and Janssen's were read
at the same meeting, and within a few minutes of
each other.

"The discovery excited the greatest enthusiasm,
and in 1872 the French Government struck a gold
medal in honor of the two astronomers, bearing their
united effigies.

" It immediately occurred to several observers,
Janssen, Lockyer, Zollner, and others, that by giving
a rapid motion of vibration or rotation to the slit of
the spectroscope it would be possible to perceive the
whole contour and detail of a protuberance at once,
but it seems to have been reserved for Mr. Huggins
to be the first to show practically that a still simpler
device would answer the same purpose. With a
spectroscope of sufficient dispersive power it is only
necessary to widen the slit of the instrument by the
proper adjusting screw. As the slit is widened, more
and more of the protuberance becomes visible, and,
if not too large, the whole can be seen at once : with
the widening of the slit, however, the brightness of
the background increases, so that the finer details of

l \->

ECLIPSES AND SOLAR ENVELOPES

the object are less clearly seen, and a limit is soon
reached beyond which further widening is disadvan-
tageous. The higher the dispersive power of the
spectroscope the wider the slit that can be used, and
the larger the protuberance that can be examined as a
whole — within certain limits, however. It is not
difficult with our latest spectroscopes, diffraction
instruments espe-
cially, to reach a
dispersion so great
that even the C
line becomes broad
and hazy, like the
b lines in an ordi-
nary instrument. FlG- 27.— HUGGINS'S FIRST OBSERVATION
OF A PROMINENCE IN FULL SUNSHINE.

In that case each

luminous point in the prominence itself is represented
in the image of the prominence, not by a point, as it
should be to give clear definition, but by a streak at
right angles to the spectrum lines.

Spectrum of the Chromosphere and Prominences.

" The spectra of the chromosphere and prominences
are very interesting in their relations to that of the
photosphere, and present many peculiarities which
are not yet fully explained. At times and in places
where some special disturbance is going on — fre«
quently in the neighborhood of spots at the times
when they are just passing around the limb of the
disk — the spectrum, at the base of the chromosphere,

143

THE SUN

is very complicated, consisting of hundreds of bright
lines. In the course of a few weeks of observation at
Sherman in 1872, the writer made out a list of two
hundred and seventy-three, and more recent observa-
tions have added largely to the number — at least
fifty lines within the limits of the visible spectrum,
and, by photography, at least eighty in the ultra-
violet. The majority of the lines, however, are seen
only occasionally, for a few minutes at a tune, when
the gases and vapors, which generally lie low, mainly
in the interstices of the clouds which constitute the
photosphere, and below its upper surface, are ele-
vated for the time being by some eruptive action.
For the most part, the lines which appear only at such
times are simply " reversals" of the more prominent
dark lines of the ordinary solar spectrum. But the se-
lection of the lines seems most capricious; one is
taken, and another left, though belonging to the same
element, of equal intensity, and close beside the first.
It is evident that the subject needs a detailed and
careful study, combining solar observations with
laboratory-work upon the spectra of the elements
concerned, before a satisfactory account can be given
of all the peculiar behavior observed.

"The lines composing the true chromosphere spec-
trum, if we may call it so (that is, those which are
always observable in it with suitable appliances) , are
not very numerous, and we give the following list,
designating them by their wave length, as given by
Rowland :

144

ECLIPSES AND SOLAR ENVELOPES

7065

.50.

Helium.

6563

.05,

Hydrogen

(Ha).

5875,

,98,

D3. (close double) .

Helium.

5316

.87.

4861

50,

Hydrogen

(H/B).

4471

,80,

/•

Helium.

4340.

66,

g (near G).

Hydrogen

(H7).

4101.85,

Hydrogen

(US).

3970

.20

(in H).

Hydrogen

(H«).

10.

3968

.56,

Calcium.

11.

3933

.86,

Calcium.

"The first line is generally very difficult to see,
though sometimes pretty conspicuous. It is in the
led, between B and a, and has a very faint corre-
sponding dark line. No. 3 has no dark line corre-
sponding as a usual thing, though occasionally one
appears, especially in the neighborhood of sun spots.
No. 9 is quite within the broad shade of the H-line,
which thus appears double in the chromosphere spec-
trum.

"The eleven lines mentioned above are invariably
present in the spectrum of the chromosphere; a much
larger number make their appearance on very slight
provocation. They are:

1'.

6678

Helium.

11'.

5183.

bi.

Magnesium.

2'.

6431

Iron.

12'.

5172,

6,.

Magnesium.

3'.

6141

Barium.

13'.

5169

&•.

Iron.

4'.

5896

2, D,

Sodium.

14'.

5167.

64.

Magnesium.

5'.

5890

2, Da

Sodium.

15'.

5018

(}.

Iron.

6'.

5363

Iron. ?

16'.

5015

Helium.

5284

Titanium? ?

17'.

4934

Barium.

8'.

5276

Chromium. ?

18'.

4924

Iron.

9'.

5234

Manganese.

19'.

4922

Helium.

10'.

5198

? ?

20'.

4919

Iron. ?

145

THE SUN

21'.

4900

Barium.

28'.

4236

Iron.

22'.

4584

Iron.

29'.

4233

Iron.

23'.

4501

Titanium.

30'.

4226

Calcium.

24'.

4491

Manganese.

31'.

4215

Stront-dm.

25'.

4490

Manganese.

32'.

4077.9.

Strontium.

26'.

4469

Iron.

33'.

4026

Helium.

27'.

4245

Iron.

34'.

3889

Hydrogen

(HO-

"It is not intended, however, to intimate that, if
one of these appears, all of them will do so, nor that
they are equally conspicuous or equally common. To
a certain degree, also, their selection by the writer is
arbitrary, for there are nearly as many more which
are seen pretty frequently, and some of them may
very possibly be found hereafter to deserve a place
upon the list rather than some that have been in-
cluded.

"It requires careful manipulation to bring out the
fainter and finer lines satisfactorily. The slit must
be adjusted with extreme care to the focal plane of
the rays under examination, placed tangential to the
solar image, and brought exactly to the edge of the
disk. A thousandth of an inch in its position will
often make the whole difference between a successful
operation and its failure, and even a slight unsteadi-
ness of the air will diminish the number of bright
lines visible by at least one half.

"As the majority of the lines are developed only by
more or less unusual disturbances of the solar surface,
it naturally happens that one very often finds them
distorted or displaced by the motions of the gases
along the line of sight (toward or from the observer) ,

146

ECLIPSES AND SOLAR ENVELOPES

as explained in a previous chapter, producing what
Lockyer calls " motion-forms. " Occasionally, also,
we meet with " double reversals, " so called, especially
in the lines of magnesium and sodium. The (dark)
lines of these substances are rather wide in the solar

in the chromosphere

FIG. 28. — DOUBLE REVERSAL, OF
THE D-LINES. (October, 1880.)

spectrum. When reversed
spectrum, the phenomenon
usually consists of a thin
bright line down the center
of the wider dark band : in
a double reversal the bright
line widens and a fine dark
line appears in its center, so
that we have a central dark
line, a bright one on each
side of it, and outside of the
bright lines a dark shade on both sides. Fig. 28
represents such a double reversal of the D-lines
observed by the writer on several occasions in
1880. The phenomenon seems to be due to the pres-
ence of an unusual quantity of the vapor at a consid-
erable density, and is the precise correlative of what
is sometimes seen in the spectrum of a sodium-flame.
The two D-lines of sodium each becomes itself double,
so that we get pairs of bright lines in place of single
lines. The electric arc often shows this still more
finely.

"At the base of a prominence, the C, F, H, and K
lines are always thus doubly reversed. Fig. 29 is from
a recent photograph of the C-line obtained at Prince-

147

THE SUN

ton, by Mr. Reed, with the large telescope and spec-
troscope. The slit was tangential to the sun's limb.

Of course, an
isochromatic
plate and a
long exposure
were required
to get such
an impression
from the
"ruby light"
of that part
of the spec-
trum. When
the slit is ail-
justed to cross
the sun's limb radially the bright lines where they
project beyond the spectrum of the photosphere
assume the "arrow-headed'' form shown in Fig. 30.
" Generally speaking, the spectrum of a prominence
is simpler than that of the chromosphere at its base.
We seldom find any
lines except C, D3, F,
gy h, H and K, at a
considerable elevation
above the photosphere,
though / is sometimes
met with. On rare
occasions, also, the vapors of sodium and magnesium
are carried into the higher regions, and once or twice

148

FIG. 29. — DOUBLE REVERSAL
(Photographed.)

Fu;. .SO.

ECLIPSES AND SOLAR ENVELOPES

the writer has seen the line No. 1 of the second list
(6678.2) in the upper portions of a prominence.

Observation of Prominences.

" When the spectroscope is used as a means of ren-
dering visible the forms and features of the promi-
nences, the only difference is that the slit is more or
less widened.

"The telescope is directed so that the solar image
shall fall with that portion of its limb which is to be
examined just tangent to the
opened slit, as in Fig. 31,
which represents the slit-plate
of the spectroscope, with the
image of the sun in position
for observation.

"If, now, a prominence ex- FIG. 31.— OPENED SLIT OP

. . . ' . . THE SPECTROSCOPE.

ists at this part of the sun s

limb (as would probably be the case, considering the
proximity of the spot shown in the figure), and if
the spectroscope itself is so adjusted that the C-line
falls in the center of the field of view, then, on looking
into the eyepiece, one will see something much like
Fig. 32. The red portion of the spectrum will stretch
athwart the field of view like a scarlet ribbon, with a
darkish band across it, and in that band will appear
the prominences, like scarlet clouds — so like our own
terrestrial clouds, indeed, in form and texture, that
the resemblance is quite startling : one might almost
think he was looking out through a partly opened

149

THE SUN

door upon a sunset sky, except that there is no variety
or contrast of color; all the cloudlets are of the same
pure scarlet hue. Along the edge of the opening is
seen the chromosphere, more brilliant than the clouds
which rise from it or float above it, and for the most
part made up of minute tongues and filaments. Usu-
ally, however, the definition of the chromosphere is
less distinct than that of the higher clouds. The
reason is, that close to the limb of the sun, where the
temperature and pressure are highest, the hydrogen
is in such a state that the lines of its spectrum are
widened and " winged, " something like those of mag-
nesium, though to a less extent. Each point in the
chromosphere, therefore, when viewed through the
opened slit, appears not as a point, but as a short line,
directed lengthwise in the spectrum. As the length
of this line depends upon the dispersive power of the
spectroscope, it is easy to see that it is possible to go
too far in this respect. The lower the dispersion the
more distinct the image obtained, but also the fainter
as compared with the background upon which it is
seen.

"Just beneath the chromosphere (at a in the cut)
the appearance is as if the edge of the sun was dark, a
phenomenon which for some time was very puzzling.
Its explanation lies in the "double reversal" of the
C-line at the base of the chromosphere, discussed and
figured a few pages back.

"If the spectroscope is adjusted upon the F-line,
instead of C, then a similar image of the prominences

150

ECLIPSES AND SOLAR ENVELOPES

and chromosphere is seen, only blue instead of scarlet;
usually, however, since the F-line is hazier and more
winged than C, this blue image is somewhat less per-
fect in its details and definition, and is therefore less
used for observation. Similar effects are obtained by
means of the yellow line near D, and the violet line near
G. With suitable precautions, using a violet shade-
glass before the
eye, and carefully
shutting out all
extraneous light,
the H and K lines
can also be used;
but visual observa-
tions in this part
of the spectrum are
extremely difficult
and unsatisfactory.

"With photog- FIG. 32. — CHROMOSPHERE AND PROMINENCES

, , , . AS SEEN IN THE SPECTRUM.

rapny the case is

the reverse — these lines are then precisely those
which can be employed most easily and conveniently.
We shall recur to this a little later.

" Professor Winlock and Mr. Lockyer have at-
tempted, by using an annular opening instead of the
ordinary slit, to obtain a view of the whole circum-
ference of the sun at once, and have succeeded. With
a spectroscope of sufficient power, and adjustments
delicate enough, the thing can be done ; but as yet no
very satisfactory results appear to have been reached.
12 151

THE SUN

We still (in visual observations) have to examine the
circumference piecemeal, so to speak, readjusting the
instrument at each point, to make the slit tangen-
tial to the limb.

"The number of protuberances of considerable
magnitude (exceeding ten thousand miles in altitude),
visible at any one time, on the circumference of the
sun, is never very great, rarely reaching twenty-five
or thirty. Their number, however, varies extremely
with the number of sun spots: during a sun-spot
minimum there are not unfrequently occasions when
not a single one can be found, though even during
those years the, more usual number is five or six —
some of which often are of considerable size. The
observations of Tacchini and Secchi have showed that
their numbers closely follow the march of the sun
spots though never falling quite so low.

"To Tacchini we owe our most complete record of
these objects, now continuous since 1872, giving their
number and distribution upon the sun, with drawings
of all that were specially remarkable. Many others
have cooperated in observations of this kind: the
Hungarian observers, Fenyi at Kalocsa, and Von
Gothard at Hereny, have given us many fine descrip-
tions and delineations. Father Perry and his assis-
tant Sidgreaves, at Stonyhurst, also deserve a special
mention.

"Their distribution on the sun's surface is in some
respects similar to that of the spots, but with impor-
tant differences. The spots are confined within 40°

152

ECLIPSES AND SOLAR ENVELOPES

of the sun's equator, being most numerous at a
solar latitude of about 20° on each hemisphere.
Now, the protuberances are most numerous pre-
cisely where the spots are most abundant, but
they do not disappear at a latitude of 40°; they
are found even at the poles, and from the latitude

FIG. 33. — RELATIVE FREQUENCY OF PROTUBERANCES AND SUN-SPOTS.

of 60° actually increase in number to a latitude
of about 75°.

"The annexed diagram, Fig. 33, represents the
relative frequency of the protuberances and spots on
the different portions of the solar surface. On the
left side is given the result of Carrington's observa-
tion of 1,386 spots between 1853 and 1861, and on
the right the result of Secchi's observations of

153

THE RFX

2,707 l protuberances in 1871 . The lengl h of each ra-
dial line represents the number of spots or protuber-
ances observed at each particular latitude on a scale
of a quarter of an inch to the hundred ; for example,
Secchi gives 228 protuberances as the number ob-
served during the period of his work between 10° and
20° of south latitude, and the corresponding line
drawn at 15° south, on the left-hand side of the figure
is therefore made Iff or .57 of an inch long. The
other lines are laid off in the same way, and thus the
irregular curve drawn through their extremities rep-
resents to the eye the relative frequency of these phe-
nomena in the different solar latitudes. The dotted
line on the right-hand side represents in the same
manner and on the same scale the distribution of the
larger protuberances, having an altitude of more than
1', or 27,000 miles.

"A mere inspection of the diagram shows at once
that, while the prominences may, and in fact often
do, have a close connection with the spots, they are
yet to some extent independent phenomena.

"A careful study of the subject shows that they
are much more closely related to the faculse.2 In
many cases, at least, facula3, when followed to the

1 The 2,767 prominences are not all different ones. If any of the
prominences observed on one day remained visible the next, they
were recorded afresh; and, as a prominence near the pole would be
carried but slowly out of sight by the sun's rotation, it is thus easy
to see how the number of prominences recorded in the polar regions
is so large.

'See page 109 [of Young].

154

ECLIPSES AND SOLAR ENVELOPES

limb of the sun, have been found to be surrounded by
prominences, and there is reason to suppose that the
fact is'a general one. The spots, on the other hand,
when they reach the border of the sun's image, are
commonly surrounded by prominences more or less
completely, but seldom overlaid by them. Indeed,
Respighi asserts (and the most careful observations
we have been able to make confirm his statement)
that as a general rule the chromosphere is consider-
ably depressed immediately over a spot. Secchi, how-
ever, denies this.

Magnitude and Classification of Prominences.

"The protuberances differ greatly in magnitude.
The average depth of the chromosphere is not far
from 10" or 12", or about 5,000 or 6,000 miles, and it
is not, therefore, customary to note as a prominence
any cloud with an elevation of less than 15" or 20"—
7,000 to 9,000 miles. Of the 2,767 already quoted,
1,964 attained an altitude of 40", or 18,000 miles, and
it is worthy of notice that the smaller ones are so few,
only about one third of the whole: 751, or nearly one
fourth of the whole, reached a height of over 1', or
28,000 miles; the precise number which reached
greater elevations is not mentioned, but several ex-
ceeded 3', or 84,000 miles. It is only rather rarely
that they reach elevations as great as 100,000 miles.
The writer has in all seen, perhaps, three or four
which exceeded 150,000 miles, and Secchi has re-
corded one of 300,000 miles. On October 7, 1880,

155

THE SUN

the writer observed one which attained the still un-
equaled height of over 13' of arc, or 350,000 miles.
When first seen, on the southeast limb of the sun,
about 10.30 A.M., it was a "horn" of ordinary appear-
ance, some 40,000 miles in elevation, and attracted
no special attention. When next seen, half an hour
later, it had become very brilliant and had doubled
its height : during the next hour it stretched upward
until it reached the enormous altitude mentioned,
breaking up into filaments which gradually faded
away, until, by 12.30 P.M., there was nothing left. A
telescopic examination of the sun's disk showed noth-
ing to account for such an extraordinary outburst,
except some small and not very brilliant faculse.
While it was extending upward most rapidly a violent
cyclonic motion was shown by the displacement of
the spectrum lines, and H and K were reversed
through its whole height.

"In their form and structure the protuberances
differ as widely as in their magnitude. Two princi-
pal classes are recognized by all observers — the qui-
escent, cloud-formed or hydrogenous, and the eruptive
or metallic. By Secchi these are each further sub-
divided into several sub-classes or varieties, between
which, however, it is not always easy to maintain the
distinctions.

"And here perhaps is the proper place to mention
that Trouvelot insists on the existence of "dark"
prominences — i.e., clouds of cooler hydrogen • that
absorb the light of the hydrogen behind them; but

156

ECLIPSES AND SOLAR ENVELOPES

Three figures of the same prominence,
seen July 25, 1872.

FIG. 34.

AS SEEN AT 2.15 P. M.

FIG. 38.
SHEAF AND VOLUTES.

FIG. 35.
AS SEEN AT 2.45 P. M.

FIG. 30.

AS SEEN AT 3.30 P. M.
Scale, 100,000 miles to the inch.

ERUPTIVE PROMINENCES
157

THE SUN

there is no proof, we think, that these are anything
but " holes. " Tacchini, on the other hand, is disposed
to assert the existence of " white" prominences, which
give a continuous spectrum, and so are not reached
by spectroscopic observation, though conspicuous to
the eye, and on the photographic plate, at the time of
a total eclipse, as in 1883 and December, 1889. But
the evidence hardly warrants confident belief in the
existence of such objects.

"The quiescent prominences in form and texture
resemble, with almost perfect exactness, our terres-
trial clouds, and differ among themselves as much and
in the same manner. The familiar cirrus and stratus
types are very common, the former especially, while
the cumulus and cumulo-stratus are less frequent.
The protuberances of this class are often of enormous
magnitude, especially in their horizontal extent (but
the highest elevations are attained by those of the
eruptive order), and are comparatively permanent,
remaining often for hours and days without serious
change; near the poles they sometimes persist
through a whole solar revolution of twenty- seven
days. Sometimes they appear to lie upon the limb
of the sun like a bank of clouds in the horizon; prob-
ably because they are so far from the edge of the disk
that only their upper portions are in sight.* When
seen in their full extent they are ordinarily connected
to the underlying chromosphere by slender columns,
which are usually smallest at the base, and appear
often to be made up of separate filaments closely in-

158

ECLIPSES AND SOLAR ENVELOPES

FIG. 43.
DIFFUSE.

FIG. 41.

FlLAMENTAKY.

FIG. 44.
STEMMED.

FIG. 42.
PLUMES.

QUIESCENT PROMINENCES.

Scale, 7o,000 miles to the inch.

159

THE SUN

tertwined, and expanding upward. Sometimes the
whole under surface is fringed with down-hanging
filaments, which remind one of a summer shower fall-
ing from a heavy thundercloud. Sometimes they
float entirely free from the chromosphere ; indeed, as
a general rule, the layer clouds are attended by de-
tached cloudlets for the most part horizontal in their
arrangement.

" The figures give an idea of some of the general ap-
pearances of this class of prominences, but their del-
icate, filmy beauty can be adequately rendered only
by a far more elaborate style of engraving.

"Their spectrum is usually very simple, consisting
of the four lines of hydrogen, and the three of helium,
with H and K. Occasionally the sodium and mag-
nesium lines also appear, and that even near the sum-
mit of the clouds; and this phenomenon was so much
more frequently observed in the clear atmosphere of
Sherman as to suggest that, if the power of our spec-
troscopes were sufficiently increased, it would cease
to be unusual.

"The genesis of this sort of prominence is problem-
atical. They have been commonly looked upon as
the debris and relics of eruptions, consisting of gases
which have been ejected from beneath the solar sur-
face, and then abandoned to the action of the cur-
rents of the sun's upper atmosphere. But near the
poles of the sun distinctively eruptive prominences
never appear, and there is no evidence of aerial cur-
rents which would transport to those regions matters

160

ECLIPSES AND SOLAR ENVELOPES

ejected nearer the sun's equator. Indeed, the whole
appearance of these objects indicates that they orig-
inate where we see them. Possibly, although in the
polar regions there are no violent eruptions, there
yet may be a quiet outpouring of heated hydrogen
sufficient to account for their production — an out-
rush issuing through the smaller pores of the solar
surface, which abound near the poles as well as
elsewhere.

"But Secchi reports an observation which, if cor-
rect, puts a very different face upon the matter.1
He has seen isolated cloudlets form and grow spon-
taneously without any perceptible connection with
the chromosphere or other masses of hydrogen, just
as in our own atmosphere clouds form from aqueous
vapor, already present in the air, but invisible until
some local cooling or change of pressure causes its
condensation. These prominences are, therefore,
formed by some local heating or other luminous ex-
citement of hydrogen already present, and not by any

1 On October 13, 1880, the writer for the first time met with the
same phenomenon. A small, bright cloud appeared on that day,
about 11 A. M., at an elevation of some 2^' (67,500 miles) above the
limb, without any evident cause or any visible connection with the
chromosphere below. It grew rapidly without any sensible rising or
falling, and in an hour developed into a large stratiform cloud,
irregular on the upper surface, but nearly flat beneath. From this
lower surface pendent filaments grew out, and by the middle of the
afternoon the v object had become one of the ordinary stemmed
prominences, much like Fig. 44.

But obviously the thing is very unusual, for in more than twenty
years of observation I have encountered the phenomenon only three
times.

161

THE SUN

transportation and aggregation of materials from a
distance. The precise nature of the action which
produces this effect it would not be possible to assign
at present ; but it is worthy of note that the spectro-
scopic observations made during eclipses rather favor
this view, by showing that hydrogen, in a feebly
luminous condition, is found all around the sun, and
at a very great altitude — far above the ordinary range
of prominences.

" Indeed, in most cases the forms and changes of
this class of prominences so closely resemble our own
terrestrial clouds that one is almost forced to believe
that they are surrounded by, and float in, a medium
which does not greatly differ from themselves in den-
sity, though it is not visible in the spectroscopic mode
of observation.

Eruptive Prominences.

"The eruptive prominences are very different-
much more brilliant and much more vivacious and
interesting. They consist usually of brilliant spikes
or jets, which change their form and brightness very
rapidly. For the most part they attain altitudes of
not more than 20,000 or 30,000 miles, but occasion-
ally they rise far higher than even the largest of the
clouds of the preceding class. Their spectrum is very
complicated, especially near their base, and often
filled with bright lines, those of sodium, magnesium,
barium, iron, and titanium, being especially conspicu-
ous, while calcium, chromium manganese, and prob-

ECLIPSES AND SOLAR ENVELOPES

FIG. 49.

PROMINENCE AS IT APPEARED AT HALF-
PAST TWELVE O'CLOCK, SEPTEMBER
7, 1871.

FIG. 46.

VERTICAL FILAMENTS.

FIG. 47.
CYCLONE.

FIG. 50.

As THE ABOVE APPEARED HALF AN HOUR
LATER WHEN THE UP-RUSHING HYDROGEN
ATTAINED A HEIGHT OF MORE THAN 200,-
000 MILES.

FIG. 51.

SPOT NEAR THE SUN'S LlMB, WITH ACCOM-
PANYING JETS OF HYDROGEN, AS SEEN
FLAMES. OCTOBER 5, 1871.

Scale, 75,000 miles to the inch.

163

THE SUN

ably sulphur, are by no means rare, and for this
reason Secchi calls them metallic prominences.

"They usually appear in the immediate neighbor-
hood of a spot, never occurring very near the solar
poles. Their form and appearance change with great
rapidity, so that the motion can almost be seen with
the eye — an interval of fifteen or twenty minutes
being often sufficient to transform, quite beyond rec-
ognition, a mass of these flames fifty thousand miles
high, and sometimes embracing the whole period of
their complete development or disappearance. Some-
times they consist of pointed rays, diverging in all
directions, like hedgehog-spines. Sometimes they
look like, flames ; sometimes like sheaves of grain ;
sometimes like whirling waterspouts, capped with a
great cloud; occasionally they present most exactly
the appearance of jets of liquid fire, rising and falling
in graceful parabolas; frequently they carry on their
edges spirals like the volutes of an Ionic column; and
continually they detach filaments which rise to a
great elevation, gradually expanding and growing
fainter as they ascend, until the eye loses them. Our
figures present some of the more common and typical
forms, and illustrate their rapidity of change, but
there is no end to the number of curious and interest-
ing appearances which they exhibit under varying
circumstances.

"The velocity of the motions often exceeds a hun-
dred miles a second, and sometimes, though very
rarely, reaches two hundred miles. That we have to

164

ECLIPSES AND SOLAR ENVELOPES

do with actual motions, and not with mere change of
place of a luminous form, is rendered certain by the
fact that the lines of the spectrum are often displaced
and distorted in a manner to indicate that some of the
cloud-masses are moving either toward or from the
earth (and, of course, tangential to the solar surface)
with similar swiftness.

" Fig. 52 is a representation of a portion of the spec-
trum of a prominence observed at Sherman on August

FIG. 52.

3, 1872, an observation to which allusion was made in
the preceding chapter. The F-line, at 208 of the
scale, must be imagined as blazingly brilliant, and
fainter bright lines appear at 203.2, 208.8, 209.4, and
212.1 (the scale is KirchhofTs), while two bands of
continuous spectrum, produced probably by the com-
pression of the gas at the points of maximum dis-
turbance, run the whole length of the figure. At the

165

THE SUN

upper point of disturbance F is drawn out into a point
reaching to 207.4 of the scale, and indicating a veloc-
ity of 230 miles a second away from us; at the lower
point it extends to 208.7, and indicates a velocity of
about 250 miles per second toward us. It was very
noticeable that this swift motion of the hydrogen did
not seem to carry with it many other substances
which were at the time represented in the spectrum
by their bright lines; magnesium and sodium were
somewhat affected, but barium and the unknown ele-
ment of the corona were not. "

An examination of the sun's limb for prominences
is made on every fair observing day at many observa-
tories. At the Italian observatories of Rome and
Catania such observations have been continued by
Secchi, Tacchini, and Ricco for about forty years. A
general discussion of this highly valuable mass of
observations is about to be published.

Prominences and the Spectroheliograph.

Since the introduction of the Spectroheliograph the
prominences can be observed much more satisfactor-
ily than before. In Plate XIV, Fig. 1 shows a large
quiescent prominence photographed by Slocum with
the Rumford Spectroheliograph at the Yerkes Ob-
servatory. The height of the prominence as shown
in the plate is 1.6 minutes of arc, or 69,000 kilometers.
Slocum states that this prominence lasted probably
continuously for at least fifty-five days in the spring
of 1910, but Evershed traces it twenty-seven days

166

PLATE XIV.

FIG. 1. — 1910, March 17. G. M. T. oh 30m. LON. 7°. LAT. + l/° TO -18°

FIG. 2. — 1910, October 10. G. M. T. 7h 56m .8.

FIG. 3. — 1910, October 10. G. M. T. 8h 6m .4.
SOLAR PROMINENCES. (Slocum.) CALCIUM (H) SPECTROHELIOGRAMS.

ECLIPSES AND SOLAR ENVELOPES

longer still. Its southern extremity remained nearly
stationary near 20° south latitude, while the northern
end varied greatly; from the equator on March 4, to
25° north latitude on March 18; then, retiring,
reached 10° south latitude on April 28. At no time
could the prominence be seen projected against the
sun's disk in Slocum's calcium spectroheliographic
observations. He saw it only on the sun's limbs.
But Evershed and Deslandres photographed it re-
peatedly; the former in H2 calcium, the latter in K3
calcium and Ha hydrogen light, appearing like a long
cloud upon the sun's disk. A similar feature is shown
in the Ha photograph reproduced in Plate VI from
Ellerman's Mount Wilson observations of April 30,
1908.

Very beautiful eruptive prominences are occa-
sionally observed with the spectroheliograph. The
two lower figures in Plate XIV show an uncommon-
ly fine quasi-eruptive prominence photographed Oc-
tober 10, 1910, in the H line of calcium, by Slocum at
the Yerkes Observatory. Although by no means as
active as some eruptive prominences, this one
changed rapidly, and there may be seen considerable
differences in form in the two exposures, separated
by a time interval of only ten minutes. The approxi-
mate position was as follows: Solar latitude 24° to
39° S; longitude 225°. Height 2.5 minutes of arc, or
108,000 kilometers.

13 167

THE SUN

Recent Flash-spectrum Observations.

Following the great discovery of Janssen and Lock-
yer in 1868, the next year brought the important dis-
covery of helium in the sun — a chemical element not
found on the earth for nearly thirty years afterwards.
Young kept up the pace by the discovery of the
" flash spectrum" at the total solar eclipse of 1870.
Setting the slit of his spectroscope where the chromo-
sphere should be, and keeping his eye prepared for
what he was about to witness, he saw, as the last
photospheric rays were extinguished, a bright line
reversal of the photospheric spectrum flash out to
view. It was not till 1896 that the flash spectrum
was photographed by Shackelton with a prismatic
camera.

The chromosphere appears as a very thin crescent,
hence its spectrum may be photographed without
slit or collimator. The appearance of such spectra
may be understood from Plate XV, Fig. 1, taken by
S. A. Mitchell at the eclipse of 1905. The spectrum
lines are each represented by arcs of circles. Where
very long arcs appear, they correspond to the great
lines of hydrogen and calcium. These elements ex-
tend much higher above the sun than others, and
hence continue in sight longer as the moon advances.

It was feared that the astigmatism, which makes
the concave grating a valuable laboratory instrument,
would render it unfit for use on the flash spectrum
without a slit. A slit for such work is undesirable,

168

ECLIPSES AND SOLAR ENVELOPES

on account of the loss of light. The work of Mitchell
in 1898, who used for the photography of stellar spec-
tra a Rowland concave grating as an objective grat-
ing, without slit, paved the way for the use of grat-
ings at the time of an eclipse. They were first used
in flash-spectrum photography in 1900.

Successful photographic observations of the flash
spectrum have been made at the eclipses of 1896,
1898, 1900, 1901, 1905, and 1908. Among the observ-
ers have been Shackelton, Campbell, Evershed,
Dyson, Jewell, Frost, Lord, S. A. Mitchell, Perrine,
and others. The observations have shown that the
flash spectrum, or spectrum of the chromosphere, is
essentially the reversal of the ordinary Fraunhofer
spectrum, but with some significant differences.
Many of the weaker Fraunhofer lines, of course, do
not appear. The lines of the two spectra in general
bear different relative intensities. Taking the lines of
any one chemical element by itself, however, the rel-
ative intensities in the two spectra are not very dif-
ferent. Lockyer, Evershed, and Dyson find in gen-
eral that the so-called enhanced or spark lines are
more prominent in the flash spectrum than in the
photospheric spectrum.1 The cause of the dis-
crepancy between the line intensities in the spectra,
as a whole, seems to be that the elements of higher
atomic weights are less prominent in flash spectra.

Dyson, in discussing the Greenwich observations

1 Frost and Mitchell were inclined to question that this is general,
but Mitchell seems now to agree that it is.

169

THE SUN

of the eclipses of 1900, 1901, and 1905, gives the meas-
ured positions of about 1200 lines, and identifica-
tions of most of them with single lines, or with blends
of several lines, found in Rowland's tables. The
range of spectrum observed is from 3295 Ang-
stroms, in the ultra-violet, to 5896 in the orange. The
average deviation of the positions from the positions
fixed by Rowland is 0.04 Angstroms, but as Dyson's
spectra are prismatic this difference is exceeded in
the green and yellow. Dyson found twenty-six strong
lines of hydrogen agreeing excellently in position
with the places fixed by Balmer's series formula.
Helium is also a prominent element. The following
are the chemical elements as they are found repre-
sented by their spectral lines :

Very strong: Hydrogen, Helium, Magnesium,
Calcium, Scandium, Titanium, Chromium, Stron-
tium.

Strong: Manganese, Iron, Yttrium, Zirconium,
Barium, Lanthanum, Cerium, Erbium, Europeum.

Not very strong: Carbon, Aluminum, Vanadium,
Neodymium.

Very weak : Nickel, Cobalt, Lead.

Possibly shown: Zinc, Lanthanum, Tantalum.

Doubtful: Silicon, Gadolinium, Prsesodymium.

Absent:1 Argon, Neon, Krypton.

Not well shown within limits of spectrum : Sodium.

The arc lines of aluminum, magnesium, barium,

1 Mitchell, however, inclines to think these elements are repre-
sented in the flash spectrum by weak lines.

170

ECLIPSES AND SOLAR ENVELOPES

zinc, and lead appear to be present, whereas their
enhanced, or spark, lines show not at all, or faintly.
In this Dyson finds these elements exceptional, for in
general it is the enhanced lines which predominate in
the flash.

The Heights of Different Metals in the Chromosphere.
By measuring the lengths of the arcs seen as flash
spectrum lines, observers have estimated the heights
to which the elements rise in the chromosphere above
the sun's general surface. From Sir Norman Lock-
yer's report of observations of the eclipse of 1898, we
have the following values. The ordinary chemical
symbols for the elements are used for short :

Element

Var-
ious

Spectrum

Not

4471

4078

4227

U.V.

3944

Quar-

Many

Flut-

Many

lines

given

trip-

tet

ing

lines

let

Includ-

4027

4216

3962

4031

lines

ing Fe

etc.

arc

lines

Mean

height

seconds. . .

13.3

7.5

6.0

4.4

3.2

2.4

3.2

1.05

to 1.4

Kilometers .

9700

7200

5400

4300

3200

2300

1800

2300

760

to 1000

Jewell,1 from observations of the eclipses of 1900
and 1901, estimates the chromospheric heights cor-
responding to separate lines of various elements. He
finds, as does Lockyer (See Ca above), that different
lines of the same elements yield widely different val-
ues. Thus, for calcium his heights range from 15,000

lPub. U.S. Naval Observatory, 2 Series, Vol. IV, Ap. I.
171

THE SUN

down to 100 miles, and for titanium from 3,500 to 100
miles. In general his results show high levels for
hydrogen, helium, parhelium, magnesium, sodium,
and ytterbium; low levels for chromium, iron, cobalt,
nickel, manganese, yttrium, cadmium, zinc, carbon
(as cyanogen), and vanadium; contradictory levels
indicated by different lines for calcium, strontium,
barium, scandium, and titanium. Most lines corre-
spond to heights of less than one second of arc (475
miles, 760 kilometers). Jewell regards the chromo-
sphere as an atmosphere of hydrogen and a few other
permanent gases, rapidly decreasing in density out-
ward, and holding as temporary constituents other
elements as products of eruptions from within, or
meteors from without.

Frost and Mitchell, from observations of the 1900
and 1901 eclipses, respectively, have also given brief
tables of the heights attained by different elements in
the chromosphere, as indicated by individual spec-
trum lines. Their results differ very little from those
above mentioned. Mitchell states that the lengths
of a great majority of the lines indicate heights not
exceeding 0.5" of arc, and would set I" of arc as the
average depth of the " re versing layer."

Jewell very pertinently calls attention to the mi-
nute quantities of substance required to produce spec-
trum lines. As some lines require less producing
substance than others, this may cause part of the dis-
crepancy between the heights estimated for different
lines of the same element.

ECLIPSES AND SOLAR ENVELOPES

Mitchell's Observations of 1905.

My friend, Prof. S. A. Mitchell, has kindly fur-
nished me, in advance of his publication, with the
following description of his apparatus, and of the
results he obtained as a member of the U. S. Naval
Observatory expedition, at the total eclipse of
August, 1905. His flash spectrum is believed to be
the best which has ever been secured.

"Mitchell used two spectrographs of high disper-
sion, both with gratings. The first was a six-inch
Rowland plane grating of 15,000 lines to the inch,
belonging to the Naval Observatory. This same grat-
ing had been used by him in Sumatra, in 1901, but
in 1905 a glass achromatic objective of five inches
aperture was used instead of a quartz lens. With this
instrument special attention was paid to the red end
of the spectrum. The other instrument was a four-
inch grating, ruled on a parabolic surface, instead of
the ordinary spherical concave surface. This grating
of 14,438 lines per inch and ten feet radius of curva-
ture was very bright in the first order on one side, and
in the estimation of Mr. Jewell it was one of the best
of Rowland gratings, and gave spectra equal in
brightness to that obtained by the ordinary six-inch
grating. This grating belonged to the Rumford com-
mittee, and was kindly loaned by Professor F. A.
Saunders of Syracuse University.

"Such a spectrograph used for eclipse work is of
the simplest form imaginable. Light from the coelo-

173

THE SUN

stat mirror, reflected horizontally, falls on the grating,
where it is diffracted, and is then brought to focus on
the photographic plate, five feet distant. Grating and
plate holder are placed in a wooden box, and if the
grating and photographic plate are perpendicular to
the diffracted beam, the spectrum is "normal." As
the spectrum was brought to a focus on a circle of
thirty inches radius, it was impossible to bend the
photographic plate, and heavy gelatine films were
used. The spectra were focussed by using a col-
limating apparatus consisting of a slit between two
concave mirrors (which were previously adjusted by
the use of a five-inch visual telescope). The spectra
were focussed visually, and test photographs were
made in order to check the ultra-violet focus. The
excellence of the focus is shown by the flash spectra,
which were photographed in the first order.

"The parabolic grating spectra extend from X 3,300
in the ultra-violet to the D lines at X 5,890 in the
orange. The plane grating spectrogram continues
in the red to the C line. The length of the spectrum
taken with the former grating is 9.5 inches. The spec-
trum is very nearly normal throughout its whole
extent; the dispersion, therefore, is such that one
millimeter is equal to 10.8 Angstrom units. This is a
dispersion about equal to that obtained by the three-
prism spectrographs attached to the great Lick or
Yerkes telescopes. As the grating at the eclipse was
used as an objective grating without slit, it had a dis-
persion a little less than a quarter of that obtained

174

ECLIPSES AND SOLAR ENVELOPES

with a 21.5 foot grating in the first order on the or-
dinary Rowland mounting.

"As is seen from the illustration (Plate XV, Fig. 1,
where unfortunately a great amount of fine detail is
lost in reproduction) the definition is excellent. The
extreme ultra-violet is not in quite so good a focus as
the region from K to D, where the definition is per-
fect. About 4,000 lines were measured in the region
from X 3,300 to X 5,900. On account of the strong con-
tinuous spectrum throughout the photographed spec-
trum, it was a little difficult to see the spectrum lines,
especially when they were faint. The spectrum lines
being curved, it was necessary to measure at the same
part of each line. Moreover, since no slit was used, it
was necessary to measure the position of the line at
the moon's edge. Evidently the height above the
sun's limb of the metallic vapor forming a given spec-
trum line has much to do with its appearance on the
photographic plate, and the middle of the measured
line will not give its exact wave length.

"Preliminary wave lengths of the flash lines were
directly obtained from the measures. These were
compared with Rowland's tables. Each line from
Rowland which was identified with certainty was
taken as a standard to obtain adjusted values of the
flash wave lengths. The smaller dispersion hi the
flash spectrum caused lines in Rowland to be blended
together; and in such blends it was difficult to know
what exact wave length to assume. Consequently, if
the flash lines were identified with single lines in

175

THE SUN

Rowland's tables, they were taken as standards.
Since the scale-value assumed was only an approxi-
mate one, and the spectrum was not strictly normal, a
Least Squares adjustment was made in a manner
suggested by Professor C. Runge, Kaiser Wilhelm
Professor at Columbia University in 1909-10. As the
result of this adjustment, the probable error of a
single determination of a wave length throughout the
spectrum is about ± 0.025 Angstrom units. The
small size of this error will be appreciated when one
remembers that a series of cusps were measured, and
that an error in measurement of a thousandth of a
millimeter, or one micron, corresponds to a discrep-
ancy of 0.01 A. U.

"Such accurate wave lengths of the flash spectra
lend the possibility of a close comparison with the
Fraunhofer spectrum. Such a comparison shows with
great certainty that the flash spectrum is but a rever-
sal of the Fraunhofer lines. Almost every line in the
ordinary solar spectrum with an intensity of 3 or
greater on Rowland's scale is found in the flash spec-
trum, in many cases two or more lines being blended
into one in the photograph of smaller dispersion.
Though all the strong Fraunhofer lines are found in
the flash spectrum, the converse is not true; for
there are many strong lines in the flash spectrum
which have no equivalent in the ordinary spectrum.
In addition to this fact, there is the further difference
that there are remarkable inequalities in intensity
between the lines of the two spectra.

176

ECLIPSES AND SOLAR ENVELOPES

"It was pointed out by Evershed that we could
easily imagine two separate gases in the sun's enve-
lope which would have absorption lines of the same
intensities, but whose emission spectra would differ
very much in intensities. A heavy gas, lying in a
thin layer above the photosphere, might absorb the
solar light exactly to the same extent as a less dense
layer extending to greater altitudes. As the moon
successively passes over layers at the time of an
eclipse, the lighter gas would give lines of the greater
intensities in the flash spectrum. As is well known,
the helium lines appear as strong lines in the flash
spectrum; they are lacking in the Fraunhofer spec-
trum. Over thirty lines of the hydrogen series have
been counted in Mitchell's 1905 spectra. In Plate
XV, Fig. 2, a portion of the spectrum is shown greatly
enlarged.

"For the purpose of a closer comparison, the re-
sults of the measures of an extent of the spectrum of
62 Angstrom units to the red side of HS, i.e. from
XX 4, 102-4, 164 are given in the following table.

"In the above region, where ninety-two lines in the
flash were measured, there are eighty-two lines in
Rowland's tables of an intensity 2 and greater. Of
these eighty-two lines, but one is with certainty lack-
ing from the flash spectrum, the Fe line (intensity 4)
at X 4,154.976. Of the ninety- two flash lines, all have
been identified with the exception of a few faint lines.
The remarkable accuracy of the wave lengths of this
flash spectrum, which far surpasses any results hith-

177

THE SUN

TABLE X. — Measures of ninety-two lines in the flash spectrum near H$.

Flash Spectrum

Wave
Length
Rowland

Numbe
of Lines
Blended

Sub-
stance

Inten
sity
and
Char-
acter

Remarks

In-
ten-
sity

Wave
Length

4102.00

4102.000

40 N

4103.10

4103.097

Si, Mn

4103.65

4103.622

4104.27

4104.288

4104.65

4104.623

Co, V,

4105.21

4105.245

— , V

4106.49

4106.502

4107.64

4107.649

Ce-Fe-Zr

4108.68

4108.687

4109.371

4109.215)

4109. 88 j

4 109. 609 1

Nd?

4109 905

4110.63

4110.691

4111.62

4111.97

41 11. '940

4112.45

4112.478

4112.89

4112.869

4113.24

4113.183

Fe, Mn

4114.00

4114.73

4ii4.'769

Fe,' —

4115.35

4115.330

4116.14

4116.138

4116.78

4116.738

V.'Nd?

4118.02

4118.008

4118.85

4118.852

Fe.'Co

4119.53

4119.550)

4119.74

4119.751V

4120.12

4120.075)

4120.35

4120.368

4120.93

(4120.973)

Helium line at 4 120. 973

4121.46

4121.477

Cr-Co

6d?

4122.02

4122.049

Fe, Ti, Cr

4122.80

4122.819

4123.45

4123.477

La, Mn

4123.93

4123.907

4124.96

4124.938

4125.93

4125.900

Fe,' —

4126.35

4126.344

4126.66

4126.673

4127.86

4127.872

4128.25

4128.251

Ce-V

6d?

4128.91

4128.894

4129.41

4129.448

Ce' —

4129.88

4129.882

4130.83

4130.804

4131.46

4132.16

ii 09 .'idol

4132.23o|

2V
10 /

4133.05

4133.062

4133.93

4133.908

Fe, Ce

4134.49

4134.492

4134.84

4134.840

2d?

4135.56

4135.529

4136.02

4136.69

4i3e!678

4137.26

4137.156)
4137. 567 j

6)
2J

4137.79

4137.809

Fe.'Ce

4138.31

4138.324

178

ECLIPSES AND SOLAR ENVELOPES

TABLE X. — Continued.

Flash Spectrum

Wave
Length
Rowland

Number
of Lines
Blended

Sub-
stance

Inten-
sity
and
Char-
acter

Remarks

In-
ten-
sity

Wave
Length

4139.08

4139.008

4139.57

4140.24

4i40.'245

Fe,' —

4141.81 ]

4141.809

4142.03

4142.025

4142.56 J

4142.542

Cr, —

4143.28

(4143.30)

Nd line at 4143. 30

4144.05

(4143.919) )

Helium line at 4143.919

4144.038 J

4144.63

4144.674

ONd?

4145.13)

4145.152

4145.37]

4145.357

4145.84

4145.840

4146.23

4146.225

4147.12

4147.145

4147.69

4147.713

Mn, Fe

4148.98

4148.948

4149.37

4149.360

4149.923)

[dentification doubtful

4150.03

4150. 056 /

00 1

4150.40

4150.411

4150.68

4151.18

4isi.'i29

Zr.'Ti

4152.23

4152.248

La, Fe, Ce

4152.68

4153.51

4i53.'542

4154.09

4154.112

Cr, Fe

5 '

4154.65

4154.667

4156.30

4156.339

Nd, Zr

4157.00

4156.970

3d?

4158.00

4157.948

4159.00

4158.959

4159.40

4159.353

4160.57

4160.53

rlasselburg gives V

4160.57

4161.23)

4161.239

4161. 65 J

4161.682

Spark line Ti

4162.79

4162.724

4163.82

4163.818

Ti.'Cr

Spark line Ti

erto published makes the identification of lines a
practical certainty. Hence, it must be concluded
that the flash spectrum is a reversal of the Fraun-
hofer spectrum, but with marked differences in the
intensities in the two spectra.

" Measures of the 1905 spectra confirm Mitchell's
1901 results that hydrogen (H), helium (He), scan-

179

THE SUN

dium (Sc), titanium (Ti), strontium (Sr), vanadium
(V), Zr, Y, Cr, Mn, Nd and Ce appear with a greater
intensity in the flash than in the photospheric spec-
trum, relative to the other elements. These eclipse
results also confirm the prominence of enhanced
lines."

Campbell's Observations.

Professor Campbell has invented, and used success-
fully at several recent eclipses, a spectroscope with a
moving plate. He begins exposures slightly before
totality comes, and as the plate keeps falling, the
spectra are produced in a continuous series at deter-
minable times; and thus are adapted to give the
whole history of the spectrum, as it changes from the
photospheric spectrum (reflected by the air) to the
chromospheric, or flash spectrum. It is well known
that Professor Campbell and other members of the
Lick expeditions have secured spectra of very high
excellence with this and other apparatus in recent
eclipses, of which the discussion is not yet com-
pleted. Professor Campbell's full publication, and
that of Mitchell, are awaited with keen interest.

CHROMOSPHERIC SPECTRA IN FULL DAYLIGHT.

Recently Adams, at the Mount Wilson Solar Ob-
servatory, has obtained many photographs of the
chromospheric spectrum in full sunlight. This
method surpasses in accuracy of wave-length meas-
urements, and may eventually rival in detail the best

180

ECLIPSES AND SOLAR ENVELOPES

eclipse " flash spectra." Adams' observations were
made with the 60-foot-focus tower telescope, and the
- 30-f oot-f ocus plane-grating spectroscope. Success de-
pended on securing excellent definition of the sun's
image, so that the spectroscope slit might be held
exactly to the edge of the limb, without the light of
the photosphere ''boiling over," so as to blot out the
bright-line spectrum. The chromosphere is a stratum
so thin that it is covered by the march of the moon at
an eclipse in a very few seconds. Accordingly, there
is not sufficient time during total eclipses for the ex-
posure of a slit spectrograph of high dispersion, and
for this reason slitless spectrographs of moderate dis-
persion have usually been employed. Consequently,
it is not practicable to get from eclipse "flash spectra''
such high precision of wave lengths as is necessary to
decide the subtler points regarding the condition and
nature of the chromosphere. Hence, the great advan-
tage of supplementing eclipse work by observations at
great dispersion in full sunlight, especially for the red
end of the spectrum, where photography requires long
exposures. Mr. Hale proposes to continue the work
begun by Adams with the 60-foot tower telescope,
and is making provision for an enlarged solar image.
He expects greatly enriched results when the 150-foot
tower telescope is available.

In the work thus far published by Hale and Adams
the number of bright lines shown is far less than
Mitchell obtained at the eclipse of 1905. Certain dif-
ferences seem to indicate that the level of the spectra

181

THE SUN

photographed in full daylight is a little above the
level best observed at eclipses. The wave length of
the bright lines found seem to be practically identical
with the wave lengths of corresponding dark lines in
the photosphere. This circumstance, as Hale and
Adams remark, does not lend support to Julius's con-
tention, which will be noted in a later chapter, that
the bright lines of the "flash spectrum" are due to
anomalous refraction of light just outside the sun's
limb. For, if this were the case, there would probably
be a shifting towards the red of their apparent wave
lengths from those of the dark lines of the photo-
sphere. But Julius thinks the margins of discrepancy
between the positions of the bright and dark lines, as
given by Hale and Adams, still leave ground for his
theory of anomalous dispersion. In order to permit
of this interpretation, however, Julius imagines "the
solar atmosphere to be honeycombed with irregular
density gradients, which may be steeper than the
underlying general radial gradient." Thus he finds
the possibility that displacements of the chromo-
spheric lines may be, now to the red, now to the vio-
let, of the Fraunhofer lines. Most observers still
retain the view that the chromospheric spectrum is
essentially the reverse of the Fraunhofer spectrum
and appears at the edge of the sun bright instead of
dark because there is no such enormously brilliant
spectrum background to dim, by comparison, the in-
trinsic brightness of the lines themselves.

182

CHAPTER V

SUN-SPOTS, FACUL.E, AND GRANULATION

Sun-spot Periodicity. — Drift. — Distribution of Sun-spots. — For-
mation and Life History. — Sun-spot Level. — Langley's Typical
Sun-spot. — Faculae. — Granulation. — Sun-spot Spectra. — Cool-
ness of Sun-spots. — Sun-spots and Magnetism. — Radial Motion
in Spot Penumbras.

ALTHOUGH occasionally seen, and recorded much
earlier without recognition of their solar origin, the
history of sun-spots as solar phenomena dates from
1610, when they were independently discovered by
Fabricius, Scheiner, and Galileo. The discovery fol-
lowed naturally from the invention of the telescope
in Holland, in 1608. There was at first some doubt
(not shared by Fabricius or Galileo) whether the sun-
spots were not planets. Indeed, sun-spots were for a
time called in France the "Bourbonian Stars."

Viewed in a telescope, or projected on a screen, the
sun-spots are plainly seen, and appear to consist of
two well-marked parts; the umbra, apparently very
dark, and the penumbra, a half-tone border around
the umbra. Sun-spots differ greatly in size, shape,
and darkness. Some large ones are TO of the sun's
diameter, or five times the diameter of the earth, and
sun-spot groups occasionally spread over an area of
u 183

THE SUN

more than ro the sun's diameter. These great spots
and spotted areas are rare.

SUN-SPOT PERIODICITY

Schwabe of Desau about 1843 discovered, as the
result of systematic observing for nearly twenty
years, that there is a periodicity in the occurrence of
sun-spots. They are most frequent at intervals of
about eleven years, and are nearly absent for a year
or two in the interim. This sun-spot periodicity was
exhaustively studied by Wolf of Zurich, who repre-
sented the spottedness by a system now called
" Wolf's sun-spot relative numbers. " These are com-
puted by the formula, r = k(Wg +/), in which r is
Wolf's number, g the number of groups and single
spots observed, / the total number of spots which can
be counted in these groups and single spots combined,
and k a multiplier which depends on the conditions
of observation and the telescope employed. Wolf took
k as unity for himself when observing with a three-
inch telescope and a power of 64. A less favored or
less assiduous observer would receive k greater
than unity, and one with a larger telescope and good
opportunities for observing would receive a fractional
value of k. Wolf's numbers seem arbitrary, but are
found by photographic comparisons to be closely pro-
portional to the spotted areas on the sun. One hun-
dred as a sun-spot number corresponds to about TTFIT
of the sun's visible disk covered by spots, including
both umbras and penumbras.

184

SUN-SPOTS, FACUL^, AND GRANULATION

Wolf, by consulting all available sources, carried
his sun-spot numbers back to 1610. His successor,
Wolfer, has kept up the series from Wolf's death, in
1893, up to the present time. In Fig. 53 and Fig. 54
the curves show the run of spottedness during all this
interval.1 It will be seen that the maxima and min-
ima are not uniformly spaced ; but so that, while a
mean sun-spot interval of 11.13 years is deduced by
Professor Newcomb, the individual periods range
between 7.3 and 17.1 years as extremes. These fea-
tures are shown in the table on page 186.

Newcomb finds the average time of increasing
spottedness, 4.62 years, of decreasing spottedness,
6.51 years. Having studied the total interval from
1610 to 1898 in three parts, he concludes that: " Un-
derlying the periodic variations of spot activity
there is a uniform cycle, unchanging from time
to time, and determining the general mean of the
activity."

The reader will note in the sun-spot curves that
there is not only a great dissimilarity in the lengths of
the individual periods, but also of their activities as
measured by the maximum number of spots observed.
Dr. Lockyer pointed out a relation between these
phenomena which has been mentioned also by Halm
and thoroughly confirmed by Wolfer. Call the time

1 In Fig. 53 the inclusion of new data leads to a modification as
follows:

Year 1800 1801 1802 1803 1804 1805 1806

Mean. . .15.0 33.7 44.1 43.0 46.8 42.5 27.3
These data are corresponding to the mid-years.
185

THE SUN

TABLE XI. — Years of sun-spot maxima and minima and maximum
intensities

Minima

Difference

Maxima

Difference

Maximum
Wolf Number

1610.8

1615.5

1619.0

'8'.2

1626.0

io's

1634.0

15.0

1639.5

13.5

1645.0

11.0

1649.0

9.5

1655.0

10.0

1660.0

11.0

1666.0

11.0

1675.0

15.0

1679.5

13.5

1685.0

10.0

1689.5

10.0

1693.0

8.0

1698.0

8.5

1705.5

12.5

1712.0

14.0

1718.2

12.7

1723.5

11.5

1727.5

9.3

1734.0

10.5

1738.7

11.2

1745.0

11.0

1750.3

11.6

'si

1755.2

11.2

1761.5

11.2

1766.5

11.3

1769.7

8.2

103

1775.5

9.0

1778.4

8.7

151

1784.7

9.2

1788.1

9.7

133

1798.3

13.6

1805.2

17.1

1810.6

12.3

1816.4

11.2

1823.3

12.7

1829.9

13.5

1833.9

10.6

1837.2

7.3

137

1843.5

9.6

1848.1

10.9

125

1856.0

12.5

1860.1

12.0

1867.2

11.2

1870.6

10.5

132

1878.9

11.7

1883.9

13.3

1889.6

10.7

1894.1

10.2

1901.6,

12,0

1906.4

12.3

interval from a minimum to the following maximum a,
and that from the maximum to the following mini-
mum b. The variations of a and of the ratio - proceed

in a sense contrary to the intensity of the outbreak of
spottedness in each period. In other words, the more
intense the outbreak of spots in any sun-spot period
the shorter the time required for its development,

186

SUN-SPOTS, FACUL.E, AND GRANULATION

187

THE SUN

both actually and as compared with the time re-
quired for its decay.

It is interesting to observe, also, that the interval
from minimum to maximum is always much less than
that from maximum to minimum. Attention will be
drawn in Chapter X to the similarity between this
characteristic and a certain type of stellar variation
exemplified in the star Mira.

The causes which produce sun-spots being as yet
doubtful, or perhaps it is better to say entirely un-
known, the causes of their periodicity and of the ir-
regularity of the periods are, of course, also unknown.
Attempts have been made to connect the period with
the times of revolution of the planets, and, indeed,
the mean length of the sun-spot period is not far from
the period of the revolution of Jupiter (11.86 years).
No satisfactory case for a connection between these
phenomena is yet made out. Schuster has recently
applied a method of mathematical analysis fitted to
bring out secondary periods which may underlie the
average sun-spot periodicity of 1 1 . 1 3 years. He finds
three well-marked periods of 11.125 years, 8.32 years,
and 4.77 years. As curious mathematical coinci-
dences he notes that the sum of the reciprocals of the
first two periods equals the reciprocal of the third,
and all three are nearly even fractions of 33|
years. He finds the relative intensities of spot-
tedness for his three periods variable, hence the
inequality of the successive total periods produced
by their combination. He inclines to attribute

188

SUN-SPOTS, FACUL.E, AND GRANULATION

sun-spots to causes outside the sun, perhaps to
meteor swarms.

Several kinds of phenomena, some solar, others ter-
restrial, are evidently closely associated with sun-
spots and share in their periodicity. Firstly, the f acu-
lae, or bright flecks on the solar surface, which are
always seen most plentifully in sun-spot neighbor-
hoods, naturally have the same seasons of maxima
and minima. Secondly, the prominences, as stated
in the preceding chapter, are most numerous at sun-
spot maxima, and decrease in number, though not
with so marked a change, as the number of sun-spots
decreases. Thirdly, the form of the solar corona evi-
dently goes through a periodic change simultaneous
with the sun-spot cycle. Thus we speak of a solar
corona having prolonged equatorial streamers of an
arrow-head shape as "a sun-spot minimum corona,"
and one nearly equally developed in all directions as
"a sun-spot maximum corona." Fourthly, the ter-
restrial auroras (northern and southern lights) follow
the sun-spot periodicity, as shown by Loomis and
many others. Fifthly, changes in the earth's magnetic
field occur in complete synchronism with the changes
of sun-spot numbers. This connection is very close,
for the agreement descends even to minute parallel-
ism, as shown by the magnetic curves plotted in Fig.
53 and Fig. 54. Great sun-spots often seem to be the
direct promoters of great magnetic disturbances
(magnetic storms) and auroral displays. Maunder
has found that the magnetic disturbances seem to

189

THE SUN

arise from restricted solar areas, not necessarily
including sun-spots, and to go out in definite
directions, or rather shafts of several degrees
diameter, which rotate with the sun.1 When such
a shaft strikes the earth a magnetic storm arises.
Such lines of influence are not, he thinks, neces-
sarily radial, but may follow coronal stream lines.
Sixthly, the earth's surface air temperature is on the
whole lower at sun-spot maximum than at sun-spot
minimum. This relation is indicated at least for
the United States in Fig. 54. The difference of mean
temperature for the earth generally, ranging from 0°.5
to 1°.0 Centigrade for a change of 100-sun-spot num-
bers, is shown by temperature statistics studied by
Koppen, Nordmann, Newcomb, Abbot, Fowle, Arc-
towski, and Bigelow. This will be further discussed in
Chapter VII. Many other terrestrial changes, in
rainfall, cloudiness, number of cyclones, panics, prices
of foods, famines, growth of trees, even flights of in-
sects, have been seriously compared with the sun-
spot changes. In some of these phenomena there
appear to be rather well-substantiated indications of
a periodicity coincident with that of sun-spots, while
such relations in many cases are probably purely fan-
ciful.

So-called " great periods" of 33, 35, 55, and several
hundred years have been proposed by various au-
thors in order to explain the variations of the lengths

1 Shearman of Toronto discovered also a periodicity of auroral
displays approximating the rotation period of the sun.

190

SUN-SPOTS, FACUL^E, AND GRANULATION

\ '

890 2 4 6 6 1900

MPERATURES AND MAGNETISM.

i States inland stations,
's).
in diurnal range )
s, Chree. j

— -

)

;

-f

.^r

{

(

X-

»•

n ^

'-c

a"" I860 2468

FIG. 54. — SUN-SPOTS AND TERRESTRIAL Ti
I. Temperature departures, Unite(
II. Sun-spot relative numbers (WoH
III. Magnetic declination f meg
IV. Magnetic horizontal force I Elli

ff~~>

'' 5

./-

•

>
i>

;

•v-

• —

)

*• —

•—x

— —

1 —

1 r

^r-

-s4-*

D m o in o
- *<3 *c> *o* '— •
li + +

o o o

SO "fr CO

! s s s

§555

191

THE SUN

and intensities of the eleven year periods, and the
changes of rainfall, times of harvest, and other
changes of terrestrial phenomena said to be indicated
by history or tradition. The tendency to groups of
three in respect to the intensity of the successive sun-
spot outbreaks has been mentioned by various writ-
ers, and may be noted in the table. This helps to sus-
tain belief in a thirty-three year period, but it will be
noted that the four maxima 1830 to 1870 were un-
commonly intense (perhaps excepting the intermedi-
ate one of 1860). The question of the reality of "great
periods" seem to require further lapse of years to de-
cide it.

SUN-SPOT DRIFT

If we imagine an observer on the moon to watch the
clouds on the earth's surface, they would appear to
him on the whole to indicate a mean rotation period
of about twenty-four hours for the earth. But he
would also discover that many, and perhaps nearly
all, of the cloudy areas had proper motions of their
own besides, so that no single cloud would give cor-
rectly the rotation period of the earth. So it is with
the sun-spots, for, after allowing for the sun's average
rotation period, nearly every spot has a motion of its
own. Carrington found a slight tendency of spots be-
tween 20° North and 20° South latitudes to approach
the equator, and outside these latitudes a more de-
cided tendency to approach the poles. Faye held
that spots persistently describe little ellipses on the
sun's surface of one or two days' period. It is said

192

SUN-SPOTS, FACUL^E, AND GRANULATION

that an actively changing spot is apt to move forward
by irregular jerks. When a spot divides, the parts are
apt to separate rapidly.

DISTRIBUTION OF SUN-SPOTS
Sun-spots very seldom occur at higher latitudes
than 40°. Within the sun-spot belt 80° wide, as thus,
denned, the distribution of spots is irregular. They
occur mainly in two zones on either side of the equa-
tor, between latitudes 10° and 30°. As regards the
northern and southern hemispheres the number oc-
curring in a very long period of years is practically
equal, but there is often a great inequality for several
years in succession. A remarkable instance of this
irregularity occurred between 1672 and 1704, when
no spots were recorded in the northern hemisphere,
and the appearance of a few there in 1705 was noted
by the French Academy as a very extraordinary
event. Newcomb draws attention in the four cycles
1856-1898 to a marked and growing preponderance
of spots in the southern hemisphere. A peculiarity
of sun-spot distribution likely to prove of great theo-
retical significance was discovered by Spoerer, and is
confirmed by Greenwich observations. There seems
to be a close connection between the latitudes of
great prevalence and the periodicity of sun-spots.
Young states the matter as follows :

" Speaking broadly, the disturbance which pro-
duces the spots of a given sun-spot period first mani-
fests itself in two. belts about 30° north and south of

193

THE SUN

the sun's equator. These belts then draw in toward
the equator, and the sun-spot maximum occurs when
their latitude is about 16°; while the disturbance
gradually and finally dies out at a latitude of 8° or 10°,
some twelve or fourteen years after its first outbreak.
Two or three years before this disappearance, how-
ever, two new zones of disturbance show themselves.
Thus, at the sun-spot minimum there are four well-
marked spot-belts; two near the equator, due to the
expiring disturbance, and two in high latitudes, due

55 18

60 18

65 18

70 18

75 18

30°
26°

100

,/'

22°

S,^

14°

••v

•^

's^

10°

-''

"""

-»»

FIG. 55. — SPOERER'S CURVES OF SUN-SPOT LATITUDE.

to the newly beginning outbreak ; and it appears that
the true sun-spot cycle is from twelve to fourteen
years long, each beginning in high latitudes before
the preceding one has expired near the equator.

" Fig. 55 illustrates this, embodying Spoerer's results
from 1855 to 1880. The dotted curves show Wolf's
sun-spot curve for that period, the vertical column at
the right of the figure, marked W at the top, giving
Wolf's ' relative numbers. ' The two continuous curves,
on the other hand, give the solar latitudes of the two
series of spots that invaded the sun's surface in those
years. The scale of latitudes is on the left hand. The

194

SUN-SPOTS, FACUL.E, AND GRANULATION

first series began in 1856 and ended in 1868; the sec-
ond broke out in 1866 and lasted until 1880. During
these years it happened that there was very little dif-
ference between the northern and southern hemi-
spheres of the sun. "

In a summary of the results of solar observations
made at Greenwich from 1874 to 1902 the Astrono-
mer Royal, Christie, gives data showing the prevail-
ing latitudes at which they occurred in different parts
of the sun-spot cycles. The maximum latitude at
which spots occurred was 42°, but they could only be
regarded as sporadic phenomena above latitude 33°.
Preceding a time of sun-spot minimum, the prevailing
spottedness occurred at low latitudes, and when the
spots reappeared after minimum it was generally at
high latitudes. The equatorial belt from 5° to -5°
was never a center of spottedness. These facts are
indicated by the following table, abridged from the
data given by the Astronomer Royal.

Year

18-*

Centers of

21°

16°

11°

9°

21°

15°

12°

8°

spottedness

19°

18°

11°

10°

20°

15°

12°

7°

Wolf's numbers

The protuberances, on the contrary, as shown by
Ricco and the Lockyers, and confirmed by Mascari,
have their zones of maximum frequency transferred
from low toward higher latitudes as the sun-spot
cycles progress.

195

THE SUN

SUN-SPOT FORMATION AND LIFE-HISTORY

As regards the formation and life history of sun-
spots, Young has described the phenomena in these
words :

" There is no regular process for the formation of a
spot. Sometimes it is gradual, requiring days or even
weeks for its full development, and sometimes a single
day suffices. Generally, for some time before the ap-
pearance of the spot, there is an evident disturbance
of the solar surface, manifested especially by the pres-
ence of numerous and brilliant faculae,1 among which,
" pores" or minute black dots are scattered. These
enlarge, and between them appear grayish patches,
apparently caused by a dark mass lying veiled below
a thin layer of luminous filaments. The veil grows
gradually thinner, and vanishes, giving us at last the
completed spot with its perfect penumbra. The
"pores," some of them, coalesce with the principal
spot, some disappear, and others constitute the at-
tendant train. When the spot is once completely
formed, it assumes usually an approximately circular
form, and remains without striking change until its
dissolution. As its end approaches, the surrounding
photosphere seems to crowd in upon and cover and
overwhelm the penumbra. Bridges of light, often
many times brighter than the average of the solar
surface, push across the umbra, the arrangement of

1 This is Secchi's view. Lockyer maintains that the spots appear
before the faculae.

196

SUN-SPOTS, FACUL.E, AND GRANULATION

the penumbra filaments becomes confused, and, as
Secchi expresses it, the luminous matter of the photo-
sphere seems to tumble pell-mell into the chasm,
which disappears and leaves a disturbed surface
marked with faculae, which in their turn subside after
a time. As intimated before, however, the disturb-
ance is not unfrequently renewed at the same point
after a few days, and a fresh spot appears just where
the old one was overwhelmed.

" The spots usually appear not singly, but in groups
— at least, isolated spots of any size are less common
than groups. Very often a large spot is followed upon
the eastern side by a train of smaller ones;, many of
which, in such a case, are apt to be very imperfect in
structure, sometimes showing no umbra at all, often
having a penumbra only upon one side, and usually
irregular in form. It is noticeable, also, that in such
cases, when any considerable change of form or struc-
ture shows itself in the principal spot of a group, it
seems to rush forward (westward) upon the solar
surface, leaving its attendants trailing behind. When
a large spot divides into two or more, as often hap-
pens, the parts usually seem to repel each other and
fly asunder with great velocity — great, that is, if
reckoned in miles per hour, though, of course, to a
telescopic observer the motion is very slow, since one
can only barely see upon the sun's surface a change of
place amounting to two hundred miles, even with a
very high magnifying power. Velocities of three or
four hundred miles an hour are usual, and velocities

197

THE SUN

of one thousand miles, and even more, are by no
means exceptional.

"The average life of a sun-spot may be taken as
two or three months; the longest yet on record is
that of a spot observed in 1840 and 1841, which lasted
eighteen months. There are cases, however, where the
disappearance of a spot is very soon followed by the
appearance of another at the same point, and some-
times this alternate disappearance and reappearance
is several times repeated. While some spots are thus
long-lived, others, however, endure only for a day or
two, and sometimes only for a few hours. "

Carrington, Secchi, Perry, Maunder, and Sid-
greaves have all noted the tendency of spots to recur
in the same positions, but not in a sense indicative of
permanent special eruptive places, as in the case of
terrestrial volcanoes. Father Sidgreaves says: "They
are indications of a more enduring state of disturb-
ance than is measured by the lifetime of a single spot,
for it is not improbable that a recurrence springs from
the same source as its predecessor. And, if this be
true, the spots must be more subject to drift than
their underlying origins, for nearly always the recur-
ring spot is found to the rear of the former position. "

According to the spectroheliographic investiga-
tions of Fox :l i i Spot birth is always accompanied by,
and generally antedated by, an eruption" (i. e., erup-
tive prominence). " In the early hours of the life of a
spot the eruption may partially or entirely cover the

lAstrophysical Journal, vol. xxviii, p. 255, 1908.
198

SUN-SPOTS, FACUL^E, AND GRANULATION

spot, and often may precede it, in the direction of
solar rotation. An eruption is seldom seen preceding
a mature single spot, but if present will be following it
at the edge of the penumbra, perhaps encroaching
somewhat. If the spot is actively growing, eruptions
are almost certain to be found on the following edge.
Eruptions accompany spots in rapid decline, being
often seen at the ends of the bridges. I think the
evidence of the Rumford spectroheliograms fairly
conclusive in showing that the spot has its genesis in
the eruption. The phenomenon of spot development
following the appearance of an eruption is so general
that it is possible, on the appearance of an isolated
eruption, to predict with certainty the advent of a
spot. When the spot is well developed it stimulates
new eruptions. " The " eruptions" mentioned by Fox
are, of course, seen with the spectroheliograph any-
where on the sun's disk, but when close to the limb
they are recognized by him to be really "the bases of
the eruptive prominences. "

THE SUN-SPOT LEVEL

The level of sun-spots is a question which has been
discussed for over a century, and often with consider-
able vehemence. In 1769, Dr. A. Wilson of Glas-
gow advocated the view that sun-spots are depres-
sions of the sun's surface. He observed that when a
spot first appears on the eastern edge of the sun the
penumbra is well marked on the side nearest the edge
of the sun, but nearly invisible on the side next the
15 199

THE SUN

sun's center, while the umbra scarcely shows at all,
being as if hidden behind a bank. As the spot ad-
vances towards the center, according to A. Wilson,
the advancing and following sides of the penumbra
become more equal, and the umbra covers an increas-
ing fraction of the total width of the spot. Having
passed the center, the spot naturally exhibits the op-
posite succession of phenomena. This progress in ap-
pearance would be conclusive evidence that spots are
depressions if it were universally admitted to be real.
Many spots are so unsymmetrical, even at the center
of the sun, as to be unfavorable objects on which to
test A. Wilson's view. Many spots alter their shape in
crossing the sun's disk quite apart from any change
due to the sun's spherical form. In the last twenty
years several very assiduous observers have published
conclusions based on very numerous observations;
and, even when discussing the spots occurring in the
same course of years, about as many disagree with A.
Wilson's view as support him. It seems most prob-
able, therefore, that the level of the sun-spot phe-
nomena seen by ordinary observation differs very
little, if at all, from that of the surrounding bright
surface of the sun.

LANGLEY'S TYPICAL SUN-SPOT

Owing to the effect of the sun's rays in heating the
surface of the earth, and thereby causing the ascent of
warm currents of air which spoil the " seeing," the
observer is at a disadvantage in studying the minute

200

SUN-SPOTS, FACULJ:, AND GRANULATION

features of the sun as compared with the moon or
other night objects. The " seeing" on the sun is gen-
erally better in the hours soon after sunrise and before
sunset, when the heating of the sun's rays is dimin-
ished both by passing through a thick stratum of air
and by striking the earth's surface obliquely. Some-
times the presence of thick haze or light uniform
cloudiness appears to favor good definition, but often
these conditions are connected with atmospheric dis-
turbances so nearly in line with the sun as to spoil the
" seeing. " Good solar " seeing" is seldom found when
a clear blue sky, a brisk breeze, and high altitude of
the sun occur simultaneously. With the hindrance
thus occasioned by the irregularities of density in the
earth's atmosphere to contend with, solar observers,
as a rule, find only comparatively rare instants when
really satisfactory views of the sun's surface may be
obtained. By combining with extraordinary skill
the impressions received in the instants of best
"seeing," which were the reward of several years of
assiduous observing, the late Dr. S. P. Langley
produced, in 1873, his famous sketch of the "typical
sun-spot," a copy of which is reproduced as the
frontispiece. This is generally conceded to represent
better than any photographs, and even better than
anyone is likely to see for himself in the telescope,
the appearance of a sun-spot and its surroundings as
seen under the best purely telescopic observation.

201

THE SUN

FACUL^E

Next to sun-spots, the most prominent solar fea-
tures, and closely associated with the life history of
spots, are the faculae, or bright patches which are
most abundantly seen near the borders of the sun's
disk. Their appearance has been likened by Young
to the flecks of foam which dot the water beneath a
waterfall. They are very prevalent in the neigh-
borhood of sun-spots, but, unlike them, they are
found all over the surface of the sun, though spar-
ingly near the poles. It is difficult to see them near
the center of the sun's disk. As stated in Chapter
III, the brilliancy of the solar surface is not uniform
all over the disk, but falls off very greatly near the
edges. Speaking roughly, the faculae, on the other
hand, may be regarded as equally bright wherever
seen on the sun's disk, and hence come out more
distinct1-/ near the edges, where the background is
less brilliant. The prevalence of faculae has maxima
and minima synchronous with the sun-spot period.

GRANULATION

Besides the sun-spots and the faculse, there is seen
under good observing conditions a general granu-
lated appearance all over the sun's surface. Many
years ago much controversy was waged over the
exact forms of the granules, some observers compar-
ing them to rice grains, others to willow leaves, and
others to bits of straw. These patches of differing

202

PLATE XVI

6 h 47 m.

7 h 37m.

PHOTOGRAPHS OF A PORTION OF THE SUN. (Janssen.)
Meudon, June 1, 1878. Interval, 50 minutes.

SUN-SPOTS, FACUL^E, AND GRANULATION

brilliance are really immense areas of 10,000 to 50,000
square miles, and are probably not of a regular pat-
tern at all, so that little insight into solar conditions
is had by the discussion of their mere forms. In
Langley's sun-spot drawing they are depicted in great
numbers, and with various shapes, quite as they are
apt to occur. Plate XVI is a reproduction of two of
Janssen's celebrated photographs of them.

SUN-SPOT SPECTRA

The spectrum of a sun-spot differs from that of the
photosphere in several significant ways. (1) As meas-
ured by the bolometer or other photometric methods,
its energy is far weaker in the violet. This is shown in
the accompanying comparisons between the intensi-
ties of the spectra of sun-spots and of the photosphere
near the center of the sun's disk. The data for the
ultra-violet spectrum are from the work of Schwartz-
child and Villiger, and the remainder from the work
of the Smithsonian observers.

Wave Lengths (\ = )

(V.320

Q/i.448

O/x.586

O/i.799

l/t.218

2/i.llS

Ratio of umbra

0.12

0.377

0.424

0.53.5

0.610

0.761

brightness photosphere

As different spots differ in darkness of their centers,
too much reliance should not be placed on the transi-
tion of relative brightness from \ = 0.320/1. to X =
0.448/4, as given above. The remainder of the data,
however, all applies to the same spot observed by the

203

THE SUN

same observers, and should, therefore, be comparable.
There are three ways of explaining the progressive
relative weakness of the shorter wave-length rays in
sun-spots. The sun-spot temperatures may be much
below those of the photosphere, there may be a
greater amount of absorption or scattering of the
light above the spots, or, finally, the phenomenon
may be due to the action of both these causes. It has
lately been made practically certain that the first-
mentioned cause, at least, is operative. This is
proved by the work on sun-spot spectra noted below.
Several observers have found that the contrast of
brightness between the sun-spots and the photo-
sphere decreases towards the sun's limb. Langley,
and also Frost, found indications that at the very
limb the total radiation of the sun-spot umbra is ac-
tually stronger than that of the photosphere. W. E.
Wilson observed that the ratio of the brightness of
the spot umbra to that of the photosphere at the sun's
center did not change from the center to ninety-five
per cent out on the solar radius, whereas the ratio of
brightness of the umbra to the surroundings increased
from TOZT to rW- He could not confirm Frost's and
Langley's result. Schwartzchild and Villiger, observ-
ing at wave-length 0.32//, in the ultra-violet, found the
ratio of brightness of sun-spots to the surrounding
photosphere at the center ten to fourteen per cent,
but close to the limb it was thirty to fifty per cent.
It has already been stated that the photosphere at
the limb of the sun is less bright than it is at the cen-

204

SUN-SPOTS, FACUL.E, AND GRANULATION

ter, and the exact amount of change has been given
for various wave lengths in Chapter III. Accordingly
it is easy to see that if, as observed by W. E. Wilson,
the sun-spot umbra remains nearly unchanged in its
intrinsic brightness wherever seen upon the sun, the
results just mentioned would tend to follow. It
seems hard to believe, however, that the radiation of
the spot-umbra at the limb could actually exceed that
of the surrounding photosphere, as observed by Frost
and Langley, and further experiments along this line
should be made.

(2) In sun-spot spectra many Fraunhofer lines are
strengthened and many weakened as compared with
the same lines in the photospheric spectrum.1 From
Adams' summary of the subject2 1 take the following
data. Calcium has sixty lines in Rowland's table be-
tween \ = 0.40/4 and A, = 0.70/u, and with one possi-
ble exception all are strengthened in sun-spots. The
strengthening increases absolutely, and also rela-
tively to the intensities of the lines affected, with in-
creasing wave length. With iron there are 1 , 108 lines
in the same interval of Rowland's table, of which 784
are affected in spots. Of these, 558 are due to iron

1 A spectrum absorption line is said to be strengthened when, by
reason of its becoming broader without becoming less dark, or by
reason of its becoming darker, or from both changes, it presents a
greater contrast to the adjoining spectrum. Weakening a spectrum
line implies an opposite change. In either case the term is relative,
and may really mean the alteration of the adjoining spectrum, with-
out change in the line itself, in such a manner that the contrast of the
line is altered.

2 Contributions of the Mount Wilson Solar Observatory, No. 40.

205

THE SUN

alone, the others being blends of iron lines with very
close lines of other elements. Of the 558 purely iron
lines affected, 300 are strengthened and 258 weakened
in sun-spots. Hydrogen has four lines in the region
under discussion, and all are weakened. The case is so
striking that it is worth giving in full:

TABLE XII. — Hydrogen spectrum in sun-spots

Line

Wave Length

Intensity

Photospheric

Sun-spot

4101.848
4340.471
4861.350
6562.835

40N
20N
30
40

1
4
10
25

i«.:

H£

The following table from Adams' publication shows
the behavior of the spot lines of thirteen different ele-
ments :

TABLE XIII. — Spectrum lines affected in

Total

Number of
Lines
Strengthened

Number of
Lines
Weakened

Percentage of Total
Number

Element

i oT/ai
Number
Lines

Com-

One

Ele-
ment

pound
Lines
and

One
Ele-
ment

pound
Lines
and

Strength-
ened

Weak-
ened

Affec-
ted

Blends

Calcium

Chromium. . .

386

200

'36

"\7

Cobalt

118

Hydrogen. . .

100

Iron. .

1108

300

i27

258

S9 '

Magnesium . .

Manganese . .
Nickel

167
251

'si

106

'{)
26

59
29

14
53

Scandium . . .

Silicon

100

Sodium

100

Titanium ....

432

247

'73

'46

"\7

Vanadium. . .

176

114

206

SUN-SPOTS, FACUL.E, AND GRANULATION

COOLNESS OF SUN-SPOTS

If the layer which produces the Fraunhofer lines
over the spots were of the same temperature that it is
over the photosphere, the lines in spots would tend to
appear weakened; because, while the emission in
the lines would in that case remain really unchanged,
the spectrum background against which they are seen
would be weakened, and approach the brightness of
the lines, as has been seen under Caption 1 . Since the
reduction of the background in sun-spot spectra is
greatest for short wave lengths, the violet lines would
be most weakened in the case we are considering.
This is, indeed, the case for hydrogen, and may be
explained in that case perhaps as a consequence of
high level, but in fact the majority of sun-spots lines
are strengthened, and this in itself may be regarded
as evidence that the " re versing layer" for most ele-
ments is cooler over spots than over the photosphere.
Besides this general consideration, there are several
others, now to be mentioned, which point to the same
conclusion.

Lines which are relatively stronger in the electric
spark than in the arc, when produced as bright lines
in the laboratory, are called " enhanced lines." Of
144 enhanced lines observed in spots, says Adams,
11 130 are distinctly weakened, none are strengthened,
while sixteen show no marked change." This almost
universal weakening of enhanced lines in sun-spots
is shown as follows, to be evidence of a low tempera-

207

THE SUN

ture in the sun-spot reversing layer. By Kirchhoff's
law (see Chapter II) emission and absorption are pro-
portional. Hence, if it requires the conditions of the
spark to produce certain emission lines strongly, it
will also require the conditions of the spark to cause
the operative gases to absorb strongly in these lines.
But spark versus arc conditions are to be regarded as
of high versus lower temperatures, a view fully con-
firmed by the experiments of Hale, Adams, and Gale
with strong and weak arcs, and those of King with
the electric furnace at high and low temperatures.
Accordingly, the weakening of the enhanced lines in
the sun-spot spectrum, in opposition to the prevailing
strengthening of lines in spots, is explained by assum-
ing that the spot vapors are too cool to produce strong
absorption of enhanced lines.

A third line of evidence showing that the reversing
layer is cooler over sun-spots is furnished by a de-
tailed comparison of the spectra of sun-spots and
photosphere on the one hand, and of low and high
temperatures in the arc or electric furnace on the
other. This comparison was begun by Hale, Adams,
and Gale, and continued by King. Adams gives in a
long table the results of such a comparison for the
lines of iron. From this table several of the most
well-marked cases, typical of strengthening, weaken-
ing, and neutrality, are given in the following table.

In general, within the error of measurement, lines
strengthened in the cool arc are strengthened in sun-
spots, those weakened in the cool arc are weakened in

208

SUN-SPOTS, FACUL.E, AND GRANULATION

TABLE XIV. — Sun-spot, hot arc and cool arc spectra

i Intensity

Intensity

Wave Length
(Rowland's)

Spot
ratio

Arc ratio

Dis-
crepancy

Sun

Spot

Hot

arc

Cool
arc

4118.708

1.2

1.3

—0.1

4291.630

0.7

0.5

+0.2

4325.939

1.1

1.2

—0.1

4461.818

0.6

0.5

+0.1

4531.327

0.7

+0.0

4939.868

' 18

0.6

0.0

5083.518

0.7

0.5

+0.2

5202.516

1.0

0.9

+0.1

5333.089

0.6

0.4

+0.2

5405.989

0.6

0.5

0.1

6024.281

1.0

0.0

sun-spots, and those unchanged in one are unchanged
in the other, and all by similar proportions. It fol-
lows from this, by a similar line of argument to that
just given for enhanced lines, that the reversing layer
is relatively cooler over sun-spots than over the pho-
tosphere*

A fourth phenomenon strongly indicating the same
conclusion is the highly conspicuous presence in sun-
spot spectra of flutings, or rythmic banded appear-
ances, immensely numerous, and characteristic re-
spectively of the spectra of titanium oxide, magnes-
ium hydride, and calcium hydride. The identifica-
tions of these flutings were discovered respectively
by Hale, Adams, and Gale, by Fowler and by Olm-
sted. These and other molecular compounds give, as
Evershed has stated, very slight and not always per-

209

THE SUN

ceptible evidence of their presence in the photo-
spheric spectrum. It is well known that high tem-
peratures tend to produce complete dissociation of
molecular compounds. The copious appearance of
the lines of compounds in the spectra of sun-spots
would be very strong evidence of the relatively low
temperature in the reversing layer above spots, even
if unsupported by the other evidences given above,
and by many other minor phenomena of which space
forbids the mention.

According to Father Cortie,1 steam also occurs in
sun-spots, for he finds water-vapor lines among those
widened in sun-spot spectra. He cites experiments,
too, which indicate that the spectrum of magnesium
hydride could not show in sun-spots if water vapor
was not also present. Evershed, however, concludes
from observations at the high and dry station of
Kodaikanal that: "On the whole, it must be ad-
mitted that the evidence for the strengthening of
telluric lines, of whatever origin, in spot spectra is
practically negligible."

An excellent photographic map of the sun-spot
spectrum, contrasted with that of the photosphere,
has been prepared at the Mount Wilson Solar Ob-
servatory and distributed to solar observers. Plate
XVII, reproduced here by the permission of the
Director, shows a section of this map including the
b group. Although no engraving can do full
justice to the original, the reader will be able to

1Astrophysical Journal, vol. xxviii, p. 379, 1908.
210

SUN-SPOTS, FACUL.E, AND GRANULATION

note for himself some of the features mentioned

above.

SUN-SPOTS AND MAGNETISM

In the year 1908, Hale discovered the existence of a
magnetic field in spots, which betrays its presence by
the widening, doubling, or tripling of a great number
of spectral lines. As stated in Chapter II, Zeeman
discovered, about 1896, that most lines of the spec-
trum are separated into two components when viewed
along the lines of force of a powerful magnet, and the
two components are circularly polarized in opposite
directions. With less powerful fields, the lines are
not clearly doubled, only widened, but their right-
and left-hand edges exhibit in this case traces of op-
posite circular polarization. Hale applied this test
of polarization to the most widened lines of sun-spots
by introducing a Fresnel rhornb to convert the sup-
posed circular to plane polarization, and found the
right-hand or left-hand edge of the lines could be cut
off at will, according to the position of the Nicol
prism used for analyzing the character of polarization
of the light. Some lines are triple in spots, but these
seeming discrepancies proved to be the best of evi-
dence of the effect of a magnetic field. For when the
same lines were examined in the laboratory they
proved exceptional, and to become triple instead of
double when viewed along the magnetic lines of force.
Hale's .brilliant discovery has cleared up one of the
most puzzling questions relating to the sun-spot
spectrum.

THE SUN

By polarization studies, Hale found that sun-spot
fields are not always of the same polarity. Very
often a pair of sun-spots quite near together are found
to be of opposite polarity. In general, the polarity
of spots in the sun's southern hemisphere is opposite
to that in the northern, but there are very numerous
exceptions to this rule, as, of course, in the case of
double spots, as just mentioned. Spots near the sun's
limb, since they present their magnetic lines of force
nearly at right angles to our line of sight, tend to show
triple lines where doublets would be seen near the
center of the disk.

The cause of the magnetic field in sun-spots is a
most interesting problem. Rowland showed many
years ago that static electric charges, in rotation,
produce electro-magnetic effects similar to those pro-
duced by electric currents in coils of wire. This seems
to point the way to a solution, for, as stated in the
account of the sgectroheliographic results in Chapter
III, the sun when viewed through the hydrogen line
Ha (C) shows curved formations (see Plate XI),
which seem to indicate spiral motion in sun-spot
neighborhoods. In such Ha photographs of double
spots, which give opposing magnetic polarity, the
curves which surround the spots seem to present the
appearance not unlike those seen among iron filings
on a sheet of paper acted upon by a pair of opposite
magnetic poles. It seems, then, not improbable
that whirling motions or vortices exist in sun-spots,
and that these carry along electrically charged par-

212

SUN-SPOTS, FACUL.E, AND GRANULATION

tides which produce the observed magnetic fields.
The impression was at first that these charges were
the so-called ions, or bodies smaller than atoms, re-
cently made known by J. J. Thomson arid others; but
great difficulty was found in accounting for their
isolation in sun-spots in sufficient numbers. It was
suggested to Mr. Hale by the writer that the mole-
cules of the compounds shown in the sun-spot spec-
trum, or perhaps even the relatively cooled elemen-
tary gases in spots, might very probably be regarded
as sufficiently different from the surroundings to pro-
duce frictional electricity, when whirled about in
the spots, just as steam becomes electrified in Arm-
strong's machine when, carrying water-droplets, it
issues from an orifice. Further discussion of the
matter will be found in Chapter VI.

«
RADIAL MOTION IN SPOT PENUMBRAS

Evershed has lately observed shifting of spectral
lines in the penumbras of spots situated at consider-
able distance from the center of the sun's limb. This
seems to indicate motion nearly radial to the center
of the spots, as if material was coming to the sun's
surface in the sun-spot centers, and then spreading
out in all directions, like smoke from a volcano.
Nevertheless, no spectroscopic evidence of motion in
spots radial to the center of the sun has ever been ob-
tained.1 Adams has lately sought to find evidences

1 As this is being published St. John has observed high level gases
moving downwards in spots.

213

THE SUN

of increased or decreased pressure in the reversing
layer over sun-spots, from shifting of lines known to
be subject to large shifts when their sources are under
pressure, but he was unable to discover evidences of
altered pressure. The significance of these facts will
be discussed in Chapter VI.

CHAPTER VI.

WHAT IS THE SUN?

Young's Views. — Halm's Views. — Schmidt's Hypothesis. — Julius'
Views. — The Author's Views.

BELIEVING that the views of the late Professor
Young probably are still shared by a majority of as-
tronomers, even after the lapse of fifteen years since
the appearance of the last revision of his work, "The
Sun, " we shall begin this chapter by quoting a part of
the summary which he gives in his Chapter IX. We
shall then take up the solar theories of Halm, Schmidt,
and Julius. In the remainder of the chapter we shall
consider still another view of the matter, which the
present writer inclines to adopt.

YOUNG'S VIEWS

Quoting from Young's "The Sun:"

"Fig. 56 is intended to present to the eye, more
clearly than any mere description, the constitution of
the sun, and the relation of the different concentric
shells or envelopes as conceived by the writer.

"The picture is an ideal section through the center.
The black disk represents the inner nucleus, which is
not accessible to observation, its nature and constitu-
tion being a mere matter of inference. The white
16 215

THE SUN

ring surrounding it is the photosphere, or shell of in-
candescent cloud which forms the visible surface.
The depth, or thickness, of this shell is quite un-
known; it may
be many times
thicker than rep-
resented, or
possibly some-
what thinner.
Nor is it certain
whether it is
separated from
the inner core
by a definite sur-
face, or whether,
on the other
hand, there is no
distinct bound-
ary between
them.

" The outer
surface of the
photosphere,
however, is cer-
tainly pretty
sharply defined, though very irregular, rising at points
into faculae, and depressed at others in spots, as
shown in the figure.

" Immediately above this lies the so-called 'revers-
ing stratum, ' in which the Fraunhof er lines originate.

216

FIG. 56. — SOLAR DIAGRAM. (Young.)

WHAT IS THE SUN?

It is to be noted, however, that the gases which com-
pose this stratum do not merely overlie the photo-
sphere, but they also fill the interspaces between the
photospheric clouds, forming the atmosphere in
which they float, and an attempt has been made to
indicate this fact in the diagram.

"Above the ' re versing stratum' lies the scarlet
chromosphere, with prominences of various forms and
dimensions rising high above the solar surface; and
over, and embracing all, is the coronal atmosphere
and the mysterious radiance of clouds, rifts, and
streamers, fading gradually into the outer darkness.

"At the center of the sun the earth is represented in
its true relative dimensions — or of the three inches
which is taken as the scale of the sun's diameter. This
scale reduces our globe to a little dot only fa of an
inch across. Around it, at its proper distance, is
drawn the orbit of the moon, still far within the pho-
tosphere, the moon herself being fairly represented
by any one of the minute points which make up
the dotted line that indicates her path.

"The central nucleus is made black in the picture,
simply for convenience, and not with any purpose to
indicate that the matter which composes it is cooler or
even less brilliantly luminous than the photosphere.
It is quite probable, indeed, that this central core
(which contains certainly more than nine-tenths of
the whole mass of the sun) is purely gaseous, and it is
of course true that, at a given temperature and pressure,
a gaseous mass has a lower radiating power, and is

217

THE SUN

less luminous, than a mass of clouds, such as those
which constitute the photosphere. But, on the other
hand, both compression and increase of temperature
rapidly raise the radiating power of a gas; and it is
highly probable that, at no very considerable depth,
the growing pressure and heat may more than equal-
ize matters, and render the central nucleus as in-
tensely bright as the surface of the sun itself.

" At the upper surface of the photosphere, however,
and all through it, indeed, the uncondensed gases are
dark as compared with the droplets and crystals
which make up the photospheric clouds. Here the
pressure and temperature are lowered, so that the
vapors give out no longer a continuous but a bright-
line spectrum, whenever we get a chance to see them,
against a non-luminous background; and, when the
in tenser light from the liquid and solid particles of the
photosphere shines through these vapors, they rob it
or the corresponding rays, and produce for us the
familiar dark-lined spectrum of ordinary sunlight.

" Although it 'may not be possible, in the present
state of science, to demonstrate that the principal por-
tion of the solar mass is gaseous, this much can at least
be said — that a globe of incandescent gas, under condi-
tions such as have been intimated, would necessarily
present just such phenomena as the sun exhibits.

"On the outer surface, exposed to the cold of space,
the rapid radiation would certainly produce the con-
densation and precipitation into luminous clouds of
such vapors as had a boiling-point higher than that of

218

WHAT IS THE SUN?

the cooling surface. These clouds would float in an
atmosphere saturated with the vapors from which
they were formed, and also containing such other va-
pors as were not condensed, and thus the peculiarities
of the solar spectrum would result. On the other
hand, the permanent gases, like hydrogen — those not
subject to condensation into the liquid form under
the solar conditions — would rise to higher elevations
than the others, and form above the photosphere just
such a chromosphere as we observe. Whether, from
the mere assumption of such a constitution for the
sun, one could work out, a priori, the phenomena of
sun-spots and prominences, is indeed doubtful; but
thus far nothing in any of them has been observed
which appears to be inconsistent with this view of the
subject — nothing, we say, unless it should turn out, as
was once maintained, that the solar surface possesses,
so to speak, ' geographical ' characteristics, evinced by
the disposition to break out into sun-spots at certain
fixed points — as if at those points there were volca-
noes or something of the sort. Of course, the fact that
the spots are distributed mainly in two belts parallel
to the solar equator, involves no difficulty, for it is
easy to conceive how, in more than one way, the sun's
rotation might lead to such a result : but peculiarities
permanently attaching to individual points on the
solar surface necessarily imply rigid connections, such
as are inconsistent with the theory of a gaseous or
even of a fluid nucleus. But while, as has been
already pointed out, there is a marked tendency in

219

THE SUN

spots to recur at or near the same points during sev-
eral solar revolutions, there is no evidence which es-
tablishes the existence of fixed spot-centers; and the
idea is to he regarded merely as a relic of the old
Herschellian theory of a solid sun. Still it is difficult
to test the notion conclusively even by means of such
extended observations 'as those of Carrington or
Spoerer, or the auroral periods of Veeder, since the
time of rotation of the solid nucleus, if it exists at all,
is unknown, and this makes the discussion difficult
and unsatisfactory.

"With reference to the constitution of the photo-
sphere there is a general agreement among astrono-
mers. A few, perhaps, still hold, as has been men-
tioned, to the idea that the visible surface is a liquid
sheet, while some believe that it is purely gaseous;
but the whole appearance of things, the details of the
granulation, the phenomena of spots and faculae, the
mobility and variability of the floccules, all better
accord with the theory adopted in these pages, which
is a necessary consequence of the hypothesis that the
sun is principally gaseous. It seems almost impos-
sible to doubt that the photosphere is a shell of clouds.
As to the precise constitution of this shell, however,
the form and magnitude of the component cloudlets,
the chemical elements involved, and the temperature
and pressure, there is room for a good deal of uncer-
tainty and difference of opinion. The more common
view, apparently — the one, certainly, which the
writer has hitherto held — is, that the clouds are

220

WHAT IS THE SUN?

formed mainly by the condensation of the substances
which are most conspicuous in the solar spectrum,
such as iron and the other metals. As to the form of
the clouds, also, it has usually been assumed that,
as a consequence of the ascending currents by which
they are formed, they are columnar, their height
being much greater than their other dimensions.

" Professor Hastings has proposed a somewhat dif-
ferent theory, which avoids some of the difficulties of
the received doctrine, though not without encounter-
ing others which seem just as formidable.

"One main peculiarity is the assumption that the
photospheric ' clouds ' are formed by the precipitation
of either carbon, silicon, or boron (the three members
of the carbon group), to the exclusion of other sub-
stances which are less refractory (have lower boiling-
points), and therefore escape precipitation.

" His idea that the stratum which produces the gen-
eral absorption at the limb of the sun is a veil of
'smoke' — i. e., of the same minute particles which
constitute the photosphere, but cooled to relative
darkness — has been already alluded to in a preceding
chapter. So far as we know, it is novel and valuable,
clearing up a good many embarrassing difficulties. It
is so obvious, on reflection, that something of the sort
must accompany the photosphere, that it is surprising
that the idea had not been thought of before. Of
course, the particles formed by condensation must,
many of them at least, be carried by the ascending
currents high above the point of their formation, and

221

THE SUN

cooled so much as to become relatively dark in com-
parison with the more vivid incandescence of the re-
gions below, just as the ascending particles of carbon,
unconsumed and cooled, constitute the smoke of a fire.

"The idea that carbon may be the main constitu-
ent of the photosphere is by no means new : it was first
seriously advanced, we believe, by Johnstone Stoney,
of Dublin, as early as 1867, mainly on physico-chem-
ical grounds, and is enthusiastically advocated by Sir
Robert Ball in his recent ' Story of the Sun. '

"As regards the 'reversing stratum' very little need
be added. Mr. Lockyer indeed denies its existence —
that is, in the sense that there is a thin stratum, close
above the surface of the photosphere, in which most
of the dark lines of the solar spectrum originate. He
maintains, on the contrary, in accordance with his
'dissociation theory,' that certain of the lines, due to
substances the most nearly elementary, and having
their molecules in the highest stage of dissociation,
originate only deep down in the solar atmosphere
where the heat is most intense; others, due to vapors
with molecules somewhat less simple, have their birth
a little higher; and others yet, due to molecules the
most complex, are produced only in the most elevated
regions of the solar atmosphere; each elevation thus
being responsible for its own special family of spec-
trum lines.

" if, however, we reject this theory as 'not proven, '
we get results not very different.

"The vapors of the photosphere and chromosphere
222

WHAT IS THE SUN?

are not to be thought of as entirely separate and dis-
tinct. All the gases are found together in the inter-
stices between the cloud-granules of the photosphere
— the unknown substance which produces the green
line in the spectrum of the corona, the hydrogen, the
calcium, and helium which characterize the chromo-
sphere, and the metallic vapors which give the re-
versing layer its peculiar properties — these all exist
together in the lower depths, unless, indeed, it may
possibly be the case that at the greater elevations
some compound bodies are formed which can not exist
in the fiercer fires below. So far as we can distinguish
between these different portions, we may define the
photosphere as the shell within which precipitation is
taking place; the reversing layer, as that lowest re-
gion of the solar atmosphere which contains sensibly
all the gases indicated by the spectroscope ; the chro-
mosphere, as the region of hydrogen, calcium, and
helium; and the corona, as that upper domain of the
solar atmosphere which becomes observable only dur-
ing solar eclipses. But the coronal gas itself is most
conspicuous and abundant right in the photosphere
and reversing layer, and the same is true of the hydro-
gen of the prominences.

"It is well, also, to bear in mind that, if any sub-
stances decomposable by heat exist upon the sun at
all, we must expect to find them in the higher and
cooler regions of the solar atmosphere. In and near
the photosphere, or underneath it, matter must be in
in its most elemental state.

223

THE SUN

"As to the mechanism of the chromosphere and
prominences, if we may use the expression, much cer-
tainly remains to be learned. In many cases, indeed,
perhaps in most, the forms and behavior of the protu-
berances are satisfactorily enough accounted for by
supposing that the heated hydrogen and its associate
vapors is simply forced up into cooler regions by pres-
sure from below — a pressure which must result from
the downward movement of the great mass of pre-
cipitated matter which forms the photosphere. But
evidently this is not the whole story. We must have
recourse to ideas of a different order to account for the
somewhat rare, but still really numerous and well-
authenticated instances when the summits of prom-
inences have been seen to rise in a few minutes to ele-
vations of two or three hundred thousand miles, the
upward motion being almost visible to the eye at the
rate of a hundred miles a second or more.

"Very perplexing, also, is the indubitable fact that
clouds of this prominence-matter sometimes gather
and form without any apparent connection with the
chromosphere below, apparently just as clouds form
in our own atmosphere, by the condensation of vapor
before invisible. On the whole, it looks very much as
if we must regard the prominences as differing from
the surrounding medium mainly, if not wholly, in
their luminosity — as simply superheated portions of
an immense atmosphere.

"But, then, we immediately encounter the difficul-
ties so ably urged by Lane, Lockyer, and others, that

224

WHAT IS THE SUN?

the existence of hydrogen of any appreciable density,
at the elevation of even a hundred thousand miles,
implies a density and pressure at the surface of the
photosphere so high as to be entirely inconsistent with
the spectroscopic phenomena there manifested — un-
less, indeed, under solar conditions, the action of
gravity upon the gases of the solar atmosphere is
modified by some repulsive force. That such a force
is at least conceivable, is obvious from the behavior
of the tails of comets; and many features in the cor-
ona point in the same direction. Of its nature and
origin we can not, however, assert anything as yet.

"Even more difficult than the problem of the
chromosphere is that of the corona. While it is some-
thing to know that the phenomenon is mainly solar,
and that, therefore, it must rank in magnitude and
importance with the most magnificent of natural ob-
jects, we have yet to find a satisfactory explanation
of many of its most obvious features. It is certainly
very complex — matter meteoric and matter truly
solar; orbital motion, solar attraction, atmospheric
resistance, and actions thermal, electrical, and mag-
netic, are probably all combined."

HALM'S VIEWS

Since the time when Young wrote, Halm has con-
tributed the following theory, designed particularly
to explain the periodicity of sun-spots.1 Halm calls

1 Annals Royal Observatory Edinburg, vol. i, pp. 74-151,
1902.

225

THE SUN

attention to the function of the so-called solar enve-
lope, and refers to the views of Langley, Pickering,
and others, that it prevents the escape of half of the
solar energy. He then considers the effect of changes
in its powers of restraint. He accepts Helmholtz's
hypothesis that the source of the solar energy lies in
the contraction of the sun, and thinks, in contradic-
tion to See's views, that the sun is already gradually
cooling. He then refers to Hastings' paper (cited by
Young) on the nature of the solar envelope, and says :
" Indeed, it seems obvious that these particles which,
while ascending from the interior to the surface, are
precipitated so as to form the luminous clouds of the
photosphere must (quoting Hastings) 'rapidly cool
on account of their great radiating power, and form a
fog or smoke which settles slowly through the spaces
between the granules ' and that ' it is this smoke which
produces the general absorption at the limb. ' ' ' Then,
to emphasize the importance of heat conservation by
the solar envelope, Halm refers to Langley's early
view (which apparently Halm has not noticed that
Langley afterwards retracted) to the effect that the
earth's temperature would fall to —200° C. if it had
no atmosphere.1

He suggests that if gravitation should be temporar-
ily too little to supply heat energy by contraction to
balance lost energy of radiation, then the layer of

1 See "Report of the Mt, Whitney Expedition" p. 123, and "The

Temperature of the Moon," Memoirs National Academy, vol. iv,
pt.2, p. 193.

226

WHAT IS THE SUN?

maximum incandescence would cool, and a lower
layer would become the new layer of maximum in-
candescence. The absorbing layer thereby increases in
thickness, so that the new layer of maximum incan-
descence dissipates less energy than the first. Thus,
at length a layer is reached which dissipates energy of
radiation as fast as gravitation supplies energy of heat.
But, when this state occurs, the outer layers will still
go on cooling, since they receive less radiation from
within than formerly. Consequently they continue to
grow more opaque, and the amount of energy of radi-
ation dissipated to space thereby becomes less than
the amount of heat energy supplied by contraction.

Halm continues: "It thus comes to pass that,
while the function of the absorbing envelope is >that
of reducing as much as possible the waste of energy
from the photospheric layers beneath, it is, by the
very nature of the process, compelled to overdo its
work, and to finally preserve too much energy within
the star. The outbreak of eruptions and the forma-
tion of spots are the consequence of an unstable equi-
librium in the photospheric layers, and take place
whenever the supply of heat from the interior is so
supplemented by the continuous reflection of heat
from the overlying atmosphere that the photospheric
layers receive more heat than is required for the main-
tenance of their thermal equilibrium. .

"The function of eruptions, consisting as they do
in the ejection of overheated photospheric matter, is
to produce a general heating and clearing up of the

227

THE SUN

cooled absorbing layers of the solar envelope. The
action of the spots consists in drawing the cooled por-
tions of this atmosphere into the hotter regions of the
photosphere. "

Halm then goes on with mathematical work aimed
to show that the consequences of these principles lead
to a periodicity of sun-spot phenomena similar, even
in its details, to that actually observed, but this part
of the paper does not seem to be so soundly based as
to add much to the merit of his views.

There is a strong objection to Halm's mathematical
analysis, which applies, also, to the several computa-
tions made by Vogel, Pickering, and others to deter-
mine the effectiveness of the so-called "solar enve-
lope," which these astronomers regard as a layer
which restrains the emission of the sun. For they
treat only the losses suffered by the direct beam
through scattering in this envelope, without taking
account of the gains which the beam acquires from
rays scattered into it by the same envelope. Their
numerical results are hence of no application to the
sun; for the light proceeding in a single direction
from any point in the "envelope" is derived from
almost a full hemisphere. Their formulae are appli-
cable only to a case like that of the earth's atmos-
phere, where the entering rays are practically all
parallel.

SCHMIDT'S HYPOTHESIS

It was in 1891 that Schmidt published his theory of
a gaseous photosphere, and he explained the appar-

228

WHAT IS THE SUN?

cntly sharp outline of the sun in a very ingenious and
interesting way. It is well known that the sun and
other objects are seen, after they get below the real
horizon of the earth, by the refraction of the air,
which curves the rays of light. The amount of curva-
ture depends on the rate of change of optical density
of the atmosphere from its outer limits to the surface
of the earth. At sea-level the difference caused by
refraction between the apparent and real positions of
heavenly bodies is about one-half a degree of arc.
Suppose the earth were to grow larger, but with the
same atmospheric densities prevailing. There would
be a certain limiting diameter, about seven times that
of the earth, where the curvature of the rays would be
just sufficient to cause them to bend entirely around
the earth in passing from the top to the bottom of the
atmosphere, so that if there were no loss of light on
the circuit a man as tall as the atmosphere is thick
might be imagined to stand on his head at the equa-
tor, arid looking directly in front of him see his own
heels all the way around the world. If the earth were
supposed still larger then all the rays leaving its sur-
face tangentially would be incurved, and reach the
surface again at some other point, without ever suc-
ceeding in escaping to space.

Schmidt conceived of the sun as a wholly gaseous
body, above the limiting size just discussed. Accord-
ingly, looking from the earth, there would be a cer-
tain diameter at which the line of sight would curve
around in an infinitely long spiral of practically con-

229

THE SUN

slant diameter. A line of sight outside of this would
pass nearly straight through the outer layers of gas,
and emerge on the opposite side of the sun in space.
A line of sight to a smaller circumference would pass
along a diminishing spiral inside the sun till it almost
reached the lesser sphere, to which it would finally be
tangent ; and there it would go around and around in
an infinite spiral course of practically constant diam-
eter. Hence, all lines of sight inside a certain limiting
circumference would give brilliant effects because
they would have an infinitely long path of incandes-
cent gas of great density to take light from ; while all
lines of sight outside this limiting circumference, hav-
ing only a limited thickness or rarified gas to take
light from, would give by comparison only negligibly
faint effects.

According to this view the solar phenomena, sun-
spots, for example, need not be regarded as superficial,
but may lie at any point between the outer limiting
sphere and the inner sphere to which the diminishing
spiral of refraction of the line of sight at length be-
comes tangent. If this is so, a sun-spot which we see
near the limb may really be somewhere on the oppo-
site side of the sun from the earth. This hypothesis
has curious consequences if we consider the apparent
rotation of the sun as measured by observing sun-
spots. For the supposed inner sphere, on which the
spot by hypothesis really lies, must go at a different
rate of rotation from the apparent rate of the sun.
Suppose the sun's equator in the plane of the ecliptic,

230

WHAT IS THE SUN?

and that a certain equatorial spot actually lay at the
limb of the inner sphere, but appeared at the limb
of the boundary sphere. After a synodic day's ro-
tation, the light pursuing the same actual path within
the sun as before would come out, it is true, at a point
just as far advanced in angle on the boundary sphere
as the point it started from was on the inner sphere;
but coming out nearly tangent to the outer sphere, as
before, it would not be directed towards the earth at
all. The path of light directed towards the earth
through a sun-spot, advanced apparently one synodic
day's march on the boundary sphere, would pursue an
entirely differently shaped spiral within the sun, and
would cut our hypothetical sun-spot sphere on its
equator, to be sure, but not, at the same angular de-
parture from the first position as would be indicated
by the appearances. But when we reach the center
of the sun's disk the line of sight is straight. Hence,
the total period of rotation of the supposed inner sun-
spot sphere must equal that of the apparent rotation
outside ; for every time the apparent sun-spot reaches
the center of the disk the real one is directly behind it.
Accordingly, the motion of the supposed inner sun-
spot sphere must be non-uniform, which seems ab-
surd. The sun-spot must, therefore, be really, as well
as apparently, superficial. An interesting result of
Schmidt's hypothesis appears, also, if we consider
spectroscopic line-of-sight determinations of solar
rotation. For the motion in the line of sight depends
on how far down in the sun we consider the light as
17 231

THE SUN

arising, so that it would seem that all spectral lines
should be widened, when viewed near the limb, unless
the material which gives rise to them is situated close
to the apparent level of the limb.

Schmidt's views have obtained considerable ac-
ceptance, but not from observers of solar phenomena.
The late Professor Keeler said:1 " According to this
theory, the sharpness of the sun's limb and the enor-
mous change of brightness at that place are not
caused by corresponding abrupt changes in the con-
stitution, density, or light-radiating power of the
solar matter, but are the result of refraction in a non-
homogeneous medium. ... In other words, the
photosphere is an optical and not a material sur-
face. . . . Various assumptions as to the mass,
temperature, etc., are here necessary, which it is gen-
erally impossible to verify, but Dr. Knopf has
shown . . . that the conditions in the case of the
sun are well within the bounds of probability. . . .
But, however difficult it may be for present theories
to account for the tenuity of the solar atmosphere,
immediately above the photosphere, and however
readily the same fact may be accounted for by the
theory of Schmidt, it is certain that the observer
who has studied the structure of the sun's surface,
and particularly the aspect of the spots and other
markings as they approach the limb, must feel con-
vinced that these forms actually occur at practically

1 Asirophysical Journal, vol. i, p. 178, 1895.
232

WHAT IS THE SUN?

the same level, that is, that the photosphere is an
actual and not an optical surface. "

JULIUS'S VIEWS

Professor W. H. Julius of Utrecht has proposed a
group of solar theories composed of ingenious appli-
cations of the principles of anomalous dispersion. It
has been abun-
dantly shown by tD03
laboratory ex-
periments that

the dispersion of
light by the L00°
vapors of metals 0999
is subject to dis-
continuities in
the regions of 7
spectrum im- 0.9%
mediately adja-
cent to their lines
of strong emis-
sion and absorp-
tion. Fig. 57 shows the anomalous two-branched dis-
persion curve of sodium vapor in the neighborhood of
the D lines, according to researches of R. W. Wood.
For comparison, the normal dispersion of rock-salt in
the same region is also given. The enormous vari-
ations of dispersion of the light on the edges of the D
lines by sodium vapor would cause the production of
dark spectral lines under certain circumstances, not

233

FIG. 57. — NORMAL AND ANOMALOUS
DISPERSION.

THE SUN

by true absorption, but by anomalous dispersion.
Julius has applied this to the explanation of many of
the solar phenomena, and the reader interested should
consult his numerous papers and also the critical
articles of Hartmann, Anderson, and others. See
The Astrophysical Journal, Astronomische Nachrichten,
et cetera.

We may briefly state two or three of Julius's ex-
planations here, and first concerning the chromo-
sphere and prominences. These objects have bright
line spectra, and appear to protrude beyond the limb
of the sun. Eruptive prominences often appear to
shoot out as rapidly as 100 miles a second! But to
Julius they are not seen by their own brilliance out-
side the sun's limb, nor do they rise with such veloci-
ties at all. The line of sight to the apparent summit
of a prominence is really, he thinks, a greatly curved
line by virtue of the anomalous dispersion caused by
the non-homogeneous density of a mass of non-lum-
inous gas existing there; and the true source of the
principal light is in the photosphere. A slight re-
arrangement of the density alters greatly the path of
the rays, and causes the impression of displacement
of the prominence at enormous speeds. Adjacent
wave lengths of the photospheric light do not reach
the observer, because not anomalously refracted.
The wave lengths of prominence spectra, if unaffected
by other causes, would generally be slightly greater
than the wave lengths of the true absorption lines of
the gases concerned, because the density must, on

234:

WHAT IS THE SUN?

the whole, diminish outside the photosphere. But
Julius regards irregular density gradients in the oppo-
site direction as of common occurrence, so that short
wave lengths will frequently occur. The displace-
ments of wave lengths themselves will, he thinks, be
almost imperceptible. Whatever may be our opinion of
Julius's explanation of the high prominences, we must,
I think, admit a considerable probability that anomal-
ous dispersion might produce many of the phenomena
of the chromosphere. But, on the other hand, if the
chromospheric gases are self-luminous, the anomalous
dispersion effects may be almost entirely masked.

Fraunhofer lines Julius regards as " absorption
lines enveloped in dispersion bands/' the latter
caused by a honeycomb of irregular density gradients
in the photosphere, and showing themselves chiefly
as the " wings" which occur with many lines. Rever-
sals of chromospheric lines he regards as evidence of
local condensations of gas, in which density gradients
in both directions occur, thus bringing both longer
and shorter wave-length dispersion bands to the eye.

Even sun-spots he attributes to refraction, but not
anomalous refraction, at least as regards their major
phenomena. He imagines local strong condensations
or rarefactions in the photosphere, and shows how
these might produce regions of diminished radiation,
on account of the re-distribution of rays, and the
return of some to the sun.1

1 See further "Regular Consequences of Irregular Refraction in
the Sun," by W. H. Julius, Proc. Roy. Acad. of Amsterdam, Meet-

236

THE SUN

Astronomers generally admit that in the sun there
may be conditions which favor the production of
phenomena of anomalous dispersion, especially in the
chromosphere. With few exceptions, however, they
believe anomalous effects negligible, and the observed
facts to be more simply and satisfactorily accounted
for on the basis of ordinary views of selective emis-
sion and absorption, such as have been given in pre-
ceding chapters. The test between the two methods
of explanation often involves the precise measure-
ment of wave-lengths, and such criteria have not thus
far been applied with such rigor as to exclude entirely
the explanations advanced by Julius. It is not im-
possible that writers on solar phenomena ten years
hence will devote much space to the discussion of
anomalous dispersion.

THE AUTHOR'S VIEWS

We must leave the reader to supplement by his
reading of the original papers these inadequate sum-
maries of the views of various investigators, and we
will pass on to the solar theory which seems to the
writer most probable. In its most general aspect
this is similar to the views stated by Secchi in 1877
for Newcomb's " Popular Astronomy." Also an
important paper by Schuster entitled "Radiation

ing of Sept. 25, 1909; "On the Origin of the Chromospheric
Light," Mooting of same Academy, Nov. 27, 1909. "Anomalous
Refraction Phenomena Investigated with the Speotroholiograph,"
by W. H. Julius, Astropkysical Journal, Dec., 1<M)S, cl cctmi.

236

WHAT IS THE SUN?

Through a Foggy Atmosphere" 1 has some things in
common with it. Still more in touch is the paper of
Schwartzchild already mentioned.2

It will be assumed: A. The sun, excepting perhaps
in sun-spots, is wholly gaseous or vaporous. Except
in sun-spots the photosphere is too hot to contain
solids or liquids. B. The density of the gases rapidly
diminishes, and their temperature rapidly falls from
within outwards across the apparent boundary of the
sun.

The view that the sun's photosphere is too hot to
contain other than gaseous constituents has been
strongly combated by J. F. Hermann Schulz,3 who
even argues that the sun is mainly liquid. He sets
the average temperature of the photosphere at 5,400°
C. (5,673° Abs.) Although admitting that the late
H. Moissan placed the temperature of his electric
furnace at 3,500° C., and stated that all known ele-
ments volatilize at that temperature, Schulz argues
that the temperature of the electric furnace is to be
set higher, even probably as high as the sun's temper-
ature, and the volatilization is not to be regarded as
complete in the furnace. His argument is that the
enormous energy of the electric current used (see
table below) had no adequate escape by conduction
or radiation, and must have raised the temperature

1 Astrophysical Journal, vol. xxi, pp. 1-22, 1905.
2"Ueber das Gleichgewicht der Sonnenatmosphare," Gottingen
Nachr., Math-phys. Kl. 1906, pp. 1-13.

3 Astrophysical Journal, vol. xxix, pp. 33-39, 1909.

237

THE SUN

of the furnace till checked by the melting and evapo-
ration of the limestone of which it was constructed.

He continues: " Now Moissan has shown that, even
at the enormous temperature attained in his electric
furnace, we have not yet reached the point at which
all terrestrial elements are truly boiling. In this re-
spect the following table is very instructive, which
Moissan gave in Comptes Rendus of February 19,
1906 (142, 430)."

TABLE XV. — Moissan's experiments on the vaporization of metals of
the iron family

Metal

Weights
Grams

Time
Minutes

Amperes

Volts

Metal
Distilled
Grams

Nickel

150

500

110

200

500

100

200

Iron

150

825
800

10
20

500
1000
1000

110
55
110

150
400

M anganese

150

500

110

150

500

110

Chromium

150

500

110

Molybdenum

150
150

10
20

700
700

110
110

0
56

Tungsten

150

800

110

Uranium

150
150
200

5
5
9

500
700
900

110
110
110

0
15
200

" Moissan further adds the following remarks:
'Molybdenum. The 150 grams were not fused by a
current of 500 amperes and 110 volts. After applying

238

WHAT IS THE SUN?

700 amperes and 110 volts for seven minutes, the
metal was fused but nothing evaporated. After
twenty minutes 56 grams were distilled. Tungsten.
After applying 500 amperes and 110 volts for five
minutes, the metal was not yet fused. After apply-
ing 800 amperes and 110 volts for twenty minutes,
boiling commenced, but only 25 grams distilled. '

" Another highly interesting paper of Moissan is:
'Sur la distillation des corps simples. ' 1 Here we find
the following statement:

" ' Gold commences to evaporate in vacuo at 1,070°.
It boils in vacuo at 1,800°, and should boil at 760 mm.
pressure at 2,530°, ' thus showing how much depends
upon the pressure under which boiling takes place.
Now all Moissan 's experiments, tabulated above,
are made at ordinary atmospheric pressure, and we
are entirely at loss to say how much the evaporation
of the various metals would have been retarded under
increased pressure, such as we might expect at the
very base of the solar atmosphere, close to the liquid
nucleus.

"Moissan tried, also, the metalloid titanium in his
electric furnace. Five hundred grams were heated
by a current of 500 amperes and 110 volts; after four
minutes vapor appeared, but after five minutes the
stuff was fused only on the surface, and carbide of
titanium had formed. Then 300 grams were treated
with 1,000 amperefe and 55 volts for seven minutes;
110 grams were distilled; the stuff itself, however,

1 Annales de chemie et de physique, (8) 8, 145-181, 1900.
239

THE SUN

had been only viscous, the surface had not become
horizontal. In his book 'Der electrische Of en/ p.
238, he says that even with a current of 2,200 am-
peres and 60 volts, the stuff in the crucible is not com-
pletely fused."

The behavior of titanium is not unparalleled. Some
substances go over from solids to gases without melt-
ing at all. In the second experiment nearly half of
the material was distilled, although the melting was
not complete. No substances are cited which failed
to become largely gaseous under a few minutes heat-
ing at atmospheric pressure with the electric oven.
True, an increase of pressure to five or ten atmos-
pheres, which may prevail in layers we can see in the
sun, would certainly have hindered the evaporation.
But if the electric oven is above 3,500° C., even 4,000°
C., it is still far beneath the photospheric tempera-
ture. For if the solar constant is 1.95 calories, as will
be shown in Chapter VII, the photosphere cannot be
at a lower temperature than 5,860° Absolute Centi-
grade, and may be much higher if its intrinsic radiat-
ing capacity is considerably less than that of the per-
fect radiator. Indeed, it seems most probable that
the photospheric temperature should be set not lower
than 6,500° Absolute. At such a temperature, pre-
vailing not minutes but milleniums, one can most
easily believe all elements are entirely gaseous.

As for the sun being mainly liquid, as argued by
Schulz, the sun's low specific gravity has led even
those who prefer to believe in a cloudy photosphere

VI n

WHAT IS THE SUN?

to regard the interior as almost wholly gaseous. We
return to our discussion.

It is required to explain : 1. Why the sun presents
a sharp boundarj^. 2. Why the enormous radiation
of the photosphere does not so far cool its surface as to
precipitate clouds. 3. Why a more or less definite
structure appears on the sun. 4. Why the spectrum
of the sun is mainly continuous. 5. Why, towards the
limb, there is a gradual decrease in brightness, and an
alteration in spectral distribution. 6. Why the solar
spectrum has dark lines.

Besides these principal requirements, there are a
thousand details of fact not necessary here to re-
hearse, which must not be hopelessly inconsistent
with any satisfactory solar theory. Finally there are
the great problems of the periodicity of sun-spots,
faculse, et cetera, the variations of solar rotation with
latitude, and the supply of the sun's energy.

(1 ) Why the Sun Presents a Sharp Boundary.

In Lord Rayleigh's celebrated mathematical in-
vestigations of the light of the sky he has shown that,
whether proceeding on the hypothesis of the elastic
solid theory of light, or on the electromagnetic theory,
the extinguishing effect on a beam of light of the
molecules of a gas, or of a collection of particles which
are small compared with the wave length of light,

may be expressed by the relation : k = ^ — ;

in which k is the coefficient of extinction, //, is the

241

THE SUN

index of refraction, and N is the number of particles,
or molecules, per cubic centimeter. Schuster has
proved the relation to be independent of theory if /A
is approximately unity. This is true for all gases.
He has applied this quantitative theory of extinction
to the atmosphere.1 For N he uses Rutherford and
Geiger's value, 2.72 x 1019 molecules per cubic cen-
timeter. If h is the height of the homogeneous at-
mosphere, that is, the height to which the atmosphere
would extend if entirely at standard temperature and
pressure, then e~kh is the fraction of light which
would reach the observer if none were lost in any
other way than by molecular scattering. From these
data Schuster calculates the extinction above sea-
level, and above 1,800 meters elevation, and compares

TABLE XVI. — Difference between observed and computed values
of atmospheric transmission

Wave
Length

Washington Observed

Computed
-"Clear"

Ml. Wilson Observed

Computed

-"Clear"

Mean

Clear

Mean

Clear

o's

0.6
0.7
0.8
1.0

0.55
0.70
0.76
0.84
0.87
0.90

0.72
0.84
0.87
0.90
0.94
0.96

—0.01
+ 0.03
0.07
0.06
0.04
0.03

0.73
0.85
0.89
0.94
0.96
0.97

0.76
0.89
0.92
0.96
0.99
0.99

0.00
0.00
0.03
0.01
—0.01
0.00

the computed values with the transmission observed
on days of mean and maximum transparency at
Washington and Mount Wilson, respectively, by
Smithsonian observers.

\ttlnrc, vol. Ixxxi, p. 91
242

1 <>()<).

WHAT IS THE RUN?

Schuster concludes that on a clear day on Mount
Wilson molecular scattering practically accounts for
the atmospheric extinction. Even at Washington
he thinks the major part of the losses in the atmos-
phere may be thus accounted for; although on the
average day something must be attributed to re-
flection and absorption of grosser dust particles.

Professor Natanson has treated the matter from
the standpoint of the electron theory. He differs in
some respects from Ray lei gh and Schuster, although
deriving a practically similar formula for scattering,
for he introduces not the number of molecules but the
number of electrons per cubic centimeter. He also
has compared theory with the observations of Smith-
sonian observers at Washington and Mount Wilson,
and finds an approximate agreement. He does not
state the conclusion in so many words, but his results
indicate that the extinction of light above Mount Wil-
son on the best days may reasonably be accounted for
by scattering of the gas itself without consideration of
dust particles.

All this has apparently a very important bearing
on our views of the sun. The temperature of the
layers from which we get the most light, as already
stated, seems to be certainly in excess of 6,000° Ab-
solute Centigrade. There are no substances, so far as
known, which can exist except as vapors in these con-
ditions. Hence, it seems reasonable to suppose that
the sun contains no solids or liquids, unless perhaps
in sun-spots, and that its substance, as we see it, and

243

THE srx

within the layers we see, is altogether gaseous. But
if this is so, how, it will be asked, can the sun present
a sharp boundary?

According to the theory of Schmidt, which has been
alluded to, this is caused by the effect of refraction.
But if Rayleigh and Schuster and Natanson are right
in attributing a substantial light scattering effect to
gases, Schmidt's theory needs hardly to be invoked,
nor, indeed, can it really be of much application. For
if, as computed by Schuster, the quantity of gas in
the vertical column of atmosphere above Mount Wil-
son is sufficient to scatter from the direct beam of
yellow sunlight six per cent of its light, a column con-
taining seventy-five times as much will suffice to scat-
ter ninety-nine per cent.

Several observers have found that the pressure in
the reversing layer for iron is about five atmospheres.
Assuming the average absolute temperature of the
photosphere to be 6,500°, and that of the air 250°, the
quantity of gas per cubic centimeter in the reversing
layer would be about ^ as great as in air at atmos-
pheric pressure. As the homogeneous atmosphere
above Mount Wilson is less than ten miles high,
seventy-five times the quantity of gas above Mount
Wilson would be found probably within 4,500 miles
of the top of the sun's reversing layer. This estimate
assumes the line of sight radial within the sun, and
regards five atmospheres as the average pressure. If,
as Evershed maintains, the pressure of the reversing
layer is only of the order of one atmosphere, still we

WHAT IS THE SUN?

must admit that the pressure increases rapidly with
the depth, so that still the estimate seems to be ample.

Hence, it seems probable that gaseous scattering
alone prevents us from seeing towards the center of
the sun, when looking directly at the middle of the
solar disk, to more than 5,000 miles below the re-
versing layer.

At the limb of the sun, the direct line of sight to a
position at the same distance radially below the re-
versing layer would traverse fully 60,000 miles of gas.
Accordingly, to obtain our column containing the
requisite quantity of gas for practical extinction of
yellow light, at the limb we should penetrate a
layer which, measured along the radius, would be
very much thinner than that required at the center
of the disk. For, even to a radial depth of only 500
miles, the direct line of sight is almost 20,000 miles.

These considerations seem to point to a reasonable
explanation of the sharp boundary of the sun. For
at the edge of the disk, owing to the oblique line of
sight, gaseous scattering will probably extinguish
almost all yellow light starting from more than 500
miles below the chromosphere, while, an even less
thickness suffices for blue or violet light. It is plain
that an indistinctness of outline corresponding to a
layer of this depth would not be readily recognized
on the solar image, since it corresponds to only about
one second of arc. Furthermore, the direct line of
sight takes in not only the nearer, but the further
solar hemisphere as well. A still thinner stratum

245

THE SUN

than 500 miles would, therefore, suffice to contribute
all the light that can be contributed to the beam di-
rectly along the line of sight. We therefore con-
clude that within a small part of a second of arc below
the reversing layer the sun would appear as a solid
body, even though entirely gaseous.1

(2) Why Is there No Cloudy Photosphere?

But it is said by Young and many others that a
cloudy photosphere must certainly exist as the result
of the juxtaposition of the hot gases of the sun with
the cold of space. Without falling back on the strong
reply that the apparent temperature of the so-called
photosphere exceeds 6,000° Absolute Centigrade, and
that no known substances can exist except as vapors
at that temperature, it may be asked whether the ab-
sence of a cloud immediately above the smoke-stack
of a locomotive in winter does not show that such a
juxtaposition of hot gases and cold surroundings with-
out forming a cloud is entirely possible. There is no
cloud formed immediate^ above the smoke-stack
because the steam there is superheated above the
boiling point. It may be urged that a little time is,
of course, required to form the cloud, and that, owing
to the rapid motion of the steam, it is carried a little
above the smoke-stack during this interval. But
this is really admitting that while the steam remains
superheated it will not form a cloud, so that all that is

1 For practical purposes of seeing, it is not the depth of the layer
which scatters ninty-nine per cent., but a much less fraction that is
in question.

246

WHAT IS THE SUN?

necessary to prevent a cloud is to supply heat to the
steam as fast as heat escapes from it, and thus to keep
it superheated.

Such a state of affairs seems to exist in the sun.
Heating is communicated from the interior to the sur-
face layers fast enough to maintain the latter above
6,000°, notwithstanding their radiation to space, and
at this temperature no cloud forms. The convey-
ance of heat from within is probably almost wholly by
repeated radiation, rather than by vertical convection
currents.1

(3) What, then, Is the Cause of the So-called ' ' Rice-grain
Structure'' on the Sun, if there Are No Clouds f
It is not to be supposed that the communication of
heat from within outwards is perfectly uniform at all
parts, for, as evidenced by the sun-spots, the prom-
inences, and the corona, there are marked defects of
homogeneity in the sun. Hence, it may readily be
supposed that some regions of the gas are a little hot-
ter than others, and that these differences of tempera-
ture will give rise to differences of brightness. By the
radiation laws, the increase of brightness is far more
rapid than the corresponding increase of temperature.
Professor J. Scheiner published, in 1895, a theory
of the solar granulation which seems very reasonable;
and which, if we consider the effects produced to be
merely regions of local cooling without actual con-
densations, would suit the theory of the altogether-

1 This is the view of Schwartzchild and also of See.
18 247

THE SUN

gaseous sun as well as it does the theory of the cloudy
photosphere.

Professor Scheiner says (quoting from a translation
in the Astrophysical Journal) : "According to the the-
ory of Helmholtz, air waves are produced when two
layers of air, differing in temperature (i. e. in density),
glide past each other, just as waves are produced by
the gliding of air over water. If the lower layer is
nearly saturated with aqueous vapor, condensations
will take place in the wave crests on account of the
diminution of pressure. Under these circumstances
the elevations or -wave crests appear as clouds, the
depressions or troughs as transparent interspaces,
and thus a more or less regular series of cirrus clouds
is produced. If the impulses resulting in wave forma-
tion act in two different directions the waves cross,
and we have the cloud effect known as a mackerel sky.
The great similarity in appearance between the solar
photosphere and terrestrial cirrus has long been rec-
ognized, and there is no doubt that the necessary con-
ditions for the application of Helmholtz 's theory to
the solar atmosphere — the existence of layers of dif-
ferent temperature, the over-saturated state of con-
densable gases (in the photosphere), and variously
directed currents in the different layers — are found in
the sun. I therefore regard the bright grains of the
photosphere as wave crests, rendered visible by con-
densation, or at least an increase of condensation, of
two crossing series of waves. "

We may adopt Scheiner 's view in the present dis-
248

WHAT IS THE SUN?

cussion, only not admitting actual condensation.
Hence, his bright grains would be our dark ones, be-
cause the cooler regions would radiate least. The
reader will see that this amendment to Schemer's in-
terpretation is rendered at least plausible by the fact
that spectroheliograms show bright and dark hydro-
gen flocculi, and of course no such a thing as a con-
densed hydrogen cloud can be thought of at solar tem-
peratures.

(4) Why Is the Sun's Spectrum Mainly Continuous?
Gases are noted for giving only line spectra, while

the solar spectrum is, on the contrary, chiefly a con-
tinuous spectrum crossed by absorption lines. In
reply to this objection it may be said that gases under
pressure give more and more continuous spectrum
along with the bright lines, even in layers of small
thickness, like those operated on in the laboratory.
(See Plate X VIII . ) Think, then, if layers many miles
thick, and under pressures of at least several atmos-
pheres, may not give a fully continuous spectrum.

(5) Why Does the Limb Fall Off in Brightness and
Grow Redder f

As stated above, the light received from near the
edge of the solar disk comes, on the whole, from more
superficial layers than that received from the center
of the disk; because at the edge we look obliquely,
and hence by a longer path, into the sun, and the scat-
tering of the molecules cuts off the view before the
deeper layers seen at the center are reached. At the

249

THE SUN

edge, the layers which are emitting light to us, being
more superficial, and hence cooler, will in consequence
give less intense light than those at the center.

Referring to Tables 7 and 8, Chapter III, it is pos-
sible to compute, either by Stefan's law or by Wien's
law, the change in effective temperature required to
account for the decrease of brightness towards the
sun's limb. As shown in Table 8, the two methods of
computation are in close accord. Extending the
result somewhat, we have the following differences of
temperature, assuming the central disk temperature
6,400° Absolute Centigrade. These may be compared
with the corresponding differences of elevation of

the lowest observable layer, assuming a depth of — -

100

radius, or 7,000 kilometers, as the limit of visibility
at the center of the disk.

Fraction of radius

from center of

sun's disk

0.0

0.1

0.2

0.3

0.4

0.5

Decrease of temper-

ature

0°

20°

45°

80°

115°

160°

Increase of elevation

of farthest visible

layer

Okm.

66km.

140km.

315km.

545km.

930km.

The small temperature gradient of the order of
1° C. per kilometer of change of level1 required for
this line of explanation seems no greater than we
should expect to exist in the sun's outer layers.

As scattering is greater for violet than for red rays,
the violet rays will come, on the average, from more

1 Mean radiating level, not lowest visible level.
250

WHAT IS THE SUN?

superficial layers than the red, both at the center and
edge. Accordingly, the diameter of the sun should be
greater if measured in violet light than if measured in
red, so far as this consideration goes. But the differ-
ence of diameter due to this cause is probably too
little to be measured. It is obscured by " boiling" of
the sun's image, diffraction, and scattered light in the
earth's atmosphere, any one of which alone probably
produces a greater effect at the limbs than that we
are considering. According to Planck's formula, the
change of intensity of radiation accompanying change
of temperature of the radiating source is greater pro-
portionally for short wave lengths than for longer
ones. Hence, it follows that the violet should be
weaker with respect to the red at the limb than at the
center of the sun. This is in accord with observation.
Whether this effect would be augmented or dimin-
ished in consequence of the fact that the effective
radiating layer for violet radiation is nearer the sur-
face than that for red at both center and edge, de-
pends on the relative change of temperature due to
this shifting of depth at the two regions. It seems
impossible as yet to determine how this would be.

(6) Why Has the Solar Spectrum Dark Lines f

All the Fraunhofer lines would really be bright if
seen against a dark background.1 They are dark only
relatively to the brighter continuous spectrum. In

1 Different persons estimate their brightness as from one-fifth to
one- tenth that of the continuous spectrum background.

251

THE SUN

these lines the selective absorption of radiation is very
powerful, and cuts off all transmission within a short
distance, so that, as. compared with the continuous
spectrum, they are emitted very near the surface of
the sun. This superficial layer in which they arise is
cooler than that which lies behind, hence its emission
is less intense, and hence the comparative darkness of
the Fraunhofer lines. As between the center and the
limb of the sun we should expect little change in the
absolute brightness of the Fraunhofer lines, because,
owing to the powerful selective absorption within
them, they are very markedly superficial phenomena
both at center and limb. Thus, but little change in
the effective depth and temperature from which they
are emitted occurs, no matter from what angle the
surface of the sun is viewed. It is not so with the
process of weakening by scattering, which requires
great thickness of gas; and hence, as we have seen,
the continuous spectrum is brighter at the center of
the sun than at the limb. Consequently the contrast
or " intensity" of Fraunhofer lines falls off towards
the limb, because they change little, while the back-
ground against which they are seen, falls off in bright-
ness.

Why Are Not all Chemical Elements Impartially
Represented by the Intensities of Their Solar Lines f

It is not to be inferred from what has been said
under (6) that there is no thickness to the " reversing
layer," or no change of its effective thickness from

WHAT IS THE SUN?

the center to the edge of the sun, but only that,
relatively to the effective thickness of the layer
which furnishes the continuous spectrum at the
center of the sun's disk, the reversing layer for
any one element is thin. Hence, we may dis- •
tinguish between high level and low level spectrum
lines. It would be expected a priori that elements
(a) of high atomic weight, (6) of high vaporizing
temperature would be found at lower levels, and
(c) that of the spectrum lines of a single element the
longer wave lengths would, so far as depending on the
relations of temperature and emission, represent
higher levels.1 It might perhaps be expected that the
reversing layer for a heavy element could lie wholly,
below that of a light one. For very low lying ele-
ments it might conceivably occur, through scattering,
that their entire spectra would disappear at the edge of
the sun, although appearing at the center. In general,
low lying elements would give weak solar spectra,
because the temperature of the emission of their lines
would more nearly approach the temperature of the
emission of the continuous spectrum background.

Referring to Chapter III, the reader will recall the
marked connection between atomic weight and in-
tensity of solar spectra. On the whole, the elements^
of less atomic weight give the strongest solar spectra.
The platinum group, of very high atomic weight, on
the other hand, is only partly represented in the solar

1 The effect of scattering would tend in the other direction, how-
ever.

253

THE SUN

spectrum. Rowland and Tatnall say,1 speaking of
these elements: "The heavier lines have been exam-
ined as to the probability of their occurrence in the
solar spectrum, and investigation has confirmed the
existence of rhodium and palladium in the sun. Ruth-
enium is doubtful" (afterwards confirmed) "and it is
most probable that there are no solar lines of appre-
ciable intensity belonging to platinum or osmium in
this region of the spectrum" (X3,000 to 4,000). "The
most intense lines of the arc spectra of rhodium and
palladium correspond to extremely weak solar lines."
This failure of solar lines is not for lack of strong lines in
the arc, for Rowland and Tatnall give many platinum
•arc lines of intensities 5 to 15, to which there are cer-
tainly no corresponding solar lines above intensity 00.

The comparison of intensities and atomic weights
given in Chapter III has some glaring discrepancies.
Carbon is found near lanthanum, although its atomic
weight is but 12. It is now believed that solar lines
attributed by Rowland to carbon are really due to
carbon compounds of considerable molecular weight,
notably to cyanogen. Glucinum and potassium fall
in strange company. But they have only one or two
lines each identified by Rowland, and these may lead
us into error. Indeed, Kayser and Runge question
the existence of potassium lines in the photospheric
spectrum.

In the flash spectrum at eclipses we have another
indication of differences of level. There again, as

1 Astrophysical Journal, vol. ii, p. 184, 1895,
254

WHAT IS THE SUN?

shown by Evershed, Lockyer, Jewell, Mitchell, and
others from measurements of the lengths of flash
spectrum arcs, the order of level agrees on the whole
with that which we have just considered.1 Again, in
Adams' work on the solar rotation, if we grant (as we
must when we recall the relative rotational velocities
observed through red hydrogen (Ha), calcium
(\4,227), and iron lines) that lower levels correspond
to slower velocities, then we find (CN2) and lan-
thanum falling in at lower levels than iron and titan-
ium, just as they appear to do from considerations of
Rowland's intensities.

The absence of lines of helium, the halogens and
other negative elements in the photospheric spectrum
is probably due to the extinguishing effect which the
metals appear to produce on the lines of such elements
when the metallic and other gases are mixed. Thus,
according to E. Wiedemarm, nitrogen and hydrogen
lines first begin to appear in a vacuum tube showing
mercury lines when the concentration of these gases
is thirty per cent. Also, common salt in a flame shows
the spectrum of sodium alone, not of chlorine.

What Causes the Differences of Character and Wave
Length for Fraunhofer Lines between the Center
and Edge of the Sun ?

The reader will recall that, after allowing for the
rotation of the sun and for an apparent general rise

1 The extremely high level of calcium H and K lines is an
anomaly not well understood.

255

THE SUN

of the brighter material toward the solar surface,
Adams l confirmed Halm's and Buisson and Fabry's
results that there is a general displacement toward the
red of the centers of most solar lines as seen near the
limbs and compared with the center. This displace-
ment is inappreciable for the more prominent lines
of hydrogen, calcium, sodium, and magnesium, and
small for the other lines of these elements. Also for
elements of high atomic weight the shifts are very
small. Iron and nickel lines show larger shifts than
those of titanium, vanadium, and scandium. En-
hanced lines as a class show larger shifts than arc
lines do. Lines strengthened at the limb show small
shifts. The displacements are greater for long wave
lengths than for short. The character of lines at the
limb is also altered. Some strong lines of the elements
hydrogen, sodium, calcium, silicon, magnesium,
aluminum, iron, chromium, titanium, and manganese
lose partially or wholly the winged appearance which
they have at the center. Many lines of all kinds of
elements are slightly widened. The enhanced lines
and lines of elements of high atomic weight are gen-
erally much weakened.

The weakening of high temperature and low lying
element lines may be attributed to scattering. At
the center of the sun we look straight down upon the
lower reversing layers and get their rays under more
favorable angles of scattering than at the limb, where
they, in order to contribute to the line of sight, must

1 Contributions of the Mount Wilson Solar Observatory, No. V.\.
256

WHAT IS THE SUN?

scatter by one or more reflections through a right
angle, or nearly so. Hence the continuous spectrum
at the limb encroaches upon their lines, diluting them
with stray light. Furthermore, what is at least
equally important, the sun's continuous spectrum is
weaker at the limb, for reasons already considered,
and would give the lines less contrast, even without
the effect we have just considered. This weakening
of the continuous spectrum at the limb contributes-
powerfully to reduce the visibility of the wings of
lines, also, because the wings are seen against a back
ground which, towards the limbs, approaches more
and more their own strength of emission. The widen-
ing of lines seems possibly a promiscuous Dopplei
effect due to their being contributed to by different
levels rotating at different velocities.

Adams explains the superior displacements of en-
hanced lines by suggesting that at the center of the
sun, the higher temperature gases are rising, the
cooler ones falling, giving for the spectrum lines in
general a rising effect; because most of the light
comes from the brighter emitting matter which is
rising. But for the lines which are high temperature,
or enhanced, lines a maximum rate of rise (greater
than that of average lines) is observed, because the
descending cooler vapors do not emit or absorb en-
hanced lines, so that for these lines there is a displace-
ment of central spectra towards the violet which ap-
pears as an increased displacement of edge spectra
towards the red. Of course, at the limb these mo-

257

THE SUN

tions of rise and fall are at right angles to the line of
sight, and, therefore, produce no Doppler effects.

Having thus cleared the ground of Doppler effects
of rotation and rise, Adams attributes the remaining
displacements to pressure depending on level. High
level lines are not displaced because emitted under
slight pressure both at the edge and center. Low
level lines are not displaced because they arise only
from thin strata at the very bottom of the layer
which is visible to us, and which must be at nearly
equal, though high pressures at both center and
limbs. Scattering does not permit us to see much
beyond the outer boundaries of such strata at center,
and at the limb we see them only faintly, and after
the rays have been one or more times reflected, hence
such spectrum lines are weak at the limb. Lines of
intermediate levels are under higher effective pres-
sures at the limb than at the center, according to
Halm's view, as adopted by Adams, because any line
of sight drawn just inside the limb has a longer rela-
tive path in the lower layers it cuts than a line of
sight drawn near the center of the disk has in the
corresponding layers. Hence, lower layers contrib-
ute proportionately more to the spectra of inter-
mediately lying elements at the limb than at the
center.

The writer must confess that he feels a little hesi-
tancy about adopting this last argument, because he
thinks that it would be necessary to consider for these
layers quite as much the rays scattered into the beam

258

WHAT IS THE SUN?

from all sides as to consider merely the line of sight.
But until the proportions contributed to a beam by
scattering from different distances, and at different
angles, and the rate of change of density along the
sun's radius, are better known than now, it seems idle
to press this objection.

Julius has explained the displacements of lines
towards the red as a simple consequence of anomalous
dispersion. But Adams shows that the lines apt to be
most powerfully affected by anomalous dispersion
show no shifts at all, and that a comparison of all the
known data as to the strength of anomalous disper-
sion for the several lines with their observed shifts at
the limb yields nothing to recommend this explana-
tion of Julius.

Why Do the Prominences and the Chromosphere Give

Bright Line Spectra?

According to the line of explanation we are pursu-
ing, the gases of these appendages of the sun are in a
condition of extremely low pressure and density, and
do not contain sufficiently many molecules contrib-
uting radiation to the line of sight to emit a strong
continuous spectrum. But for the spectrum lines of
powerful selective emission, their radiation is suffi-
ciently considerable to reveal their forms.

What of the Characteristic Forms and Occasional
Immense Velocities of the Prominences f
Although not shared by all, there has always been a

hesitancy among many of those who regard prom-

259

THE SUN

inences as real protruding masses of bright gas, and
not, after Julius, as mirage effects, to trust their ob-
servations, both direct and spectroscopic, that these
gaseous masses are bodily projected at such rates as a
hundred miles a second. It is hard to imagine on
purely mechanical grounds how such velocities could
arise. In spectroscopic determinations the motion
observed in prominences is apparently tangential to
the sun. W. A. Michelson of Russia has suggested
that in this case we may really have moderate motion

across the line of sight,

as illustrated in the accom-
FlG 58 panying Fig. 58. Let as be

a line of sight, and let a

mass of gas whose front is AA, rise to positions BB
and CC. Then the source of light moves effectively
from a to b to c, giving an apparently enormous
motion in the line of sight, which is really a much
smaller motion across the line of sight. Whatever
may be thought of this explanation, it, of course, has
reference only to apparent enormous tangential ve-
locities.

As for apparent enormous radial motions, we all
frequently see wisps of cirrus clouds stretch across the
sky in a twinkling, as it were. This does not, of
course, indicate motion of translation from one end
of the wisp to the other, but rather the rise of a
trough of cooling, which causes a precipitation of
cloud almost simultaneously along its whole length.
Young and others state that detached prominences

260

WHAT IS THE SUN?

sometimes form without preliminary attachment to
the sun's chromosphere. Perhaps eruptive promi-
nences are formed somewhat as cirrus clouds are, by
the rise from the sun of some disturbance fitted to
arouse emission almost simultaneously within large
masses of previously non-luminous hydrogen and
calcium gases, which lie where a prominence is about
to appear. Perhaps an electrical excitation would be
most reasonable. The apparent tremendous velocity
of the outbreak could then be explained by reference
to the accompanying dia-
gram, Fig. 59. Let the line
of sight be in the line of
the arrow I, AB the photo-
sphere, and ab the trough which suddenly arouses
emission beginning at a and proceeding almost im-
mediately to 6. The lower end, a, is lost in the glare of
the photosphere, and the prominence appears to rise
from c to b in the very brief time needed to extend to
b the influence of the trough. If the line of sight
had been in the direction of the arrow II the promi-
nence would have appeared detached. The writer
does not venture to recommend this suggestion very
strongly.

Professor E. Pringsheim1 has suggested an explan-
ation of these enormous observed prominence veloci-
ties that appears very reasonable. He refers to ex-
periments of J. Stark, who has found Doppler dis-
placements of the order of magnitude which occur in

1 E. Pringsheim, " Physik der Sonne," Leipzig, 1910, pp. 225-228.
261

THE SUN

prominences, when observing the so-called " canal
rays," which are curious electrical discharges ob-
tained through rare gases by special contrivance.
This indicates that the positively charged atoms,
which give the light, may travel with velocities like
those which appear in the prominences, when forced
by ordinary differences of electrical potential in very
rare gases. Pringsheim fchen draws attention to the
fact that the comet of 1843 passed at perihelion within
3' or 4' of the photosphere (TJT the solar diameter)
without being affected by the resistance of the
material encountered. This proves the existence of
a sufficient, indeed, of an extraordinary degree of
vacuity there. It is not known that sufficient
variations of electrical potential exist in the sun's
neighborhood, but those which exist in the earth's
atmosphere are abundantly sufficient to drive elec-
trons with prominence-like velocities in vacuum, ac-
cording to Pringsheim's computations. The exist-
ence of similar potential gradients near the sun
seems not improbable.

Julius's explanation of prominences through anom-
alous dispersion we have already noted, but it
requires us to admit the propagation of disturbances
to the apparent tops of the prominences, and to
believe that the gases at such enormous heights are
dense enough to produce appreciable anomalous re-
fraction.

262

WHAT IS THE SUN?

What Is the Corona?

Comets pass through the corona without sensible
retardation. Hence, its matter, whether it be purely
gaseous or partly meteoric dust, must be very rare.
Since the corona gives the Fraunhofer spectrum in its
outer part, it must contain reflected photospheric
light. Its substance in its inner parts may well be hot
enough, by virtue of proximity to the sun, to give
light of incandescence. Its form suggests the auroral
streamers, and inclines one to think that, as the au-
roral light is of electrical luminescence, so may a part
of the coronal light be. As the aurora gives bright
lines in its spectrum, so, also, does the corona. The
proportions of the mixture of these three varieties
in coronal radiation is unknown, but, according to
bolometric work, the mixture gives almost the same
spectral distribution in the inner corona as photo-
spheric radiation. This suggests that light of lumi-
nescence and of reflection together predominate over
incandescence.

On the other hand, the results obtained by repre-
sentatives of the Lick Observatory at several eclipses
have led Campbell, Perrine, and Lewis to express the
opinion very definitely that the inner corona shines
mainly by ordinary incandescence, due to the heating
of its particles on account of the absorption by them
of photospheric radiation. Neglecting the bolometric
results, this conclusion would be perfectly reasonable
and it may yet prove that there is some error in these
19 263

THE SUN

latter results which may explain the discrepancy.
Yet the conditions of the bolometric work at Flint
Island were so satisfactory, and the observations so
concordant, that this seems rather improbable. Per-
haps some new line of explanation may suit all
parties.

The electrical explanation of the coronal form and
brightness is receiving much attention. Professor
Pringsheim devotes much space to it in his new work,
"Physik der Sonne." In a recent article, however,
Professor R. W. Wood1 has sought to explain both
the polarization and light emission of the corona as
the effects of the passage of the powerful sun-rays
through comparatively cool metallic vapors, thereby
exciting fluorescent light in them. By laboratory
experiments with light so excited in vapors of sodium,
potassium, and iodine, he finds the percentage of po-
larization similar to that in the corona. He states
that the spectrum of mixed vapors would be contin-
uous, at least for low dispersion. The fluorescent
spectrum is, in fact, made up of thousands of fine
lines arranged in groups and bands, and gives no re-
semblance to the bright line spectra of the same ele-
ments. These lines lie so closely packed as probably
to escape detection with low dispersion spectroscopes.
Any color of fluorescence may occur, according to the
kind of vapors mixed, and their proportions. Wood
thinks it quite possible that the coronal green line is

1 Astrophysical Journal, vol. xxviii, p. 75, 1908.

WHAT IS THE SUN?

not a bright line of some unknown substance, but
rather a yet unrecognized fluorescent line from some
well-known element.

Schaberle has long maintained a mechanical erup-
tion theory of the coronal form.1 He traces back the
probable courses of the streamers, and locates them
in centers of eruption on the sun's disk. His views
agree well with the hypothesis that the coronal bright-
ness is mainly of incandescence.

The cause of the change of form of the corona with
the sun-spot cycle is unknown.

Importance of Temperature.

It will be noted that in the solar hypotheses we are
recommending the temperature plays a most promi-
nent part. First of all, the existence of a cloudy pho-
tosphere is denied because the temperature of the
photosphere is shown probably to reach 6,500°, for it
is highly improbable that solids or liquids can exist in
such conditions. Secondly, the presence of the so-
called " granulations " is regarded as evidence of dif-
ferences of temperature in the radiating gas — dif-
ferences which would naturally be expected in an im-
mense globe of gas giving off tremendous amounts of
radiation from its surface, and known to present irreg-
ularities of rotation and cyclonic motions in addition.
Thirdly, the darkening towards the limb is regarded
primarily as a temperature effect, secondarily due to

1 See Lick Observatory Contributions, vol. iv, 1893.
265

THE SUN

scattering. Owing to scattering, the effective radi-
ating layer must necessarily be nearer the surface,
and hence cooler, at the limb than at the center of the
disk. We say it must be nearer the surface: For,
travelling obliquely, a ray must become extinguished
by scattering in the gas at the limb, before it reaches
the same radial depth that it does if travelling radi-
ally at the center. Fourthly, the darkening at the
limb would naturally be greater for violet than for
red rays, firstly, because with all incandescent bodies
a fall of temperature causes more decrease of radiation
for short rays than for long; and, secondly, because
molecular scattering is greater for violet rays than for
red, and hence at the sun's edge the effective radiating
layer for the violet will be more near the surface than
will that for the red. Fifthly, the Fraunhofer lines are
regarded, not as dark, but as very bright, intrinsi-
cally. They only appear dark because, owing to pow-
erful selective absorption of the gases which give rise
to them, they cut off completely the light from be-
hind, and the observer sees only a relatively thin and
superficial layer of the sun, when viewing it by the
light of the Fraunhofer lines. The reversing layer is
hence colder, and its radiation less intense than that
of the continuous spectrum background which comes
from deeper layers of the sun. Sixthly, the contrast
of the Fraunhofer lines with the background of spec-
trum decreases as our view approaches the edge of
the sun's disk, because the Fraunhofer line region is so
thin and superficial that its temperature is nearly the

266

WHAT IS THE SUN?

same at the edge as at the center; whereas for the
continuous spectrum background, the effective radi-
ating layer rapidly approaches the sun's surface as we
look nearer the limb, and hence its radiation de-
creases, owing to the fall of temperature.

What of Sun-spots ?

We now trace the importance of temperature in the
explanation of sun-spot phenomena. In accordance
with the Mount Wilson observations of Hale, Eller-
man, and St. John, we may regard sun-spots as vor-
tices, and, as indicated by the spectrum work of Ever-
shed, we must conclude that in the Fraunhofer line
region the motion along the spiral is from within out-
ward. We may imagine that these vortices are sim-
ilar in form to water-spouts seen at sea, with the
trumpet-shaped part at the top, and the whirl carry-
ing matter from below outward. In such circum-
stances there would be a great cooling of the gases,
owing to their rapid expansion as they approach the
limb. This cooling (as appears from the discovery of
lines due to the copious presence of calcium and mag-
nesium hydrides, arid also of titanium oxide in sun-
spot spectra) carries the temperature down to per-
haps 3,500°, which is low enough for the formation of
liquids, and perhaps some solids.

These dissimilar substances, by their friction (per-
haps even by their very formation) we suppose may
give rise to charges of electricity, which, being carried
round rapidly in the stem of the vortex, produce the

267

THE SUN

effect of currents of electricity as shown by Rowland.
Hence they give 'rise to the magnetic field, which
Hale finds is a feature of a sun-spot. The top of the
vortex, we assume, corresponds in level nearly with
the upper Fraunhofer line region. There the cooled
matter spreads out some distance in spirals which
grow in radius so rapidly as to be almost radial to the
umbra. As there is no longer further rise and expan-
sion, at length the matter becomes warmed, by con-
tact, to the temperature of the surroundings. The
stem of the vortex is the umbra of a sun-spot, the
spreading top is the penumbra.

The peculiarities of the sun-spot spectrum and the
causes of these peculiarities have been dealt with at
considerable length in Chapter V. We may summar-
ize them as the peculiarities attending, (a) diminished
temperature as compared with the photosphere, (6)
the action of strong magnetic fields. The sun-spot
spectrum has been shown by Hale and Adams to be
the type of the spectrum of the red stars. Since we
now know that, as regards the characteristics for
which this comparison holds, the sun-spot spectrum
results from the mere cooling of the photospheric
material, this relation is very significant, and indi-
cates distinctly one step in the process of stellar evo-
lution. We shall recur to this in Chapter X.

We have noted particularly in Chapter V the re-
markable behavior of the hydrogen lines in sun-spots.
They are all weakened, and the shorter wave-length
lines most weakened, as cohipared with the photo-

268

WHAT IS THE SUN?

spheric hydrogen spectrum. The lines of other ele-
ments are generally strengthened in spots. This
anomaly seems explainable as due to the high level of
hydrogen. This gas is relatively unaffected in posi-
tion, and in fact, as St. John observed, is sucked in-
ward and downward rather than whirled outward
and upward by the cyclonic motion in sun-spots.
Hence, its temperature and (in consequence) its radi-
ation is rather increased than lowered by the presence
of spots. The continuous spectrum background
against which we see the hydrogen lines is, however,
weakened in spots, and thereby the contrast of the
hydrogen lines is diminished. In other words, they
are weakened. Owing to the lower temperature, the
energy spectrum, that is, the continuous spectrum
background, in sun-spots as at the sun's limb, is
weaker in the violet as compared with the red than is
the ordinary solar spectrum. Thus, in spots, the radi-
ation in the violet hydrogen lines approaches more
nearly the brightness of the spectrum background
.than that in the red lines. Hence, the comparatively
greater weakening of the shorter wave-length hydro-
gen sun-spot lines follows.

In the center of the sun-spot vortex there is a ten-
dency to form a vacuum. Into this partial void is
sucked the superincumbent matter, which is the high-
level hydrogen of the chromosphere and prominences.
Hence occurs the inwardly directed radial motion of
this gas shown by the Ha spectroheliograms at Mount
Wilson. Between the Ha level gas, which is going in-

269

THE SUN

ward, and the Fraunhofer line gases, which are going
outward, there must exist a quiescent region. Hence,
the lower level hydrogen and the HI and H2 cal-
cium spectroheliograms show little or no evidence of
stream lines or other phenomena of stream motion.
The failure of Adams to discover differences in the
pressure of the reversing layer over sun-spots may
be regarded as confirmatory of the superficial char-
acter of the reversing layer, and of the absence of
either elevation or depression in the general sun-spot
level.

As for the cause of the formation of sun-spots, that
is all conjecture. They are generally preceded by
faculae and, according to Fox, by eruptive promi-
nences. Perhaps the faculie, which on our tempera-
ture hypothesis we regard merely as regions of su-
perior temperature, may be formed first, owing to the
presence above them of prominence or coronal mat-
ter. Such formations above would impede radiation,
and hence would cause the regions below to be over-
heated. Being overheated, they would tend to ex-
pand, and by expansion would cause the rise of ma-
terial from below, owing to reduced pressure above.
In this outflow a rotation would usually be set up,
just as in the escape of water from a spout, and thus
the sun-spot would be formed. Once formed its
vortical motion would tend to continue, and would
naturally remain for considerable time. Hale has
noticed that the vortices of most sun-spots of the
southern hemisphere go in one direction and those in

270

WHAT IS THE SUN?

the northern in the opposite. This indeed is what
would be expected in consequence of the different
rates of rotation of the sun at different latitudes.
But it would also be expected that accidental local
circumstances, irrespective of this general cause,
might sometimes determine the rotation in opposite
senses. This also is in line with observation.

As to the general cause of the periodical changes (1)
of the form of the corona, (2) of the areas of faculaB
and (3) of the sun-spot numbers, these also, are things
as yet altogether uncertain. We have already noted
Halm's theory of sun-spot periodicity as a conse-
quence of internal conditions. Schuster and others
have suggested exterior influences as the operative
causes of the periodicity. For instance, the periodic
returns of swarms of meteorites, and the periodic
returns of certain planetary configurations have been
mentioned.

As for the variable rates of solar rotation at dif-
ferent latitudes and depths, these have been regarded
by Wilsing, Sampson, Wilczynski, and Moulton as
vestiges of some ancient actions in which the sun fig-
ured with outside celestial bodies.

What Supplies the Solar Energy f

Lastly comes the greatest problem of all: What
maintains the solar temperature despite the sun's
enormous losses by radiation? These losses stagger
expression in figures. At 90,000,000 miles (145,000,-
000 kilometers) the average radiation is about 'two

271.

THE SUN

calories per square centimeter per minute. Hence
the total emission of the sun is about

22
2X4X y X (14,500,000,000,000)2 calories per minute!

Ordinary fires of coal are kept up by the combination
of carbon with oxygen. Except in sun-spots no com-
binations are going on in the sun. It is so hot there
that most compounds would separate into their ele-
ments, instead of elements uniting with the evolution
of heat. If the sun had no continuous supply of heat,
but, like a piece of metal lying on the blacksmith's
anvil, had been cooling off, there would have been
a marked decrease of the earth's temperature within
historical times. Geologists show that the earth has
not varied more than a few tens of degrees from the
present temperatures for probably 50,000,000 years.
Indeed, in that remote past the earth's temperature
appears to have been a little higher than it is now.
Assuming that the sun emitted its present quota of
radiation during all that interval, the problem of its
source of supply has been, at least until very recently,
insoluble. Since the discovery of the breaking up of
radio-active materials to produce elements of lower
atomic weight with the evolution of heat, as for in-
stance in the production of helium from radium, per-
haps no such difficulty ought to be regarded as in-
superable. It is objected that radium and uranium
lines are not found in the solar spectrum. We have
seen, however, that the lines of the elements grow
more and more feeble as the atomic weight increases,

WHAT IS THE SUN?

and that this seems to be due to the fact that the
heavy elements lie at low levels in the sun. Hence it
is not surprising that uranium (238.5) and radium
(226.4) should not show spectrum lines even if these
elements are present in the sun. Dr. G. F. Becker,
however, thinks radium and uranium to be elements
which form only at temperatures much below those
of the sun. At all events, it is more satisfactory, if
possible, to account for the solar heat by known
causes, rather than to invoke radio-activity of un-
discovered materials.

There are certain circumstances of geology which
may indicate a diminished radiation of the sun in an-
cient times. Although palms used to flourish in the
arctic zones, it does not appear that the tropics were
then much hotter if any than now. As Manson in-
sists, this uniformity of climate from the poles to the
equator seems hard to reconcile with the present
zonal distribution of temperature, if the sun were then
as now the principal source of heat, and its effects
then, as now, zonally distributed. On the other
hand, there is accumulating evidence that glaciation
has occurred more than once over great regions of the
tropics, and most notably in the Permo-Carbonifer-
ous period. In that remote period, far antedating
the so-called " glacial" or Pleistocene period of com-
paratively recent times, glaciation prevailed in Aus-
tralia, Southern Africa, Hindustan, and perhaps in
other tropical regions. It was no mere sporadic
mountain-top affair, but probably a phenomenon of

873

THE SUN

more imposing extent than even the glaciation of the
Pleistocene Period.1 As will be shown in the next
chapter it seems very difficult to see how such a sub-
tropical glaciation as this could have come about if at
that time the sun's output was substantially as great
as now. It does not help us to suppose that the poles
of the earth were then shifted so that these countries
were sub-arctic. The area involved is so vast that the
glaciation would still extend further from the sup-
posed pole than did that of the Pleistocene period.
Besides, this would bring one pole in the vicinity of
Mexico, and the Permian deposits of Texas do not
justify the inference of a polar climate there.

It seems worth considering if the Permian and the
still earlier tropical glaciations which geologists are

1 According to Chamberlin and Salisbury (" Geology," volume ii,
pages 636, 634. Henry Holt & Co., 1906): "The known Permo-
Carboniforous glaciation of Australia, India, and Africa is found in
two zones, the one north and the other south of the equator. In
neither zone have the limits of glaciation been accurately deter-
mined, but in the former it is known to have extended from latitude
18° to about 35° and probably still further north, while in the latter
it is known to have extended from latitude 21° to 35°. In an
equatorial zone about 40° in width glaciation has not been dis-
covered. The glaciation of these various countries has a range of
about 130° in longitude. Glacial conditions must therefore have
prevailed over an area, or at least about the borders of an area many
times as large as that covered by ice in the northern hemisphere
during the Pleistocene glacial period." Speaking of the Australian
glaciation they say: "It is not to be understood that the phenom-
ena here described are restricted to high altitudes; rather they are
known chiefly at low levels, descending in some places nearly to the
sea. The altitude of this region is not only low now, but it was
probably low during the glaciation as shown by the relation of the
glacial deposits to the marine beds."

2*3 I

WHAT IS THE SUN?

now recognizing and also the generally prevailing
similarity of polar and equatorial climates in early
epochs do not all point to one of the following
hypotheses :

(A) Perhaps the sun in those early times was not
so nearly exclusively as now the earth's source of heat,
and the earth itself still retained so much heat that its
life was practically independent of the sun except for
light. In a later chapter we shall see that under
other favoring conditions by no means all of our pres-
ent light supply is necessary to promote maximum
plant growth, and that the red end of the spectrum,
which would suffer least reduction by a decrease in
the solar temperature, is highly efficient for plant
growth. Perhaps, then, the sun has been gradually
growing in temperature and emission, and in the Per-
mian times had not then become the practically ex-
clusive source of heat to the earth's surface. We
may, then, briefly consider if Permian glaciation was
perhaps due, as Manson has suggested, to a very mod-
erate elevation of land areas within a region of a still
prevailing low-lying cloud mantle, with accompany-
ing snowy precipitation. The great, and it seems to
me insuperable, difficulty which this hypothesis en-
counters is to explain in any reasonable way how the
earth's temperature could be maintained for millions
of years without depending so completely on the sun
that the explanation of the uniformity of climates
fails. Furthermore the aridity of climate indicated
by the great Permian deposits of salt and gypsum

275

THE SUN

does not speak for the existence of a thick cloud man-
tle. These difficulties will be further discussed in the
next chapter.

(B) Perhaps the sun in very ancient times had not
yet altogether condensed to a pronounced nucleus,
but still existed as a nebula of very considerable size,
so that the earth was illuminated and warmed from
all directions, or (if no part of the nebula inclosed
the earth) at least from nearly a hemisphere.1 This
of course would promote uniformity of temperatures
from the equator to the poles. If thus receiving ra-
diation from a very large solid angle, the intensity of
the radiation need have been only very slight indeed
to maintain the earth's temperature. Such radiation
might be furnished by a cloud of small particles (not
gases) comprising the nebula. Even if they gave no
considerable radiation of their own, they would re-
flect that of the hotter solar nucleus. On either hy-
pothesis (A) or (B) the radiation of the ancient sun,
or solar nebula, to outside space may have been con-
siderably smaller than the total of the present solar
radiation.

If either of these views or a combination of both is
acceptable, it relieves the problem of solar radiation
of much of its difficulty. We may then suppose that
50,000,000 years ago the total emission of solar radi-

1 This suggestion was made by Chambcrlin about twelve years
ago. In Plate XXVI, Fig. 1, is shown a spiral nebula on edge. The
bright region at its center is seen to extend out of the plane of the
spiral so as to fill a large sphere.

276

WHAT IS THE SUN?

ation was considerably less than the present emis-
sion, and that it increased slowly for ages, reaching
approximately the present output at about the Pleis-
tocene period, which was perhaps not over 100,000
years ago.

Helmholtz in 1853 proposed a source of solar en-
ergy supply which is everywhere recognized as cer-
tainly very considerable. He pointed out that the
shrinking together of the sun converts potential en-
ergy of position into heat, just as the falling of a
stone converts its potential energy of position finally
into heat. Several authors have made computations
of the quantity of energy which would be available
from this source. Their results have generally been
based on the assumption that the sun was originally
a nebula filling a sphere whose diameter was the orbit
of Neptune. It appears that the condensation of
such a nebula having the mass of the sun would
have furnished thus far about 25,000,000 times as
much energy as the sun now loses each year. (This
estimate is based on a "Solar Constant" of 2.0
calories per square centimeter per minute.)

According to Helmholtz 's view, a contraction of
about 250 feet per year in the sun's diameter would
suffice to sustain the present solar radiation. At this
rate it would require about 10,000 years to reduce
the apparent diameter of the sun by one second of
arc, so that, so far as telescopic observation is con-
cerned, the contraction theory is tenable, for a change
of -T0- second in the solar diameter is unrecognizable.

277

THE SIX

From calculations of Newcomb the sun will require to
have shrunk to one half its present size if it maintains
its present rate of radiation for about 7,000,000 years
longer. As shrinking cannot go on indefinitely, nor
can the supply of heat from this cause have been in-
finitely great, we must, from this point of view, re-
gard the duration of life depending on the sun's rays
as having had a beginning in the remote past, and as
tending towards an end at some remote time in the
future.

There has been much question in recent years
whether Helmholtz's hypothesis of the sun's energy
supply is adequate to account for the duration of life
upon the earth revealed by the geological record.
Joly has estimated from the volume and salt contents
of the ocean, compared with the rates of discharge
and salinity of the rivers, that the earth's geological
age is about 80,000,000 years. G. F. Becker has re-
cently revised the calculations, with allowance for a
more rapid discharge of salt in earlier periods, and
finds about 50,000,000 years. On the other hand,
many geologists think the thickness of the earth's
deposited strata requires us to admit more than 100,-
000,000 years. The duration of the sun's radiation at
present rate of output apparently cannot have been
supplied by shrinking alone for more than 25,000,000
years. But, as has been said, it seems plausible that
the solar radiation was formerly less considerable than
now. If so, we may lengthen several fold the dura-
tion of the supply by contraction of sufficient solar

278

WHAT IS THE SUN?

radiation for purposes of supporting life on the earth,
leaving the question of the earth's temperature main-
tenance under the supposed circumstances to be dis-
cussed in the following chapter. On these grounds we
may regard Helmholtz's contraction hypothesis as
adequate to satisfy the requirements of geology and
physics in regard to the source of the sun's energy.
Whether or not radio-active processes are, or have
been, considerable sources of solar energy is not yet
determined.

CHAPTER VII

THE SUN AS THE EARTH'S SOURCE OF HEAT

Causes of Low Temperature at High Altitudes. — Measurement of
the Intensity of Sun Rays. — Dependence of Solar Radiation on
Air Mass. — The Transmission of the Atmosphere. — The "Solar
Constant of Radiation. "—The Light of the Sky.— The De-
pendence of the Earth's Temperature on Radiation. — Fluctuation
of Solar Emission. — Geological Temperatures.

NEARLY all of the heat of the earth's surface comes
directly from the sun's rays. The heat of coal and
wood and the energy of water power and wind, from
which heat may be derived, are indirectly the effect
of solar rays either of present or past times. Occasion-
ally a person is met with whose mind works so curi-
ously as to lead him to deny that the sun is hot. Such
an one almost invariably calls attention to the fact
that as we ascend a mountain, or are carried up by a
balloon, the temperature falls. Thus, although we
may be actually approaching the sun, the heating
effects of the solar rays become less obvious. Of
course the elevation possible for man to attain is in-
significant compared with the radius of the earth's
orbit, so that no change of solar radiation ought to be
appreciable from the change of distance to the sun
involved in climbing a mountain. But a consider-
able increase in the intensity of the sun's rays attends

280

THE SUN AS THE EARTH'S SOURCE OF HEAT

the mere ascending above the lower dusty part of the
atmosphere. Hence there is some excuse for surprise
at the decrease of temperature? observed at high alti-
tudes, which occurs notwithstanding the increase in
the direct solar radiation.

One secret of this paradox lies in the fact that the
sun's rays heat only objects which absorb them.
Highly transparent objects like glass, or the air, de-
rive little heat by being shined upon; for the rays pass
through them almost unchanged. Absorbing sub-
stances like lamp-black, on the other hand, almost
entirety destroy the rays and convert their energy of
vibration into heat. Upon the surface of the earth
the air is in contact with such an absorbing substance,
namely the ground, and is warmed by contact with it.
At high altitudes the free air has contact with no ab-
sorbing substance to warm it, and as it transmits sun
rays with great freedom it derives only a little heat
from them directly. It contains, moreover, ozone,
carbon dioxide, and water vapor which all radiate
freely long-wave rays and thus dissipate to space the
heat gained. Consequently the high air is cold, and
cools whatever it blows upon. Its cooling action on
the surfaces of mountains is greater on account of the
high winds which prevail.

Rising currents warm the upper air less than they
would do but for the decrease of atmospheric density
which occurs with increasing altitudes. For the air
currents which rise from the heated surface of the
earth expand in rising, and by expansion are some-

281

THE SUN

what cooled. A factor of considerable influence tend-
ing to cause lower temperatures on elevated inland
table lands, like the plateau of Thibet, is the compar-
ative lack of water vapor in the air above. The
water evaporated from the Indian Ocean can hardly
reach the plateau of Thibet because in rising through
the free air to such a great height it is so much cooled
as to be mostly precipitated. Water vapor, while
nearly transparent to light, and indeed to perhaps
eighty-five per cent of all .the rays which the sun
sends, is on the other hand a powerful absorber of the
rays of great wave length which are emitted by a
comparatively cool body like the earth. Hence at
low altitudes where water vapor is plentiful in the
air, it is a considerable hindrance to the escape of
earth rays to space. In the comparative lack of water
vapor at high altitudes of interior regions of large
continents, the cooling of the ground by radiation to
space is much more rapid than at sea level, and hence
lower temperatures prevail. In the case of steep and
rough mountains the configuration of the ground is con-
ducive to low temperatures because it diminishes the
radiation per unit area received from the sun, while
increasing the area affected by the cooling winds.

We may therefore attribute the coolness of the free
upper air to its transparency, its considerable radi-
ating capacity, and its expansion ; the coolness of the
rugged mountains to their contours, and to the con-
tact of the cool winds; the coolness of the elevated
inland plateaus to the dryness of the air above them ;

THE SUN AS THE EARTH'S SOURCE OF HEAT

at the same time recognizing that they all three are
receiving more intense rays from the sun than is the
earth's surface in general.

MEASUREMENT OF THE INTENSITY OF SUN-RAYS

That which the sun sends to the earth in such abun-
dance used to be considered as three distinct things,
namely: Actinic or chemical rays; light or visible
rays; heat or invisible rays. These distinctions are
now known to be misleading, for the rays which affect
modern photographic plates extend in the spectrum
from far beyond the farthest violet to far beyond the
farthest red, and the rays which can produce heat in-
clude all these, and many more, still further beyond
the red. All rays may be totally transformed to pro-
duce heat, however they may differ in their effects
upon the eye, or on different chemical substances.
All these rays travel with equal velocity in free space,
and this velocity is about 300,000 kilometers (186,-
000 miles) per second. That which so travels is not
a material substance, but waves, similar in some re-
spects to the waves which travel on water, or on a
stretched rope. That which distinguishes red light
from blue light is the length of the wave, or the num-
ber of complete waves executed per second. The
wave lengths of visible light vary from about 0.0004
millimeter in the violet to 0.0007 millimeter in the
red; and the corresponding numbers of vibrations
per second from 750 to 430 millions of millions. But
there have been recognized by means of photography

283

THE SUN

rays of wave length only 0.0001 millimeter and wave
frequency 3,000,000,000,000,000. By delicate heat
measuring apparatus rays of wave length 0.06 milli-
meter and frequency 5,000,000,000,000 have been
recognized. All this, and perhaps a wider range of
spectrum, is probably included in the sun beams as
they leave the sun, but our atmosphere prevents
some of the shortest and longest of them from reach-
ing the surface of the earth.

Since it is upon the supply of these sun rays that
heat, light, power, and the growth of all living things
upon the earth depends, the measurement of the in-
tensity of the total supply, and the determination of
the different varieties which compose it, are of first-
rate interest and importance.

We measure the intensity of solar radiation by the
heat which it will produce when completely absorbed
on a surface at right angles to the rays. A conven-
ient unit for measuring solar heating is the calory per
square centimeter per minute (see Chapter II). The
maximum intensity of solar radiation as measured
near sea level at Washington when the sun is not more
than 45° from the zenith usually ranges from 1.15 to
1.45 calories per square centimeter per minute on
cloudless days, depending on the clearness and dry-
ness of the air. At Mount Wilson in California, over
one mile above sea level, the values observed range
from 1.45 to U>2 calorics; and on Mount Whitney in
California, nearly throe miles in altitude, the ob-
served values reach 1.75 calories.

284

THE SUN AS THE EARTH'S SOURCE OF HEAT

Fig. 60 shows the march of intensity of sun rays
during the forenoon of July 6, 1910, on Mount Wilson.
The horizontal scale gives zenith distances, the verti^

1.4
1.2
1.0

" -

>•

CALORIES. P ?

)° 10" 20" 30° 40° 5O* 60° 70° 80* 90
ZENITH DISTANCES.

FIG. 60. — MABCH OF INSOLATION (Mount Wilson).

cal scale calories per square centimeter per minute.
The decrease of intensity at the smallest zenith dis-
tance observed is caused by increased humidity, due
to the springing up of a sea breeze about eleven

285

THE SUN

o'clock. The individual observations are given in
Table XVII, page 287.

Formerly radiation was regarded as three distinct
entities, namely : actinic or chemical rays ; visible or
light rays; obscure or heat rays. As already stated
this view is an error now generally abandoned, and all
radiation comprised in these three categories is rec-
ognized as of the same fundamental kind, differing
only as to wave length. The reader will therefore
recognize that Table XVIII is not intended to re-
vive this ancient classification, but only to fix our
ideas of the amount of solar radiation found in the
regions (1) where ordinary photographic plates are
most sensitive, (2) where the eye is the most sensitive,
and (3) in the infra-red spectrum. These facts are
given for the beam outside the earth's atmosphere,
and as it reaches Mount Wilson and Washington
under different angles of zenith distance. The num-
bers express the radiation, within the stated regions
of wave length, in calories per square centimeter per
minute. The zenith distances selected are 0°, 60°, 70°
32' and 75° 32', for which the "air masses"1 are 1,2,
3, and 4.

DEPENDENCE OF SOLAR RADIATION ON AIR-MASS

It is not possible to express satisfactorily the de-
crease of intensity of the direct solar beam, depending

1 The "air-mass" is the ratio of tho length of the path of the sun's
rays in the atmosphere to the corresponding length if the sun were
vertically overhead. It is closely expressed by the secant of the
zenith distance for zenith distances less than 75°.

286

THE SUN AS THE EARTH'S SOURCE OF HEAT

287

THE SUN

'O 00 CO O

<N CO iM 00

d d d d

o o o o

o o

00 £

T-l 00

<£> O

d i-L

o o o ^

<N O

T-l IO

d d

i-H IO CO ^

d d d ^H

o ^H

o p p

(N CO

o d

o o

000^

id 4.3 8 «

3332

Q »O O O

4.0 id id id

288

THE SUN AS THE EARTH'S SOURCE OF HEAT

on the decreasing elevation of the observing station.
For the lower layers of the air contain a load of dust
and water vapor which changes in quantity, quality,
and distribution from hour to hour and day to day.
In short, the variation of the solar beam in the lower
atmosphere does not proceed according to any fixed
law of relation to the barometric pressure.

As for the change of the intensity of the direct solar
beam for different zenith distances of the sun, that
may be well expressed by Bouguer's exponential for-
mula, e = e0a se ' z, as explained in Chapter II, pro-
vided we deal with homogeneous rays (rays which
are practically of one wave length), and observe them
at a single station on a clear day. If we imagine the
atmosphere to be made up of a great number of shells
concentric with the earth, and the shells of such thick-
ness as to contribute equal amounts to the barometric
pressure, each of the upper shells will transmit to the
shell next below practically an unchanging fraction of
the intensity the shell receives of a homogeneous ray.
But when the ray reaches a layer within one or two
miles of sea level the fraction transmitted continually
decreases from shell to shell owing to the increasing
load of dust carried by the lower layers.

The total thickness of the atmosphere necessary to
be considered as affecting solar radiation is less than
one hundred miles, and is so small compared with the
earth's radius that the shells may be regarded as prac-
tically parallel planes, except when we deal with rays
entering the atmosphere at very great zenith dis-

289

THE SUN

tances. Atmospheric refraction, too, may be neg-
lected in these computations for rays whose zenith
distance is not above 75°. Hence we may assume
that for zenith distances less than 75° the ratio of the
length of the path of the ray in each shell to the thick-
ness of the shell is constant, and equal to the secant
of the zenith distance. Under these restrictions (as
shown in Chapter II) the exponential formula of
Bouguer serves to determine the intensity, e, of mono-
chromatic rays at different zenith distances, even
though we do not know the change of transmission
from layer to layer. For as the sun rises higher and
higher the thickness in every layer changes in the
same proportion. In thought we may go even further,
and, with the sun in the zenith, imagine that the
thickness in every layer should be reduced simultane-
ously in equal proportions until no air remains. In
other words, we can, after the secant reaches its min-
imum value, unity, substitute another function of the
quantity of air in each shell, which we imagine to be
decreased in equal proportion in all layers until no
more atmosphere is left. Thus we may determine
the intensity, e0, which our monochromatic ray would
have outside the earth's atmosphere.

The quantity, a, which appears in the formula, is
the fraction of the intensity outside the earth's at-
mosphere which remains in the beam as it reaches the
observer at the earth's surface. This quantity is
called the atmospheric transmission coefficient. It
differs with the altitude of the observer and the clear-

290

THE SUN AS THE EARTH'S SOURCE OF HEAT

ness of his sky. It differs also for rays of different
colors; increasing, generally, as we pass from short
wave lengths to longer ones. There are, however, cer-
tain rays which suffer powerful selective absorption in
the gases and vapors of the earth's atmosphere, and
for such rays the transmission coefficients are very
small. Absorption bands play a very great part in
the red and infra-red spectrum, where the bands of
oxygen, water vapor, and carbon dioxide are principal-
ly found. This is made clear in the accompanying
illustration, Fig. 61, which shows two successive ob-
servations made on Mount Wilson by the bolometer
of the relative intensity of the rays in the solar spec-
trum of a 60° flint-glass prism. At places marked * the
sun rays were cut off so as to give the base line, or line
of zero radiation. At places marked | the sun rays
were altered in intensity so as to keep the curve within
the bounds of the plate. The heights above the base
line are proportional to the energy of the spectrum
rays. The length is proportional to the prismatic de-
viation. Fraunhofer lines show as depressions of the
curve. Prominent Fraunhofer lines are indicated by
their letters. These energy curves, or holographs,
were made on Mount Wilson as a part of a series of six
such curves obtained at different solar zenith dis-
tances in a single forenoon. They were made to de-
termine the transmission of the atmosphere at all
parts of the spectrum. From such observations the
distribution of solar radiation as it would be outside
of our atmosphere is computed. We have studied in

291

THE SUN

THE SUN AS THE EARTH'S SOURCE OF HEAT

Chapter III the significance of such work in regard
to the sun's temperature.

We could not determine the intensity outside the
atmosphere if the transparency of the air varied much
during the several hours required to complete the
series of holographs. Fortunately there is the follow-
ing criterion for the excellence of any given day in
this respect : In the course of the usual reductions,
logarithms of the heights above the base-line (corre-
sponding to intensities of radiation at given wave
lengths are plotted against zenith distances. The
results should show straight lines. Fig. 62, p. 294,
shows how well this test is met by the Mount Wilson
conditions. The tangent of the inclination of such
lines gives the logarithm of the transmission at ver-
tical sun, which we have called a. Values of a for
a given wave length are of course greater for Mount
Wilson than for Washington. By dividing the aver-
age Washington values by those for Mount Wilson
we obtain the average transmission of the mile of air
nearest sea level, as it is above Washington. We
shall see in the following tables that the loss in passing
through this last mile of the air is almost the same as
the entire loss above Mount Wilson.

Bouguer's formula is not exactly applicable to pyr-
heliometric measurements of the total radiation (or
summation of all rays of all wave lengths) of the sun.
It fails because rays suffer unequal extinction in the
atmosphere, some being almost completely extin-
guished in the upper air owing to the action of water

293

THE SUN

vapor and other selective absorbents. Hence for
these rays the intensity at the earth's surface does

^2.538

40
49
25
17
762
SO

2
6

\
\

$s*

2
6

£779

32.928

NNX

o LOGARITHM WTEc

:3.\28

S3.I20

^"^

— o —
^0^-

•&

"---<c

"~^

1 — —

N^.'£

'V.TC,

>-OSECANTZI2 24 3.6
FIG. 62. — ATMOSPHERIC TRANSMISSION PLATS.

not alter much with the zenith distance. Neverthe-
less the exponential formula holds approximately

294

THE SUN AS THE EARTH'S SOURCE OF HEAT

even for total radiation, except that the logarithmic
plats like those given in Fig. 62 are slightly curved,
and if continued on in straight lines to zero atmos-
pheric thickness they fall below the real intensity of
total solar radiation outside the earth's atmosphere,
as obtained from spectrum observations. Pouillet,
however, determined transmission coefficients for the
total solar radiation, and was thereby led to his cele-
brated value 1.76 calories per square centimeter per
minute for the solar constant of radiation. Radau,
and later Langley, showed clearly that, on account of
the differences of transmission for rays of different
wave lengths, we must observe the transmission of
each color by itself, and determine what the intensity
of each separate color would be outside the atmos-
phere. Langley first applied this procedure experi-
mentally. Following his method we may sum up the
area included under the solar spectrum energy curve
outside the earth's atmosphere, and compare it with
the area for the corresponding curve at zero zenith
distance of the sun. Thus we may find the actual
vertical transmission of the atmosphere for the total
radiation. Knowing by measurements of the pyr-
heliometer the intensity of total radiation for any ob-
served zenith distance we can determine how many
heat units the area of the corresponding spectrum-
energy curve represents. Summing up in similar
terms the area as it would be outside the earth's at-
mosphere we may obtain the true " solar constant."

21 295

THE SUN

TRANSMISSION OF THE ATMOSPHERE

In the following table there are given the mean
results for the vertical transmission of total solar
radiation, according to observations of the Astro-
physical Observatory of the Smithsonian Institution.

TABLE XIX. — Transmission for total solar radiation

PLACE

Washington

Mount
Wilson

Mount
Whitney

True transmission

0.699

0.817

0.896

Apparent transmission

0.787

0.894

0.960

Table XX, opposite, gives the atmospheric trans-
mission for vertical rays, and for the zenith distances
whose secants are two and three, respectively, for rays
of various wave lengths.

THE SOLAR CONSTANT OF RADIATION

From the mean results of the Washington observa-
tions of 1902 to 1907, the Mount Wilson observations
of 1905 to 1910, and Mount Whitney observations of
1909, 1910, all corrected to the absolute scale of heat,1
the total intensity of solar radiation outside the
earth's atmosphere at the earth's mean distance from
the sun (called the " solar constant" of radiation),

1 Values published in Volume II of the Annals of Astrophysical
Observatory of the Smithsonian Institution were given on a pro-
visional scale of pyrheliometry differing about five per cent, from the
true one, and are here given as reduced to true calories.

296

THE SUN AS THE EARTH'S SOURCE OF HEAT

dddd

tOl^ O CO
l^co co **

OiOiOiOi

dddd

§o^§§
dddd

OiOi CO CO

dddd

TjH <M GO O

tO to CO Oi
Oi Oi GO GO

dddd

CO CO Oi F^-

<M Oi >O •*
Oi GO 1> GO

dddd

GO i

OOOO

=oc5«^«^
ocOqq

dodo

§co

S a

0000

O^ O^

oooo

:28

' CO CO

O CO <N

(N CO CO t^

»O <M CO O
T-H CO Oi iO

GOr^ ** co

CO *O C^l iO

fl «s §

Illl

CO Oil> GO

dddd

^ CO ^ Oi

•^ to CO Oi
Oi Oi 1> 1>

dddd

CO "* (M tO

dddd

t^ CO CO ^O

dddd

CO<M i-i t^"

CD -M

2 .

.agd.t5

"SF1 §3
S^'-S S

§ 'C t*

297

THE SUN

as expressed in calories per square centimeter per
minute is as follows :

PLACE

Washington

Mount Wilson

Mount
Whitney

Date

1902-1907

1905

1906

1908

1909

1910

1909

1910

Observations.

113

Mean solar
constant.. .

1.960

1.925

1.921

1.929

1.896

1.914

1.959

1.956

In 1909 and 1910 observations were made simul-
taneously on Mount Wilson (elevation one mile) and
Mount Whitney (elevation nearly three miles) by
Smithsonian observers, with the following results:

DATE

1909 Sept. 3

1910 Aug. 12

Aug. 13

Aug. 14

Mount Wilson

1.943

1.924

1.904

Mount Whitney

1.959

1.979

1.933

1.956

We see that notwithstanding the differences in al-
titude of the observing stations, and the differences
of atmospheric transmission above them, there is
good agreement between the computed values of the
" solar constant" of radiation. Prior to 1905 this
quantity was in great doubt, as numbers ranging
from 1.76 to 4.10 had been given for it, and the ac-
cepted value then was 3.0. The opinion now seems to
prevail that no considerable change from the Smith-
sonian result of about 1.95 calories per square centi-

298

THE SUN AS THE EARTH'S SOURCE OF HEAT

meter per minute will come from future experiment-
ing.

Expressed in another way, the measurements indi-
cate that if the sun's rays could be completely em-
ployed to melt ice exposed continuously to them at
right angles, they would suffice to melt a layer 426
feet thick in a year.1 Such a layer at the earth's mean
distance, if it entirely surrounded the sun, would
weigh 4 X 1025 (4 followed by 25 ciphers) tons, and
the complete melting of it each year would represent
as many heat units as the burning of 4 X 1023 tons
of anthracite coal. This, then, is a measure of the
sun's yearly output of radiation.

THE LIGHT OF THE SKY

It must not be inferred from the tables given on
a preceding page that only 81.7 per cent of the sun's
radiation reaches the Mount Wilson level at vertical
sun. That, to be sure, is the average result for the
direct solar beam, but the sky supplies an appreciable
addition of indirect rays even on Mount Wilson. At
sea level the sky light is a still more considerable por-
tion of the total radiation, but as yet not very ex-
actly measured. The relative brightness of the sun
and sky differs greatly according to the manner in
which the rays are received. Owing to the great ex-
tent of the sky, it is not possible, when receiving rays

1 As the earth has four times the area of its cross-section, we may
say that the sun's rays are capable of melting an ice shell covering the
earth to an average thickness of 106 . 5 feet annually.

299

THE SUN

simultaneously from its whole extent, to have them
all fall at right angles to the absorbing surface.
Hence the sky light is at a disadvantage with respect
to sunlight, unless we observe the brightness from
every part of the sky by itself and then sum up the
results. From bolometric measurements of 1905
and 1906, made by the Smithsonian observers, and
reduced in this manner, it appears that the total sky
radiation on Mount Wilson computed at normal in-
cidence, and including all wave lengths, is from eleven
to twenty per cent of the total direct sun radiation.
Both sun and sky rays are in this estimate supposed
to be received at right angles to the absorbing sur-
face, and the sun to be not over 50° from the zenith.
The percentages depend on the clearness of the sky,
increasing with the haziness. If we make the as-
sumption that the sky shines on a horizontal surface,
and the sun upon n surface normal to the beam, these
percentages become 5.2 and 7.7. If both sun and sky
rays are supposed to shine on a horizontal surface,
the ratio varies of .course greatly from hour to hour.
Professor Exner has derived formulae for the rela-
tive brightness of the sun and sky on the hypothesis
that the sky light is all due to scattering from parti-
cles which are small as compared with the wave
length of the rays.1 He has found it necessary to make
some rather rough simplifying assumptions. Never-
theless his computations fall in pretty well with such

1 Sitzungsbericht, d. K. Aknd.d. Wissen., Wien., M. N. Klassr.
CXVIII, Ha, 1909.

300

THE SUN AS THE EARTH'S SOURCE OF HEAT

observations as are available. In the following
tables taken from Exner's publication, z is the zenith
distance of the sun, H the intensity of sky light and S
that of sunlight, both being measured on a horizontal
surface. At normal incidence outside the atmos-
phere the intensity of sunlight is taken as unity. The
quantity p is the transmission coefficient of the
atmosphere above the observer for a vertical ray.
From Smithsonian observations we see that p = 0.6
would correspond to wave length 0.43//, (violet) at
Washington 0.35/z. (ultra-violet) at Mount Wilson.
Correspondingly, for p = 0.75 we have 0.59/*, (yellow)
at Washington and 0.41//, (violet) at Mount Wilson.

TABLE XXI. — Sunlight and sky light. (Exner.)

P = 0.6

p = 0.75

S + H

S
H

S + H

S
H

80°

0.241

0.009

0.250

0.04

0.136

0.032

0.168

0.24

70°

0.245

0.077

0.322

0.31

0.138

0.147

0.285

1.06

60°

0.252

0.180

0.432

0.72

0.141

0.282

0.423

2.00

50°

0.259

0.289

0.548

1.12

0.146

0.408

0.554

2.79

40°

0.268

0.394

0.662

1.47

0.151

0.528

0.679

3.50

30°

0.276

0.484

0.760

1.75

0.155

0.625

0.780

4.03

20°

0.281

0.547

0.828

1.95

0.158

0.693

0.851

4.38

10°

0.285

0.582

0.867

2.04

0.160

0.731

0.891

4.57

0°

0.288

0.600

0.888

2.08

0.162

0.750

0.912

4.63

The change of H with the zenith distance of the
sun is not as great in these tables as it should be.
This appears from the following measurements of
Roscoe for which we may assume p =0.6.

The units employed by Roscoe are not the same as
301

THE SUN
TABLE XXII.— Sunlight and sky light. (Roscoe.)

80° 9'

70° 19'

58° 46'

47° 47'

36° 51'

25° 46'

0.038

0.062

0.100

0.115

0.126

0.138

0.000

0.023

0.052

0.100

0.136

0.221

those employed by Exner, so that for easier compari-
son the results of Roscoe may be multiplied by 2 or
2.5. As there is much difference for different days
and for different stations in results of this kind, Ex-
ner's computations seem to be near enough at least
for giving a general idea of the state of affairs. Indeed
the following summary, which I translate from Wies-
ner's description of his photographic observations of
light received on horizontal Surfaces,1 fits Exner 's
results for short wave lengths very well :

"The direct sunlight, which is sometimes twice the
intensity of diffused light, may also sink to zero. —
For solar altitudes less than 19° (z = 71°) the chem-
ical intensity of the sunlight as compared with dif-
fused daylight is negligible. With increasing solar
altitude the intensity of the direct sunlight gains in
comparison with the diffused light. The solar alti-
tude for which S = H seems not to be constant even
for apparently clear sky, and for one and the same
station. For cloudless sun the equality of direct and
diffused light occurs generally when the solar alti-
tude is about 57° (z = 33°), yet with clear sky it was

'Vienna Academy. Denkschriften, Bd. 64, 1807.

303

THE SUN AS THE EARTH'S SOURCE OF HEAT

observed once at 33° (z = 57°). Since the intensity
of the direct beam may reach twice that of the dif-
fused, the total combined chemical effect may be
threefold that of the diffused light."

Exner also gives computations of the relative
amounts of combined direct and diffused light receiv-
able on vertical surfaces facing respectively South,
North, West, East, compared with the amounts re-
ceived on a horizontal surface. The southern expos-
ure only of the vertical surface is supposed to receive
some direct sunlight, designated by 2. V8, VN, Vw,
VE designate the diffused illumination of the vertical
surface toward the South, North, West, and East.
S and H have their former meanings.

TABLE XXIII. — Sky light on a vertical surface. (Exner, Schramm.)

Computed, p=0.8

Observed by W. Schramm

2 + Vs

Vw VE

2 + V8

S + H

S + H S + H

S + H

85°

1.43

0.537

0.461

2.73

0.560

0.542

0.604

2.21

0.263

0.241

3.41

0.268

0.397

0.386

1.69

0.148

0.146

1.81

0.258

0.331

0.351

1.25

0.099

0.104

1.32

0.147

0.223

0.204

0.92

0.075

0.083

0.976

0.118

0.195

0.175

0.68

0.062

0.071

0.749

0.091

0.131

0.139

It is not probable that the influence of the direct
sunlight was wholly absent in Schramm's VE and Vw
observations. Apart from these there is a pretty
good agreement of observed and computed results,
as the following summary also indicates,

303

THE SUN

Ratios of average vertical illumination to that of
the North.

2 + Vs

Vmean

Observed

7 64

1 00

1 26

1 29

2 80

Computed

6 91

1 00

0 94

2 45

Exner computes from p = 0.6 and z = 40°, the
following :

S + H

2 + Vs

Vmean

S + H

S + VB

v w— VE

Vmean

0.662

0.480

0.106

0.120

0.207

3.2

4.5

As we have stated at some length in Chapter VI,
Schuster has employed Lord Rayleigh's theory of
the scattering of light by particles small compared
with the wave lengths, to compute the transmission
of the direct beam of sunlight. He assumes that the
loss in the atmosphere is wholly from the scattering
caused by the molecules of the air. He finds close
agreement between the computed and observed re-
sults for excellently clear days at Mount Wilson and
Washington. This seems to indicate that the dust
load of the atmosphere plays a subordinate part in
affecting the solar radiation on the best days, and
that under such conditions as are found ordinarily
at Mount Wilson, and occasionally at Washington,
nearly all the light of the sky is due to the diffuse

THE SUN AS THE EARTH'S SOURCE OF HEAT

reflection or scattering of sun rays by the molecules
of air. Somewhat similar conclusions have been
reached still more recently by Natanson, except that
he considers the matter from the point of view of the
electron theory.

The light of the sky is very much richer in the vio-
let rays than that of the direct solar beam. From
Smithsonian experiments the following values are
taken for different wave lengths, assuming about
equal intensities in the extreme red for a sunbeam
and a skybeam, and giving the approximate normal
spectral distribution for both as observed at the sur-
face of Mount Wilson.

TABLE XXIV. — Sun and sky light. Relative brightness for different
wave lengths on Mount Wilson.

WAVE LENGTH -*•

0.422/w,

0.457M

0.491ft

0 . 556/u.

0.614/*

0 . 660M

Sunlight; z = 50°

186

232

227

211

191

166

Skylight
Ratio

1,194
6 92

986
4 25

701
3 09

395

1 87

231
1 21

174
1 05

At the sea-level, especially in cities and other dusty
localities, the proportion of blue in sky light is usu-
ally much less than that given above; for particles
large as compared with the wave length of light, such
as occur in dust, do not act in the same way as small
particles and molecules. Large particles, by reflect-
ing sunlight, tend rather to diminish than to increase
the relative proportion of the intensity due to rays of
short wave length.

Skylight is brightest near the sun and near the hori-
305

THE SUN

i— i
d

B i

.!!

°o

00
CO

COCO

iCt-

i— i

r*» 01

O l^-

d t^

i-i <N'

cod

I> l-H

•^

00
1C

coc^

01 Cl
COO

CO ui

dei

|2
~

CO i-H

1-1

R
Q

T-l

t£

•^

r^[^

O5 O

rH (M

00 i-H

>-<

•1

2§

CO®

<N d

>C i— i

o>~

co'^

28§

5 2

1
£

i-H

: g

"S J

[Vj [^

^ — s

^ G

? 5

I-H !

G
$

-D :

1
§

-+J
o

i— i

>— i

•

1— 1

300

THE SUN AS THE EARTH'S SOURCE OF HEAT

zon. The results shown in Table XXV on the total
radiation are from bolometric measurements at
Mount Wilson, and at Flint Island, a coral island
in the Pacific, near the equator, lying 400 miles
northwest of Tahiti.

The measurements are reduced as if the sun were
in the zenith.

Summing up columns* (V) and (VI) we find the
average brightness of a portion of the sky equal in
angular area to the sun as compared with the bright-
ness of the sun: First as received on a surface at
right angles to the beam in both cases; second with
the skylight received on a horizontal surface and the
sunlight received normally. The measurements in-
clude all the rays transmissible by a glass plate three
millimeters thick. The results are as follows:

TABLE XXVI. — Ratio of total radiations: Sky to sun

STATION

FOR EQUAL AREAS

FOR WHOLE SKY

Normal
incidence

Sky on
horizontal

Normal
incidence

Sky on
horizontal

Flint Island

636X10-8

302X10-8

0.67

0.32

Mount Wilson

176X10-8

69X10-8

0.18

0.072

Thus according to these measurements (which how-
ever are not sufficiently numerous or exact) at sea
level the sky furnishes to a horizontal surface thirty-
two per cent as much radiation as the direct high sun.
At 1,800 meters elevation only 7.2 per cent.

307

THE SUN

THE DEPENDENCE OF THE EARTH'S TEMPERATURE
ON RADIATION

The temperature of the earth seems to be main-
tained at present almost wholly by the absorption of
solar radiation. It is thought by some that the
earth's temperature is rising slowly. If so, this
would indicate that the sum total of the earth's sup-
plies of heat exceeds its losses. But this change of
temperature, if real, is so exceedingly slow that we
may practically say that the earth's heat income and
heat outgo balance. The outgo, neglecting the rela-
tively trifling effects of vegetable and other storage
processes, is made up wholly of the earth's radiation
of long-wave rays to space. It has been shown that
the absorptive effects of atmospheric water vapor,
carbonic acid and ozone combined prevent nearly
or quite nine-tenths of the rays which are emitted
at the earth's surface from escaping directly to
space.

Hence the earth's effective radiating layer may be
regarded as situated in the atmosphere, and as being
chiefly the water vapor layers at several miles ele-
vation, whose average temperature is about — 10° C.
The still higher situated effective radiating layers of
carbonic acid and ozone gases, whose average temper-
atures reach as low as — 60° C, also radiate freely in a
few limited regions of spectrum. We shall not be
far astray, therefore, if we regard the average tem-
perature of the earth's radiating layer as not above

308

THE SUN AS THE EARTH'S SOURCE OF HEAT

260° of the absolute centigrade scale.1 The constitu-
ents of this radiating layer, namely, water vapor, car-
bonic acid and ozone, owing to their powerful selec-
tively absorbing properties for rays proper to this
temperature, must also be nearly perfect radiators
for the same temperature. Admitting that their
radiating power is perfect, the emission of the as-
sumed radiating layer at 260° C. absolute is, by
Stefan's law (see Chapter II), about 0.34 calories per
square centimeter per minute.

In order to sustain this average rate of loss of heat
over the whole surface of the earth, solar radiation,
shining effectively over only the area of the earth's
cross-section, must be absorbed at four times this
rate, or 1.36 calories per square centimeter per min-
ute. Of the energy represented by the solar con-
stant (1.95 calories) about thirty-five per cent is re-
flected away according to the Smithsonian determi-
nation of the earth's " albedo. " The remainder is 1.27
calories, and nearly suffices to furnish the heat above
computed as lost from the earth. The difference
(0.09 calories) may possibly mean that an appreciable
quantity of heat is furnished by terrestrial sources,
such as radio-active processes. However it seems
quite reasonable to suppose that the difference may
be accounted for (1) by assuming that the earth's
effective radiating layer is not a perfect radiator, so
that its radiation falls short of the 0.34 calories per
square centimeter per minute which a perfect radi-

1 Water freezes at 273° of this scale, and boils at 373°.
309

THE SUN

ator at 260° C. absolute would emit; or (2) that the
effective radiating temperature is below 260° C.
absolute.

The surface temperature of the earth reaches 310°
absolute C. in the tropics, and at the poles falls as
low as 220°, and its effective mean temperature is
287.2° absolute C. or + 14.2 C. It exceeds the tem-
perature of the radiating layer by over 25°. This is
in large measure for the same reason that the gar-
dener's hot beds, or the steamfitter's asbestos-covered
pipes, exceed the temperature of their surroundings.
For the sun's rays shine through the atmospheric
vapors readily, and warm the earth's surface. The
escape of its heat, as we have seen, is hindered by the
atmosphere. Hence the earth's surface temperature
rises sufficiently to force a flow of heat out to the
effective radiating layer. If it was not for the blank-
eting effect of the water vapor of the atmosphere, the
earth's mean surface temperature would probably be
nearly 20° C. below freezing, providing the reflecting
power of the earth was not changed. But if there
was no water vapor in our air, the sun's rays would
reach the earth's surface with at least ten per cent
greater intensity on cloudless days than they do now.
Since clouds would then be absent there would be
about 1.75 calories instead of 1.27 as now available
to warm the earth. Consequently, the earth's mean
temperature, if water was absent, would be about
277° absolute or + 4° C. But there would then be
a much greater range of temperature between night

310

THE SUN AS THE EARTH'S SOURCE OF HEAT

and day and between summer and winter than there
is now.

Upon the moon there is no atmosphere, and by the
observations of Lor*d Rosse, of Langley and of Very,
the moon's sunlit surface falls from about the tem-
perature of boiling water nearly to that of liquid air
within the short duration of a total lunar eclipse.
Quite otherwise is the state of affairs on the earth.
In the following table is given the yearly average of
the daily range of temperatures which occurs at sev-
eral stations on the earth.

TABLE XXVII. — Yearly means and mean daily temperature depar-
tures. (Centigrade.)

HOUR -»

Lati-
tude

Mid-
night

Tirnbuctu
Port au Prince . . .

16°49'N
18°34'N

-4M

-2°. 6

-5°. 6
-3°. 2

-6°. 8
-3°. 7

-7°. 7
-3°. 8

-2°. 8
-0°.6

+3°. 2
+2°. 9

+6°. 9
+4°. 7

HOUR -*

Noon

Mean

Timbuctu
Port au Prince . . .

+6°. 9
+4°. 7

+8°. 5

+4°. 5

+7°. 4
+3°.l

+3°. 4
+l°.l

-0°.l
-0°.8

-2°. 4
-1°.3

-4°.l
-2°. 6

29°. 2
25°. 9

Even a polar night of five months' duration in
which the sun is continuously below the horizon pro-
duces no such range of temperature on the earth as a
total lunar eclipse of a few hours' duration does upon
the moon. Witness the following mean temperatures :

Fort Conger. Latitude 81° 44'. Temperatures Centigrade.

Jan.

Feb.

Mar.

April

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

-39°.0

-40°. 1

-33°.5

-25°.3

-10°.0

+0°.4

+2°.8

+1°.0

-9°.0

-22°.7

-30°.9

-33°.4

311

THE SUN

These examples indicate how slowly the temperature
of the earth falls towards the absolute zero when the
solar radiation is utterly cut off. The delay cannot
be attributed to the influence of the inner heat of the
earth. From the rise of temperature at increasing
depths in the earth (about 1° C. in 28 meters) taken
in connection with the observed conductivity of rock
(about 0.0042 calories per second per centimeter cube)
it is calculated that the heat supplied to the surface
from within is but 0.00010 calories per square centi-
meter per minute, which would suffice to keep a per-
fect radiator at only 34° absolute (Centigrade) tem-
perature and could not be expected to keep the earth's
surface above 40° absolute, or -233° C. Even this is
far above what the moon and all the stars combined
could do to supply the place of the sun.

The following table1 of the mean monthly temper-
atures (Centigrade) gives some idea of the yearly
ranges of temperatures on the earth at various sta-
tions in the Northern Hemisphere. Several pairs of sta-
tions at nearly the same latitude, but one inland, the
other oceanic, are contrasted to show the influence of
the oceans in reducing fluctuations of temperatures.

The reader will notice how much smaller are the
yearly ranges of temperatures for oceanic stations
than those for the inland stations. It is also appar-
ent that the yearly range increases with the latitude.
This is due in part to the growing disparity of the

1 The data are taken mainly from various publications of J. v.
Hann.

312

THE SUN AS THE EARTH'S SOURCE OF HEAT

STATION

Ver-
khoy-
ansk

Fort
Conger

St.
Louis

(U.S.A.)

P.
Delgada

(Azores)

Tim-

buctu

Port
au
Prince

Bogota

Jaluit
(Mar-
shall
Islands)

Latitude N

67°34'°

81°44'°

38°38'

37°45'

16°49'°

18°34'°

4°31'

5°55'

Elevation
(meters)

173

250

2660

January. . .
February. .
March. .
April . .
May. . . .
June. . .
July. . . .
August.
September.
October . . .
November.
December .

-51°.0
-45°.3
-32°.5
-13°.7
+ 2°.0
+ 12°.3
+15°.o
+10°.l
+ 2°.5
-15°.0
-37°.8
-47°.0

-39°.0
-40°. 1
-33°.5
-25°. 3
-10°.0
+ 0°.4
+ 2°.8
+ 1°.0
- 9°.0
-22°.7
-30°.9
-33°.4

- 0°.8
+ 1°.7
+ 6°.2
+13°.4
+18°.8
+24°.0
+26°.0
+24°.9
+20°.8
+14°.2
+ 6°.4
+ 2°.0

+14°.l
+13°.9
+14°.l
+15°.4
+1G°.G
+18°.9
+21°.3
+22°.0
+20°.9
+ 18°.9
+16°.9
+ 15°.l

+21°.8
+23°.8
+28°. 1
+32°.5
+35°.0
+34°.2
+32°.7
+31°.l
+31°.8
+31°.0
+26°.8
+21°.4

+24°. 1
+24°.G
+25°. 1
+25°.9
+26°.0
+27°.l
+27°.G
+27°. 3
+26°.7
+26°.3
+25°.6
+24°.4

14°.2
14°.4
14°.S
14°.7
14°.8
14°.5
14°.l
13°.9
13°.9
14°.4
14°.7
14°.5

27°. 1

27°.2
27°.0
26°.9
26°.9
26 .8
26°.8
26°. 9
26°.9
27°. 1
27°. 1
27°.0

Yearly
Range

66°.5

42°.9

26°.8

8°.l

13°.6

3°.o

0°.9

0°.4

Yearly
Mear^

-16°.7

-20°.0

+ 13°. 1

+ 17°.3

+29°.2

+25°.9

14°.4

27°.0

longest and shortest days at higher latitudes, and in
part to the more rapid change in the intensity of
illumination with change of zenith distance of the
sun at high latitudes. At the equator the days and
nights are always equal, and the secant of the zenith
distance of the noonday sun varies only from 1 to
0.917. At latitude 45° N. the length of day varies
from eight hours, thirty-four minutes to fifteen hours,
twenty-six minutes, and the secant of the noonday
zenith distance from 0.930 to 0.366. The value of
the secant of the zenith distance influences the result
in two ways, first, as it measures the length of the
path of the rays in the air, second, as it measures
their weakening on a horizontal surface in conse-
quence of obliquity.

313

THE SUN

A minor influence which affects the yearly march
of the solar radiation is the change of the sun's dis-
tance from the earth. This causes an increase of
nearly seven per cent in the earth's heat supply from
July to January; and combined with the sun's march
in declination it produces two maxima and two
minima of radiation in the tropics. At Bogota
temperature maxima occur March to May and Oc-
tober to December — minima August to September,
and January, all occurring a little after the corres-
ponding maxima and minima of radiation.

By taking the three factors of solar distance,
obliquity and the daily duration of sunlight into
account, formulae have been devised for comput-
ing the " effective insolation" as it is sometimes
called. This is the intensity of the uniform beam,
which if received continuously at normal incidence
would yield an equal supply of radiation to that
which is really effective on a horizontal surface.
In such computations atmospheric losses are usu-
ally neglected, but on the other hand diffused sky
radiation is also neglected. We may also imagine
an hypothetical earth equal in size and similar in
motions to the real earth, but a perfect absorber and
radiator; thin as an egg shell; perfectly conducting
of heat from east to west, but perfectly non-conduct-
ing from north to south. The temperature of such a
structure can be computed for all times and latitudes
by Stefan's law (see Chapter II). When such com-
puted temperatures are compared with those actu-

314

THE SUN AS THE EARTH'S SOURCE OF HEAT

ally observed on the earth, it is found that no real
stations show as great yearly ranges of temperature
as the corresponding hypothetical ones. For Tim-
buctu and some other desert stations the observed
range is over half the computed. For average inland
stations the ratio is about three-tenths; for average
coast stations, one-fifth; for average island stations,
one-twelfth; at Apia, in Samoa, only one twenty-
fifth. For the hypothetical earth the percentage
change of absolute temperature is everywhere one-
fourth of the percentage change in solar radiation
which would cause it.

The accompanying illustration, Fig. 63, gives the
march of the " effective insolation" at the north lat-
itudes 17° 40' and 5° 10', and also the yearly change
of temperature at Timbuctu (16°. 49' N.), Port au
Prince (18° 34' N.), Bogota (4° 31') and Jaluit (5° 55').
The curves show how much the effect of solar change
may be modified by local conditions, and especially
how considerable are the delays which occur at
oceanic stations between solar causes and their ter-
restrial temperature effects. Thus, while at Tim-
buctu, an inland station, the maximum and minimum
temperatures attend closely the minima of effective
insolation, they are so far delayed that the maximum
temperature occurs three months after the insolation
is at its maximum at St. Louis, Senegambia, a coast
station nearly west of Timbuctu. Such facts should
be taken into consideration when seeking by studying
temperature statistics to determine if fluctuations of

315

THE SUN

32'

JAN. APR. JULY OCT.

FK;. <;:<. INSOLATION AND TERRESTRIAL TEMPERATURES.

316

THE SUN AS THE EARTH'S SOURCE OF HEAT

solar radiation probably occur. Long solar periods
like the eleven-year sun-spot period may be expected
to affect the temperatures of all terrestrial stations
nearly simultaneously. Not so temporary solar
changes of a few days or months. These could be
expected to show their influence only at inland sta-
tions, and preferably cloudless ones. The lag of tem-
perature minima behind the inducing solar radiation
minima is about twenty days for average continental
stations, only ten days for particularly favorable
ones, but reaches two months or more in the case of
many island ones.

FLUCTUATION OF SOLAR EMISSION

, Numerous attempts have been made to see if ter-
restrial temperatures, by their departures from the
normals, indicate fluctuations of the sun's emission
of radiation. Koppen concluded from such investi-
gations, published in 1873, 1880, and 1881, that the
earth's temperature is higher at sun-spot minimum
than at sun-spot maximum. This conclusion is con-
firmed by Stone, Gould, Nordmann, Newcomb,
Abbot and Fowle, Arctowski, Bigelow, and others.
Taking a general view of their results with those of
Koppen, we may conclude that for a change of 100
sun-spot numbers of Wolf's scale, which is about the
average range of sun-spot activity, there is a change
of the mean temperature of the earth of about 0.7° C .
The cause of this cannot be in the mere darkening of
the disk of the sun by the areas covered by the spots,

317

THE SUN

for if they were perfectly black all over the change of
solar radiation corresponding to 100 sun-spot num-
bers would be only — — , which is not J of what is nec-
500

essary to produce the observed change of tempera-
ture. There must therefore be other changes of the
sun (or possibly in the earth's atmosphere or in the
intervening space), as yet not understood, which at-
tend the increase of sun spots and which exceed in
their effective reduction of solar radiation the direct
influence of the darkness of the spots themselves.

It is only within the last five years that there have
been direct measurements of the solar radiation suf-
ficiently complete and accurate to show whether
there are frequent changes of the sun's emission suf-
ficiently large to affect the earth's temperature notice-
ably. The reader might be inclined to suppose that
a mere analysis of the deviations from normal tem-
peratures at numerous stations would be sufficient to
practically verify or disprove frequent variability
of the sun. Indeed there are found numerous in-
stances of departures from normal temperatures
which seem to indicate something of the kind, but yet
the evidence of different localities is so contradictory
and confusing that careful meteorologists reserve an
opinion on the matter. Recently the idea has gained
some adherents that a few per cent of increase of
solar radiation during a period of several months need
not necessarily affect the temperatures of all stations
on the earth in the same direction, but might make

318

THE SUN AS THE EARTH'S SOURCE OF HEAT

some warmer and others cooler temporarily, so that
the effect may thus be obscured. This contradiction
of effects, as these observers think, may be due to
changes of cloudiness and of the circulation of the
atmosphere. Statistical studies of temperatures by
Arctowski, Bigelow, and others support the view that
certain large regions are on the whole balanced in
their temperature changes, so that the same dis-
turbance which warms one region, cools another, as
if there were waves of effect superposed upon the
earth. Apparently the meteorological condition of
the earth is so complicated by the relative configura-
tion of land and sea, cloudy and clear areas and hot
and cold regions that we cannot expect to determine
solar changes with certainty by climatic investiga-
tions, and must rather, turning the matter about,
first determine the solar changes by direct observa-
tions, and then search out their terrestrial effects.

Up to 1905 the measurements of solar radiation
made in different countries by different investigators
were so much at variance as to make it seem highly
unlikely that sufficiently accurate knowledge of the
solar emission could be obtained to lead to the dis-
covery of variability of the sun. But the Smithson-
ian observations at Washington, Mount Wilson, and
Mount Whitney as given on a previous page are so
highly concordant, and seem to be so probably com-
petent to fix the intensity of the solar radiation out-
side the atmosphere within about one per cent, that
there now seems to be a good prospect of discovering

319

THE SUN

fluctuations of solar radiation if they exceed one per
cent from the mean value. Heretofore the observa-
tions have not been kept up continuously, but ef-
forts are being made to remedy this defect by the
establishment of one or more additional " solar con-
stant" observing stations.

From measurements of the Smithsonian observers
on Mount Wilson during six months of each of the
years of 1905, 1906, 1908, arid 1909, as given in Fig.
64, it appears probable that fluctuations of the sun at
irregular intervals of several days, and sometimes of
several months, are not uncommon. Apparently the
amplitude of such changes sometimes reaches ten
per cent, and seems frequently to reach from three to
five per cent. But notwithstanding that the reality
of' these changes is attested by various evidences,
such as the continuity of a change during several
days of consecutive observing, which the reader can
see for himself in the observations of 1908 and 1909,
Fig. 64, and the fact that the fluctuation of radiation
depending on the yearly change of the solar distance
can be easily recognized,1 though it amounts to but
three per cent during the six months covered by the
Mount Wilson work, yet the supposed solar variabil-
ity can hardly be said to be conclusively shown until
another station as well equipped as Mount Wilson
supports the conclusion by simultaneous measure-
ments.

If occasional variations of ten, or even five, per

1 The values platted in Fig. 64 are reduced to mean solar distance.
320

THE SUN AS THE EARTH'S SOURCE OF HEAT

321

THE SUN

cent, shall be found to exist in the solar radiation,
they may well be expected to produce noticeable
effects on the earth's climate. We have seen that it
is difficult, if not impossible, to seek the cause from
the observed climatic effects, owing to their great
complexity. But when the variations of the sun
become accurately observed, and thus the action of
the cause is known, the tracing of the climatic effects
will be a matter of great interest and importance.
Indeed, some comparisons of temperature statistics
with the " solar constant" values have been made al-
ready, and indicate a probable connection between
the two. But much more work is needed for certainty.
A solar change of five per cent continued for six
months might well alter the mean temperature of
inland stations by 2° C., or 3.6° F., and this would
make the difference between an unusually hot and
an unusually cold season. Its influence in temperate
zones on the length of season favorable to vegetable
growth would be very noticeable, as will be more
clearly shown in the following chapter.

GEOLOGICAL TEMPERATURES

It is generally believed that from the Cambrian to
the Pleiocene a genial climate usually prevailed at
the poles, and, moreover, without evidences of ex-
traordinarily high temperatures at the equator. This
state of affairs seems to be inconsistent with the view
that the sun controlled the temperatures then in the
same manner that it does at present.

322

THE SUN AS THE EARTH'S SOURCE OF HEAT

On the other hand, the sub-tropical glaciation L
which interrupted this condition of uniform temper-
atures during the Permian- period seems inconsistent
with the present value of the solar constant of radia-
tion. If the constant is two calories per square cen-
timeter per minute, the average 2 insolation per
square centimeter per minute in latitude 28° is 0.55
calories. The largest possible part of this supply
would be lost for purposes of heating if the earth was
completely cloudy. According to observations of
Smithsonian observers this maximum possible loss,
at this tropical latitude of high sun, would be less than
forty-two per cent, leaving at least 0.32 calories ab-
sorbed. This remainder would maintain a perfect
radiator at 254.3° absolute Centigrade, or —18.7 C.
If the radiating layer were not perfectly " black" it
would have a higher temperature than this, and any
contribution of heat from the earth's interior would
also tend to raise its temperature. Under the per-
fectly cloudy condition we are considering, the upper
part of the cloud layer would absorb most of the ab-
sorbable radiation from the sun, but its own outward
radiation would be restrained, then as now, by the
water vapor, carbon dioxide and ozone lying higher.
Accordingly there would be a "blanket" or " hot-
house" effect similar to that which now "exists, and
which now raises the surface temperature of the
earth nearly 30° C. above the temperature of the

1 See description quoted near the end of Chapter VI.

2 For night and day for the whole year.

323

THE SUN

radiating layer. Hence we may suppose that the
upper layer of the clouds would also have been nearly
30° C. above the temperature ( — .18.7) which has just
been computed for a perfectly radiating layer, mak-
ing the cloud temperature about + 10° C. Accord-
ing to Kirchhoff's law the earth within such a mantle
would also be at the same temperature, or higher if
warmed at all by internal heating. Thus it seems
unlikely that a perfectly cloudy earth could have
been glaciated at latitude 28° while a solar constant
of 2.0 calories prevailed. If the earth was not per-
fectly cloudy the conditions would have been less
favorable for glaciation and more like -those of the
present time.1 The matter of Permian tropical glaci-
ation is still more perplexing when we consider that
there was no glaciation simultaneously in temperate
and polar regions.

Reverting now to the hypothesis called (A) in
Chapter VI : If in Permian times and still earlier the
solar radiation alone was far too little to maintain the
surface of the earth or its clouds above freezing, then
glacial conditions would have been produced in those
times by a very moderate rise of land level. For sup-
posing conduction from within and radio-activity to
have been considerable sources of earth heat, and the
outer layers of the clouds not greatly warmed by the
sun, the thickness of the water vapor bearing stratum

1 The preceding argument does not tend to show that high tropical
mountains might not be glaciated. The reasons for cold tempera-
tures on high mountains have already been explained.

324

THE SUN AS THE EARTH'S SOURCE OF HEAT

would have been much less then than now, since the
vertical temperature gradient of the atmosphere
would have been far more rapid. Accordingly a
much less degree of elevation than now would suffice
to reach levels of comparatively free radiation to
space, and air currents cold enough to cause snow.

On this hypothesis we may represent the contribu-
tory influences which maintained the terrestrial cli-

TIME.

FIG. 65. — HYPOTHETICAL TEMPERATURE DIAGRAM.

mate in geologic time by the accompanying Fig. 65.
The upper curved line indicates the earth's tempera-
ture, the lower curved line what it would have been
if the sun had been the sole contributor of heat to the
earth. No attempt is made to draw the figure to
scale either in time or temperature, but only to illus-
trate the idea proposed.

The grave difficulty with our hypothesis of a low
intensity of solar radiation in early geological periods
is of course the question how the earth's surface tem-

325

THE SUN

perature was generally maintained above freezing.
We do not, to be sure, argue that no heat at all came
from the sun, but only that, while increasing, it had
not in Permian times reached perhaps three-fourths
its present intensity. No difficulty arises in sup-
posing that in the very earliest geological times the
earth's own heat sufficed completely to maintain its
surface temperature. The difficulty, lies in supposing
that the earth could have still contributed appreci-
ably after the enormous lapse of ages, estimated
roughly at 50,000,000 years, to Permian times. This
difficulty appears to me insuperable.

We will turn now to hypothesis (B) stated in Chap-
ter VI.1 According to Laplace's theory of the origin
of the solar system we are to suppose that when the
earth was formed the sun was expanded so as almost
to fill the orbit of the earth. Other nebular hypothe-
ses recognize the probable existence of much meteoric
nebulosity in the solar system at that epoch. It is to
be supposed that at that stage the sun itself was a
combined structure of nebulosity and condensation,
such perhaps as we see in the Pleiades stars (see
Plate XIX), and was not spherical, but still its
polar diameter was very much greater than now.
Under such circumstances the sun, with its outlying
appendages, as viewed from the earth would sub-
tend a great part of a hemisphere, so that its rays
would be nearly equally diffused all over the earth's
surface. Such a state of affairs would have pro-

1 This hypothesis was suggested by Chamberlin about 1898.
326

THE SUN AS THE EARTH'S SOURCE OF HEAT

moted the uniformity of temperatures we have
already noted.

Besides this, the 'intensity of the sun's radiation
must have been very small; for otherwise, coming
from so large an angle, it would have melted the
earth. If the rays came to the earth from a hemi-
sphere, and the sun then filled the earth's orbit, the
total output of the sun to space need not have ex-
ceeded half its present amount. Thus our hypothe-
sis may release us from the difficulty of understanding
how the sun could have radiated so long without the
temperature of the earth showing marked change.
For in early times we imagine the solar radiation of
very slight intensity, but the angle subtended by the
sun very large. With increasing solar density the
angle diminished, but the output of solar radiation
increased. We may even suppose considerable neb-
ulosity existed all about the earth, and that this neb-
ulosity, by reflecting solar rays, and by sending some
long wave rays of its own, helped to diminish the
rigor of the demands which geology induces us to
make on the sun in ancient times. Hence our hypo-
thesis (B) relieves us of the difficulty of the problem
of the supply of solar energy during the enormous
lapse of geological time.

As regards the possibility of tropical glaciation:
We may suppose that the full maintenance of ordi-
nary temperatures required formerly, as it does now,
the cooperation of the blanketing effect of the water
vapor of the earth's atmosphere; and that in addition
23 327

THE SUN

to this the earth's internal sources of heat were then
of some appreciable importance in maintaining its
surface temperature. The earlier the period we con-
sider, the greater we may suppose the contribution
of the earth's own heat, and the less the requirement
of the sun. But we may assume that all three fac-
tors, solar radiation, terrestrial conduction and the
blanketing effect of the earth's atmosphere were re-
quired to maintain genial temperatures in the Per-
mian period. As we have assumed that solar radia-
tion was nearly uniformly distributed over the earth's
surface, because of the large angle subtended by the
sun and of the reflection of radiation by still existing
outlying nebulosity about the earth, glaciation at the
tropics was accordingly no more difficult to bring on
then than glaciation elsewhere. Hence, a regional ele-
vation of land areas, or any other means of bringing
about a reduction of the efficiency of the atmosphere
as a blanket, in any locality, would have produced
local glaciation. Snow and ice once formed, would
help to perpetuate themselves by their high reflecting
power.

There are several ways in which the efficiency of
the atmosphere as a blanket may be altered. One of
these is by a considerable reduction of the atmos-
pheric humidity, and this, though somewhat unfav-
orable to great rainfall, would still be in line with the
known pronounced aridity of the Permian period.
But decreased humidity generally brings with it de-
creased cloudiness, which permits more solar radi-

328

THE SUN AS THE EARTH'S SOURCE OF HEAT

ation to be received, and thereby tends to raise in-
stead of depress temperatures. Hence it may be
that we are to look for the supposed regional altera-
tion of temperature rather in a considerable increase
of cloudiness, due to change in the relative arrange-
ment of land and oceans.

But, in whichever of the several ways suggested,
or in still others, the local decrease of temperature
might have been brought about, the hypothesis (B)
is evidently highly favorable to the explanation of
tropical glaciation, since it makes it just as easy to
produce ancient glaciations in the tropics as in polar
regions. When the Pleistocene period arrived we"
assume that the sun had so far shrunk that its influ-
ence was then, as now, zonal. We may further sup-
pose, if we choose, that the sun's radiation was less
then than now, and that this combined with other
causes to produce the Pleistocene glaciation. It is
well known that one of these causes was a consider-
able elevation of the glaciated areas.

Our hypothesis (B) seems to relieve us at once of
three formidable difficulties, and enables us to under-
stand: 1. How the sun has continued to suffice for
terrestrial needs throughout geological time. 2. Why
earlier geological periods were characterized by uni-
formity of climate irrespective of latitude. (3) How
it was possible to have tropical glaciation at all dur-
ing the Permian and earlier epochs; but especially
without evidence of simultaneous overwhelming gla-
ciation over all the temperate and polar zones of the

329

THE SUN

earth. Unfortunately the most favoring foundation
of hypothesis (B), namely, the Laplacian nebular
hypothesis, is now strongly attacked on dynamical
grounds, as we shall see in Chapter X. The substi-
tute proposed by Chamberlin and Moulton seems
less adapted to the line of explanation just given.
For it leaves little opportunity for the development
of solar heat by contraction, and besides does not
permit us to assume so widely spread sources of the
earth's supply of radiation in ancient times. Never-
theless Chamberlin himself put forth the rudiments
of hypothesis (B) about twelve years ago. We need
not yet despair that a nebular hypothesis may be
proposed as suitable to our purpose as to other re-
quirements.1

1 A reference to See's views is nuule in ( 'lutpler X.

CHAPTER VIII

Plant Requirements. — The Assimilation of Carbon by Autotrophic
Plants. — Etiolation. — Plant Geography. — Light Requirements
of Plants. — Heliotropism. — Plants as Energy Accumulators.

THE vegetable kingdom varies so widely in forms,
habits, and every characteristic of its members, that
the reader must not expect in this chapter a discus-
sion of all the sun's functions with respect to all
plants. But the higher plants, such as everybody
sees in the forests and fields, and which provide not
only food for man and beast, but countless materials
for building and the arts, are directly and indirectly
dependent in many interesting ways on the sun's
radiation. The subject of plant growth is so full of
cases of extraordinary adaptations that it is hard to
avoid digressing from the story of purely solar influ-
ences to speak of some of these ; and perhaps readers
may pardon a few such excursions from the main
highway of our subject.

PLANT REQUIREMENTS

The higher plants require carbon, oxygen, hydro-
gen, nitrogen, sulphur, phosphorus, potassium, cal-
cium, magnesium, and iron. Living vegetation con-

331

THE SUN

tains a very high percentage of water; but both of its
constituents, oxygen and hydrogen, also enter largely
into more complex compounds with carbon. As re-
gards their methods of obtaining carbon, plants are
classified in three groups: (1) Autotrophic, or the
self-nourishing, which obtain it through their leaves,
under the influence of light, from the carbonic acid
gas of the air. (2) Saprophytes, or scavengers, which
take it, in part at least, through their roots from de-
caying vegetable and animal organisms. (3) Para-
sitic plants, which take nearly all their nourishment
from living vegetation on which they fasten them-
selves. We shall practically confine our attention to
the first class, and when we use the term plant for
short in what follows, we shall mean generally auto-
trophic plants.

All plants are largely composed of water and most
of them employ it profusely in their vital actions.
A large birch tree, according to Von Hohnel's figures,
may send into the air through its leaves in one day
eighty pounds of water, which it has gathered mainly
from the soil by its roots. If 200 such trees grew on
an acre their water output in a season would perhaps
reach 1,500 tons. While not all trees and plants are
proportionally as free in using water as this, or indeed
can be, still they all require it, and depend on the sun
not only to keep water in the liquid form, but also to
promote the atmospheric circulation which promotes
the rainfall. These two functions, first maintaining
a proper temperature, second inducing a sufficient

332

THE SUN'S INFLUENCE ON PLANT LIFE

rainfall, are in this age almost wholly solar. Formerly
it may possibly have been otherwise.

From determinations of Konig we learn the follow-
ing percentage compositions of some of the common
vegetables.

TABLE XXVIII. — Composition of food products

Non-nitrog-

Water

Fat Ether
Extract

Nitrogenous
Material

enous (as
Carbohy-

Wood
Fibre

Ash

drates)

Wheat (grain) . .

13.65

1.75

12.35

67.91

2.53

1.81

Potato tubers . .

75.48

0.15

1.95

20.69

0.75

0.98

Beetroot

87.61

0.11

1.09

9.26

0.98

0 95

Lettuce (leaves)

94.33

0.31

1.41

2.19

0.73

1.03

The various chemical substances mentioned above
as plant requirements, and also some others, occur in
weak solution in the water which plants so plenti-
fully absorb through their roots. We cannot enter
into that profoundly interesting and difficult question
how this fluid rises to the tops of such immense trees
as the Sequoia and the Eucalyptus, which sometimes
reach heights of 500 feet, and in which the action of
gravity would tend to produce outward pressures
within the roots of fifteen atmospheres. Suffice it to
say that in some manner the fluids obtained from the
ground do reach all parts of the plant, and the water
copiously passing through the leaves evaporates.
This is called transpiration. The carbonic acid of the
air entering the leaves during light action in a manner
to be described later, is altered and combined with the
various elements transported from the roots. Com-

333

THE SUN

plex nourishing compounds produced in the leaves
descend to all living cells of the stem and roots, and
after undergoing further transformations even re-
ascend in spring to start the new growth of shoots,
leaves, and buds.

The various elements are not available to the plant
in all their chemical combinations, and in some com-
binations may even be poisonous. Without going
into details, it will be interesting to note the case of
nitrogen. This element, found uncombined as a gas
in air, is rather inert chemically, and none of the
higher plants seem able to make use of it in its free
state. Ammonia, too, though prevalent as a product
of decay in the soil, and existing also in the air, is not
nourishing to most of the higher plants. Nitrites are
said to be poisonous in moderate concentration, al-
though in very dilute solution perhaps useful. Ni-
trates, then, are to be regarded as the principal nitro-
gen sources for autotrophic plants. In agriculture
the removal of crops withdraws available nitrogen
from the soil faster than ordinary processes can pro-
duce it, hence the use of fertilizers containing salt-
peter. But the leguminous plants, such as peas,
beans, clover, alfalfa, etc., are said to be able to use
free nitrogen, and it is customary to plow under green
crops of such nature to improve the soil. Careful
researches have shown that in reality certain micro-
organisms, often present in the soil, cause the forma-
tion of nodules on the roots of these leguminous
plants, and that atmospheric nitrogen is only assimi-

334

THE SUN'S INFLUENCE ON PLANT LIFE

lated when these micro-organisms inhabit the nod-
ules. Different Leguminosce require different species
of micro-organisms for a successful partnership of
this kind. The micro-organisms combine the free
nitrogen into forms useful to the plant, and the plants
supply other materials, perhaps carbohydrates, for
the micro-organisms. This is but one of many in-
stances in which the higher plant forms are proved to
depend upon the activities of the lower, quite as much
as the lower on the higher. Quite recently it has be-
come possible to purchase at large seed stores cultures
of the proper micro-organisms, with instructions for
multiplying them, so that when sowing a field with
clover or alfalfa, for instance, the cultures may be
mixed with the seed so that it may be certain that the
crop will not suffer from lack of nitrogen.

THE ASSIMILATION OF CARBON BY AUTOTROPHIC
PLANTS

Many plants, among which are corn and others of
the most valuable food plants, will thrive in water
cultures as well as in the soil, although the supply of
carbon through their roots is made impossible. Hence
the source of the carbon which is a fundamental ele-
ment of all organic life must be, in such cases, the air.
If grown in darkness, although in all other respects
the conditions are retained identical, no considerable
gain of carbon occurs and the plants remain white,
for no chlorophyll is formed. We find, then, that
carbon dioxide of the air is taken in under the influ-

335

THE SUN

ence of light, and is acted upon to form complex com-
pounds with hydrogen and oxygen, such as hexose,
sugar, starch, and also the nitrogenous carbon com-
pounds which go to compose vegetation. This ac-
tion is found to require the green chlorophyll bodies
of the living plant cells, and chlorophyll, as we have
said, is not produced without light. Oxygen is given
off in the chemical transformations and escapes from
the leaves. The process of absorbing carbonic acid
and transforming it with evolution of oxygen, as just
described, is called assimilation.

The evolution of oxygen may be demonstrated in a
simple way by cutting a branch of the water plant
Elodea Canadensis and placing it in a tube of water
charged with carbonic acid gas. If retained in very
dim light for some time nothing easily noteworthy
occurs, but when well illuminated a stream of bub-
bles escapes from the cut ends. By inverting a test
tube, previously filled with water, the gas may be col-
lected, and will be found by testing it with a glowing
coal to consist largely of oxygen. Quantitative experi-
ments have been made by counting the bubbles given
off in this manner, and it has thus been shown that
their number is usually nearly proportional to the in-
tensity of the light. In darkness no oxygen is given
off, but carbon dioxide is slowly evolved instead. This
reverse process is called respiration.

As already stated, an essential condition for carbon
assimilation is the presence in the plant of the green
substance of chlorophyll. This is found in most

336

THE SUN'S INFLUENCE ON PLANT LIFE

plants almost solely in the leaves, so that these are
the great organs of assimilation. Chlorophyll in alco-
holic solution has a fine fluorescence. It appears
green by transmitted light and red by reflected light.
The spectrum of crude chlorophyll in alcoholic solu-
tion is characterized by six absorption bands. Three
are in the violet and merge together in strong chloro-
phyll solutions. The other three occur in the green,
yellow, and red respectively. By treating the alco-
holic solution with benzine the crude chlorophyll,
which is a mixture, will yield in benzine solution a
blue green dye which seems to be the more important
component. This itself is complex, and contains
among other constituents one which is called phyllo-
porphyrin, and differs only a very little chemically
from the hsematoporphyrin of blood. But however
curious and interesting chlorophyll may be, its spe-
cial function, the promotion of carbon assimilation,
does not go on except the chlorophyll be in the living
plant cells. Artificial chlorophyll bearing cells will
not answer.

It has been shown that for every volume of carbon
dioxide operated on by the plant, an equal volume of
oxygen is liberated. Among the principal products of
the reaction is glucose or starch. Such facts may
imply some such actions as are expressed in the fol-
lowing symbolic manner.

6CO2 + 6H2 O = Cn H12 O6 (Glucose) + 602;

or 6CO2 + 5H2 O = C6 H10 O6 (Starch) + 602.

337

THE SUN

Starch is readily demonstrated as being produced in
many plants during light action, but plants of dif-
ferent families vary greatly in the quantity of it they
produce. Indeed, as we shall see many times, the
different plants behave so differently under given
conditions that hardly a single general fact can be
stated, in regard to which some kinds do not exhibit
exceptions. As one person is repelled by coaxing and

moved by argu-
ed ment, while another
goes only as senti-
ment dictates, so
the plants seem to
have their diverse
characters, and two
kinds may react
oppositely to the
same stimuli.

The organs of
carbon assimilation
are the leaves, and
in these the portals of access are the little open-
ings called stomata. These exist in most plants most
plentifully on the under surfaces of the leaves, al-
though found in some only on the upper surfaces and
in still others on both. They are very minute slit-
like orifices, so small that a needle prick is a huge hole
in comparison with one of them. A single leaf of a
sunflower may have no less than 13,000,000 stomata.
Fig. 66 (after Schwendener) gives a general idea of

FIG. 66.— STOMATA. (Schwericlener.)
A, Cross-section on tn, n. B, Plan view
of half-stoma omitting parts outside a,
6. C and D, Closed and open stomata.

THE SUN'S INFLUENCE ON PLANT LIFE

the form and surroundings of these minute but nec-
essary organs in the Amaryllis. There are special con-
trivances called guard cells adapted for opening and
closing the stomata. These guard cells, when dis-
tended by containing much liquid, or when shone
upon by strong light, cause the stomata to open wide,
thus promoting the assimilation of carbon dioxide,
and also the transpiration of water vapor, of which
we shall speak later.

Although very numerous, the combined area of the
stomata, even when wide open, is hardly more than
one per cent of the area of the leaves, so that it was
long a mystery how so much carbon dioxide could
pass through them. This question was solved by
Brown and Escombe (1900). They found that when
carbon dioxide is admitted through an orifice to a
medium capable of absorbing the gas as fast as re-
ceived, the amount which diffuses through the open-
ing decreases with the diameter, not with the area of
the opening. This seeming paradox is explained by
supposing the velocity of flow to increase as the open-
ing decreases, so that a smaller hole accommodates
the diffusion not merely from directly above, but also
from the side areas which were before served by the
larger hole. The observers found that their strange
new law held for numerous openings as well as for
single ones, provided the openings are separated by
distances as great as eight or ten times the diameter
of the holes. From this it follows that a surface
pierced, like a leaf, with extremely numerous but very

339

THE SUN

small openings can admit the passage of a gas by dif-
fusion at almost as rapid a rate as if the whole area of
the surface were one hole. This extraordinary dis-
covery raises our admiration of this excellent con-
trivance of Nature, whereby the whole area of the
leaves of a plant is as if available to transmit nour-
ishment from the air, and to permit the escape of
water vapor, although in reality nearly all of this
area is actually closed, to protect the delicate cells
within.

The rate of assimilation of carbon, or what is al-
most proportional to it, the rate of gain in dry weight
of a plant, depends on various factors. Of these we
may mention first the concentration of the carbon
dioxide in the air. Although, according to Eber-
meyer's estimates, a square mile of forest uses up over
500 tons of carbon dioxide in a year, so that the
demands of the plant life of the world are really
enormous, there is nevertheless an almost steady,
and everywhere nearly uniform, percentage of car-
bon-dioxide in the air — about three parts in 10,000.
The steady drain of plant life is to be set over
against the production of carbon dioxide by the res-
piration of animals, the burning of wood and coal
and other sources of supply, but it is surprising that
the atmospheric proportion remains so nearly uni-
form as it does. Geologists are by no means of the
opinion that this proportion has always been the same
as at present. It is therefore of interest to inquire how
the assimilation varies with the concentration of car-

340

THE SUN'S INFLUENCE ON PLANT LIFE

bon dioxide. There seems to be some disagreement
between investigators as to the precise optimum con-
centration, but all are agreed that the rate of assim-
ilation of carbon increases steadily with the concen-
tration of carbon dioxide up to a concentration of at
least more than ten fold the present. Under such con-
centrations the rate of assimilation may reach more
than twice its usual value. According to some, the
increase of assimilation is even directly proportional
to the increased concentration of CO2, within these
limits. While, of course, this change has no practical
interest while the carbon dioxide of the air remains
constant, yet it may have been of considerable prac-
tical importance to vegetation in past geological
epochs when the air was more highly charged.

Temperature is a still more important agency in
regulating the growth of plants. Assimilation may
be recognized with some plants at temperatures
of several degrees below freezing, but practically
speaking all growing plants of the higher forms
must be maintained at temperatures between 0° and
50° C. The increase of the rate of assimilation for
most plants is very rapid from 0° up to a temperature
of about 35°, and at higher temperatures than this
there is a still more rapid decrease in the rate. It is
an interesting question whether the principal forms
of vegetation could flourish on any planet if the mean
temperature lay below 0°, or above 50°. Although we
cannot answer this question absolutely, still it seems
probable that the answer must be in the negative.

341

THE SUN

At all events, we see not only how entirely our own
lives now depend on the sun, but even on that nice
balance between the receipt of solar radiation and the
emission of terrestrial radiation, in which even the
amounts of water vapor and cloudiness are important,
as stated in the preceding chapter.

We now take up the dependence of carbon assimi-
lation on light, deferring the consideration of other
effects of light on growth. Plants raised in darkness
do not become green. The formation of chlorophyll
and the assimilation of carbon require radiation be-
tween wave lengths 0.39/t and 0.77/4. Experiments
on the relative effectiveness of rays of different wave
lengths are not altogether satisfactory. They have
been confined to a few kinds of plants, and great dif-
ficulty is found here, as in physics and astronomical
work, in separating a sufficiently intense nearly
monochromatic beam of light, and in measuring its
intensity. Investigations were made about thirty
years ago on the relative efficiency of the different
rays by Reinke and by Engelmann. They agree in
fixing the wave length of maximum effect in the red at
about 0.65//, to 0.70/A, but Engelmann found a secon-
dary maximum in the blue at 0.48/-6, not found by
Reinke. Engelmann's observations distinguish be-
tween the assimilation of the upper and lower sides
of a leaf capable of such action, and he finds the posi-
tion of maximum effect shifted distinctly towards
shorter wave lengths for the surface which receives
its illumination through the leaf. This result de-

342

THE SUN'S INFLUENCE ON PLANT LIFE

pends, it is thought, on the strong absorption by chlor-
ophyll for red rays, for thereby the light which pene-
trates the leaf is greatly weakened in its longer waves.
Undoubtedly the relative activity of different wave
lengths of light in promoting the assimilation of car-
bon is closely associated with the absorption spectrum
of chlorophyll, as indeed would appear from Engel-

042

6*50

$66

FIG. 67. — PROMOTION OF CARBON ASSIMILATION BY LIGHT (full curve)
AND ABSORPTION OF LIGHT (dotted curve) IN GREEN LEAVES.
(Engelmann.)

mann's results, given in Fig. 67. There is needed
much more research in this difficult field. It would
be greatly promoted by the introduction of means for
obtaining nearly monochromatic light of well deter-
mined and adequate intensity, covering considerable
areas suitable for plant growth under otherwise nat-
ural conditions.

24 343

THE SUN

Kniep and Minder l have recently made observa-
tions by the bubble method with Elodea Canadensis
on the assimilation of carbon dioxide in lights of dif-
ferent colors. They used sunlight filtered by colored
solutions so as to select red light (wave length 0.62/4
to some point in the infra-red not determined), green
light (wave length 0.512/t to 0.524/*), or blue light
(wave length 0.35/1, to 0.50//,) at pleasure. In each case
the light could be reduced to a fixed intensity as meas-
ured by a Rubens thermopile, so that they could in-
vestigate the rates of assimilation under equal inten-
sities of total radiation for each of the three colors.
They found the green light of no effect in producing
assimilation. It was like no light at all. The red
and the blue they found equally effective. Hence
their experiments tend to support Engelmann's, as
they indicate two wave-length regions efficient to pro-
duce assimilation. We must wait for more elaborate
experiments before we shall know just how the effi-
ciency varies with the wave length, and whether all
plants are best promoted by the same rays. It is
clear, however, that as the red end of the spectrum
predominates in direct sunlight at the earth's surface,
whereas the violet end greatly predominates in sky
light, a plant may be made to assimilate carbon pre-
dominatingly by red or blue light according to
whether it grows in direct sunshine or not. This
may offer a method of evolving new plant forms as
we shall see in the next section.

1 Zeitschrift fur Botanik, vol. i, pp. 619-650, 1909.
344

THE SUN'S INFLUENCE ON PLANT LIFE

Experiments have been made on the dependence
assimilation on the total intensity of light irrespective
of wave length. Results very naturally differ for
plants of light-loving and shade-loving habits. In
general, the rate of carbon assimilation increases
nearly directly proportionally to the intensity of the
light, but this ratio of course cannot persist for ex-
tremely high intensities, because, first, of injury to
the plant, and, second, of deficiency of other promot-
ing elements, especially carbon dioxide.

ETIOLATION, OR EFFECTS OF DEFICIENCY OF LIGHT

Plants grown in darkness or weak light tend to
have long stem internodes and leaf stems, and small
white or yellow leaves. These and other effects of
deficiency of light on plant growth are termed etiola-
tion. As already stated, the higher plants do not in-
crease much in dry weight unless exposed to light, so
that experiments on the effects of complete darkness
on growth are mainly restricted to such species as
have large food stores in their seeds or tubers. The
object served by natural etiolation is at length to
bring the leaves of the plant to suitable illumination,
as is seen by the tall tree stems and climbing vines in
closely growing forests. In experimental work this
result cannot, of course, be reached, but nevertheless
the tendency is plainly shown.

Although, as stated above, the effect of- darkness
or very weak illumination is to restrict the leaf area,
leaves grown in moderate light are larger and thinner

345

THE SUN

than those grown in full sun. This form of etiolation
is of importance to the tobacco industry, since large
thin leaves are preferred and command much higher
prices. Accordingly, in Connecticut and Florida
large fields of tobacco are now grown shaded by tents
of open-meshed cloth to promote this improvement.
Other desirable results obtained by this device con-
sist in the greater and more uniform humidity of the
air and soil and the prevention of disastrous winds
and hail.

With many kinds of plants the buds will not de-
velop if the light is too weak, and there are besides
many other effects embraced by the general term
etiolation. Curiously enough red light, which, as we
have seen, is highly effective for promoting carbon
assimilation, in many cases behaves like darkness in
respect to etiolation. It is thus possible to grow
plants under conditions favorable to their adequate
nourishment, and at the same time to greatly alter
their forms by etiolation effects. This interesting
feature perhaps offers opportunities for promoting
the evolution of desired forms in useful plants.

PLANT GEOGRAPHY

In natural surroundings there is a very great range
of light intensity, and with it a great range in temper-
ature and moisture. These circumstances produce
very marked effects on plant life. In the tropics
abound regions of great rainfall, from 100 to 500 cen-
timeters (40 to 200 inches) annually, with the aver-

346

THE SUN'S INFLUENCE ON PLANT LIFE

age cloudiness as high as fifty to sixty per cent. The
mean temperature being also high, 25° to 28° C.,
there results a high atmospheric humidity. Such re-
gions are the home of the tropical rain forest, which,
viewed from a ship presents a noticeably irregular
skyline filled with every shade of green, but tending
toward the more sombre hues. Flowering trees are
occasionally conspicuous. The interior of such a
forest teems with a varied mass of vegetation from
the ground to the top of the highest trees. Vines and
rugged ferns abound, so that the traveler's way is
almost impassable. Fruits and flowers are plentiful.
Parasitic and saprophytic plants revel among the
luxurious surroundings. On viewing a tall tree one
can scarcely distinguish which is its own foliage and
which that of the dependent vines and parasites that
load its trunk and limbs almost to breaking.

Sharply contrasting with such scenes as this are
the sub-tropical deserts like the African Sahara.
Here also the temperature is high, but variable, rang-
ing perhaps 20° C. in a single day. Rainfall may be
as slight as 5 centimeters annually, but more often
reaches 10 to 20 centimeters. The scanty vegetation
is provided with extraordinary contrivances to re-
duce as far as possible the loss of water by transpira-
tion. Leaves are small, thick, glossy and waxy, their
stomata protected heavily. Thorns abound. The
roots run very deep so that even at one or two meters
in depth they have hardly diminished at all in size.
As a rule only small plants and shrubs are found.

347

THE SUN

Some varieties have special reservoirs for the storage
of water.

The periods of rest are not conspicuous in the trop-
ical vegetation, for they are not governed by temper-
ature, though perhaps often by rainfall. Tropical rest
periods are frequently localized to single trees or parts
of trees; but in the temperate and arctic zones there
occurs in winter manifestly a general cessation of
growth. Not all temperate and arctic trees cast
their leaves, but they generally rest from the growth
of shoots during the cold months. Askenasy has in-
vestigated these matters at Heidelberg for the gean
tree, which may serve as a type for other broad-
leaved trees. The season of activity lasts from about
mid-April to mid-October. It comprises the period
of growth of foliage, April-May, during which next
season's foliage buds appear; then, May-September,
follows the period of assimilation during which stems
and roots enlarge and the next season's flower buds
are formed; then comes the period of decline ending
in the fall of leaves. During the summer the growth
of next season's buds is slow, and ceases altogether
from October to early February. Then a growth
begins and becomes more and more rapid. Although
a warm March greatly accelerates the development,
a warm October cannot start growth. From the end
of November, development may be forced by hot-
house conditions. During the rest period chemical
changes of the reserve material go on, and it is indeed
transported between different organs of the tree,

348

THE SUN'S INFLUENCE ON PLANT LIFE

Temperate forests contrast with the tropical rain
forest described above in the relative absence of vines,
parasitic vegetation, and undergrowth, although in
moist regions herbs and shrubs are not lacking. The
evergreen conifers are more and more in evidence at
higher latitudes, but these become dwarfed toward
the arctic zones. The growth period of arctic flora
is limited to about two months, but is favored by the
fact that the sun is then continually above the hori-
zon. All varieties start their growing almost simul-
taneously, and reach their flowering stage almost
together, within a couple of weeks. Although the
mean temperature of the air during the growing pe-
riod may be 5° C. or more, the soil is frozen almost to
the surface.

Wiesner has made extensive photographic re-
searches to. determine the light requirements of
plants. He employs a modification of the method of
Bunsen and Roscoe. A normal photographic paper
is prepared by soaking in three per cent common salt
solution, drying in darkness, soaking five minutes in
twelve per cent silver nitrate solution, and drying
again in darkress. A normal tone or grade of dark-
ening is prepared by coating a paper with a mixture
of one part lampblack in 1,000 parts zinc oxide. When
the photographic paper reaches this shade by expos-
ure to light for one second, the light is said to be of
unit intensity of the Bunsen-Roscoe scale. Those au-
thors showed that for equal blackening of the photo-
graphic paper the intensity of the light, between wide

349

THE SUN

limits, is inversely as the time required. Hence if n
seconds are required to produce normal tone, the
light intensity is 1 / n Bunsen-Roscoe unit. To
avoid inconveniently long exposures in deeply shaded
places, and to allow sufficient time for accurate re-
sults in strong light, Wiesner introduced a gradation
of shades, forming a kind of tone scale, which he
standardized in terms of the normal tone.

By such procedures Wiesner has measured the
light action due to direct sunlight, and to diffused
light at Buitenzorg (Java), Cairo (Egypt), Vienna
(Austria), several stations in Norway, and Advent
Bay (Spitzbergen) . His measures were made on days
varying in brightness from cloudlessness to rain and
snow. The Vienna measurements extend for several
years. He has made observations in the open, in
leafy tree crowns, and under the shadow of thick for-
ests. It is not possible to give here any adequate
summary of this very extensive work, but the reader
may consult the original articles of Wiesner.1

Some of Wiesner's results are as follows: The max-
imum total illumination at Vienna was 1.50 B-R
units; at Buitenzorg, 1.61. At Vienna the mean
midday value ranges from 0.1 B-R unit in January to
0.96 in July. At Buitenzorg in December and Janu-
ary the midday values range from 0.65 to 0.85. Rain
or snow diminishes the light total to one-tenth or less

1 Especially in Sitzungsberichte Wien. Akad. Math. Naturw. KL,
Bd. 102, I, 1893; 104, I, 1895; 109, I, 1900; 113, I, 1904. Also
Denkschriften of the same Academy, Bd. 64, 1897; 67, 1899.

350

THE SUN'S INFLUENCE ON PLANT LIFE

of its normal value. At Vienna the ratio of direct
sunlight to diffused skylight action is very variable,
but an average value for several hours near midday is
about unity. On half cloudy days the total light
action is almost as strong as on cloudless days. On
completely cloudy, but not stormy, days the total
light action is reduced from three- to five-fold.

Taking the total direct and diffused light action in
the open as the basis of reckoning, Wiesner compares
with it the light action found in the crowns of trees
and elsewhere. Calling the first value I, the second
i, he calls the ratio (i/\ = L) the relative photic ration.
When the leaves are beginning to form in spring, be-
fore they get large enough to cast deep shadows, the
values of L within tree crowns and under trees are
not greatly less than unity. But later in the summer,
when leaves are full grown, and next season's leaf
buds are forming, these ratios become much smaller.
For instance for the white birch (Betula alba),
Wiesner finds:

Date

Observed values of I

L 1 Da-v>s 1
**t Minimum J

Total daylight

In tree crown

April 16

0.834

0.333

2.5

May 1

0.875

0.219

1
4

May 14

1.122

0.142

1
8

May 29

1.200

0.109

351

THE SUN

This rapid increase of the shading action of forest
trees, as they develop their leaves, determines the
nature and habit of the underbrush. Generally the
leaves of the underbrush present a scattered, or flat
array, so as not to shade one another. Often the
undergrowth has the habit of rapid development of
leaves and blossoms in early spring, before the over-
growth is fully leaved.

In arctic regions the vegetation, almost without
exception, requires practically all the available light.
This depends, no doubt, on the coldness and short-
ness of the season of growth. Values of L much
below unity seem to be insufficient for arctic plants.
This may explain the absence of tree forms there.
Whereas in the tropics, and even in temperate zones,
most plants have means for reducing the light action
on their leaves, no such contrivances are common in
the frigid zones.

The range of light requirements is indicated by the
following values of the relative photic ration and
total light action within the crowns of trees in full
leaf. The terms (Max.) and (Min.) refer to the max-
imum and minimum daily values of the quantities
concerned.

Among underbrush growing in a shade so deep that

L = — he found beeches, maples, and other well

growing saplings. Grasses in the temperate zones

were found, although not blooming, when L = .

352

THE SUN'S INFLUENCE ON PLANT LIFE

Common 'name

L (Min.)

I (Max.)

Remarks

Boxwood

0.012

•

Beech

108
1

0 015

Isolated tree

Maple

85
1

0.030

Isolated tree

White poplar

0.086

Isolated tree

Pine
White birch

15
1
11
1

0.118
0 144

Enclosed tree
Enclosed tree

Ash

9
1

0 224

Enclosed tree

Larch

5.8

0.260

Isolated tree

Blackthorn

5
1
1.3

0.722

Blooming but not leaved

Some tropical grasses survive L =

100'

Lichens

were found in the tropics which had the photic
ration only L = — -. Many forms of tropical or-

chids, epiphytes, and other shade-loving plants, were

found to thrive under photic rations from — to — .

10 ou

We cannot dwell longer on the interesting work of
Wiesner. From it we see how unnecessary it is, for
many forms of vegetation, that the light should be of
the full intensity which is now available in the open.
Indeed, Wiesner remarks that in experiments made
by rotating plants, so as to get equalized illumination,
the buds will develop and leaves be fully grown under
illuminations far below the minima observed under

353

THE SUN

natural conditions. In the natural state the well-
illuminated buds grow at the expense of their less
favored neighbors, and as their leaves expand they
tend still further to suppress the undeveloped buds.
In view of all this, and in view of the hypothesis (B)
advanced in Chapters VI and VII, which treated of a
more uniform illumination assumed to be formerly
prevailing, it is interesting to speculate whether the
great vegetation of the Carboniferous era was not
produced under a far more feeble illumination than
that which now prevails.

Considering the present lack of exact experiments
on the efficiency of different wave lengths of light to
promote plant growth, the photographic experiments
of Wiesner are perhaps all that are yet demanded.
But we can easily see the advantage which would re-
sult to plant physiology if such an instrument as the
the spectrobolometer could be employed in skilled
hands to determine the relations of wave length and
intensity of light to carbon assimilation and etiola-
tion, for numerous plant forms.

HELIOTROPISM

It is well known that different plants vary as re-
gards the angles which their organs present to the
direction of strongest light. A nasturtium, for in-
stance, if principally illuminated from one direction,
will expose almost every leaf and flower with its face
broadside toward the light. Plants within a room
bend toward the window. Some species which live

354

THE SUN'S INFLUENCE ON PLANT LIFE

in dry and cloudless regions present their leaves edge-
wise to the strongest illumination. Such adaptations
as those we have mentioned, and others, are em-
braced under the term heliotropisrn. Different plant
organs differ in respect to the matter, so that botan-
ists distinguish organs which are orthotropic, and
those which are plagiotropic, according as they tend
to lie in the direction of the principal light or at some
other angle with respect to it. Also orthotropic
organs may grow in a positive heliotropic manner,
i. e. towards the source of light, or the contrary.
Roots are usually negatively and stems positively
orthotropic, while leaves may be regarded as plagio-
tropic.

It was supposed by De Candolle (1832) that helio-
tropisrn was a simple consequence of different rates
of growth between strongly and weakly illuminated
parts of an organ. It had been found (as already
stated under etiolation) that stems grown in darkness
exceeded in length those grown in light. Further-
more, it has been shown that plants increase in stat-
ure faster by night than by day. See, for instance,
the following measurements by Kraus on the growth
of a species of bamboo at Buitenzorg, Java, in
twelve-hour intervals :

Date

Dec. 4

Dec. 5

Dec. 6

Dec. 7

Dec. 8

Growth by day ....

10.5 cm.

4.5 cm.

8cm.

8.5 cm.

12cm.

Growth by night . . .

16cm.

15 cm.

16 cm.

12.5cm.

355

THE SUN

On such grounds De Candolle assumed that helio-
tropic curvature was simply the effect of the retard-
ing influence of light on the growth of that side of the
stem most strongly illuminated. This simple ex-
planation may have some justification, but it is not
adequate to explain the facts. For plant organs
which curve away from the light also grow faster in
the dark. Furthermore, the same organ may react
either positively or negatively or not at all according
to the intensity of the light, as shown by the experi-
ments of Oltmans. This author is of the opinion
that the best intensity of illumination for the general
welfare of the organism is that at which it exhibits no
heliotropic curvature. Direct sunlight is too bright
to induce heliotropic curvature in most plants, hence
they do not as a rule turn their leaves from east to
west with the progress of the sun, although in the
case of the sunflower this occurs with the blossoms.

It seems that illumination acts rather as a stimulus
than as a force in producing heliotropism, for the
effect may be produced by brief light action and the
actual curvature take place in the appropriate direc-
tion after the light has been withdrawn. Further-
more, the reaction does not necessarily occur where
the light is applied, but the stimulus may be trans-
mitted some distance from the sensitive recipient
organ to the position where curvature takes place,
although the part of the organ where it becomes
curved is shielded entirely from the action of the
light.

356

THE SUN'S INFLUENCE ON PLANT LIFE

Heliotropism is without doubt of great value to
plants in enabling them to adjust their leaves most
advantageously to increase or reduce the illumina-
tion in which they find themselves. It is especially
valuable to many compound leaved plants subjected
to the powerful heating effects of the direct rays of
the unclouded sun. They open out their leaves in
the early morning or during cloudy weather, but tilt
them up edgewise in the hot sun, thus reducing the
effective area for heating. Such plants, though of
large leaf area, may thrive in the driest regions.
Other plants have their leaves permanently set at
such an angle as to receive a minimum of direct sun-
light.

On the other hand many plants growing in com-
paratively weak light, and some sun-loving plants in
the open, turn the broad sides of their leaves toward
the strongest light. The negative heliotropism of
the roots of plants is of advantage, for by it they may
be saved from growing out of the soil.

PLANTS AS ENERGY ACCUMULATORS

The energy now available in coal and oil was for-
tunately preserved for our use in the decomposed veg-
etation of former ages. Extraordinary luxuriance of
vegetation is thought to have prevailed in those
ancient times, and we now use the accumulated en-
ergy of solar rays emitted long before the existence
of man. Attempts to employ solar energy by arti-
ficial engines will be referred to in the next chapter,

357

THE SUN

but none of them can as yet be compared in econom-
ical success with the natural process of storage always
occurring in vegetable growth. Artificial processes
are for the moment far more efficient, but not in pro-
portion to their great cost, and none of them, like the
natural process, stores permanently the energy re-
ceived. Most solar engines transform solar radia-
tion immediately into heat, and this is gradually lost.
Growth transforms solar radiation immediately into
chemical energy, and this may be indefinitely pre-
served.

Various estimates have been made of the efficiency
of plants as transformers of energy. Pfeffer (1871)
computed from Boussingault's work that a square
meter of Nerium leaf surface formed starch at the
rate of 0.000535 grams per second. Assuming the
product formed to have a heat of combustion of 4,100
calories per gram, he found 2.2 calories per square
meter per second to be the amount of energy eon-
served. The amount of energy received from the sun
would depend on the time of day, inclinations of the
leaves, moisture of the air, etc., but might be esti-
mated at about 150 calories per square meter per sec-
ond in ordinary conditions near sea-level. This
would give an efficiency of about 1.5 per cent.

Brown, in his Bakerian Lecture (see Nature, vol. Ixxi,
p. 522), summarized some careful experiments on the
efficiency of the sunflower. He made estimates of
the temperatures of leaf surfaces and of their thermal
emissivity. The latter in still air was about 0.015

358

THE SUN'S INFLUENCE ON PLANT LIFE

calories per square centimeter of leaf surface for 1° C.
difference of temperature from the surroundings.
Leaves evolve carbon dioxide in darkness in their or-
dinary process of respiration. For a sunflower leaf
respiring 0.7 cubic centimeters of carbon dioxide per
100 square centimeters per hour, the respiration
causes a rise of temperature of the leaf in still air of
0°.019 C. above its surroundings. This effect is
therefore practically negligible. Not so the effect
of transpiration and evaporation of water, especially
in windy surroundings, for this may alter the temper-
ature of leaves by several degrees. The absorption
coefficients of leaves of various plants in ordinary
sunlight were determined. These range from sixty-
five to seventy-eight per cent, and for the sunflower
leaf was found to be 68.6 per cent. Such values
would probably differ according to the quality of the
light. The rate of absorption of carbon dioxide by
the plants was measured. Air was drawn through
the glazed case containing the leaf specimen and the
carbon dioxide contents of the air after passage was
compared with that of air unaffected by the plant.
Various concentrations of carbon dioxide were experi-
mented upon, and it appeared that up to concentra-
tions six times the normal, the rate of assimilation
was proportional to the concentration of carbon diox-
ide in the air. The material formed by the plant was
assumed to be hexose, whose heat of combustion is
3,760 calories per gram. The rate of assimilation
seemed to be independent of the intensity of the
25 359

THE SUN
light, until this was reduced as low as 0.04 calories

per square centimeter per minute, or say -— of or-

dinary sunlight. Hence the efficiency of the plant
appeared much higher under weak illumination. In
some cases the efficiency found was as high as five per
cent, but not often above 1.7 per cent.

Two numerical illustrations will show the charac-
ter of the results. Both deal with the sunflower leaf
(Helianthus annuus). In the first case the activity
of the leaf did not suffice to expend all the solar energy
it absorbed and the leaf was above the temperature of
the surroundings. In the second case, owing to the
high temperature, the fully opened stomata, and the
low humidity prevailing, there was rapid transpira-
tion and the contrary state existed. The numbers
given in the first part of Table XXIX apply to the
energy reaction per square centimeter of leaf surface
per minute. In the latter part of the table is given
the disposal of the leaf in percentages of the solar
energy received, plus the heat energy received from
the surroundings.

It appears from such investigations as have been
made that plants may store up as chemical energy in
round numbers one or two per cent of the energy of
solar radiation which shines upon their leaves. This
may seem a very small efficiency, but on its results
accumulated through former ages have depended the
great manufacturing achievements and the comfort-
able winter warmth of our dwellings for many years.

360

THE SUN'S INFLUENCE ON PLANT LIFE

TABLE XXIX. — Economy of Helianthus annuus

Case A

Case B

Total solar radiation received
Amount absorbed
Amount of energy used vaporizing
water

0.2569cal.
0.1762

0 1243

0.2746cal.

0.1884

0 3668

Amount of energy used in photo-syn-
thesis

0 0017

0 0033

Amount of energy lost by cooling
Velocity of wind in meters per minute. .
Temperature of leaf above surround-
ings

+0.0502

428

+0°.43C.

-0.1817
200

-1°.84C.

Energy used in photo-synthesis
Energy used in transpiration

per cent.
0.66
48.39

per cent.
0.72
80.38

Solar energy transmitted by leaf

31 40

18 90

Heat energy lost to surroundings

19.55

In the combination of water power and electricity we
seem now to be passing in a measure away from the
dependence on coal and steam, but there is little ques-
tion that both coal and oil will long remain in exten-
sive use to remind us of our dependence on the
growth of ancient vegetation and its transformation
of solar radiation into chemical energy.

CHAPTER IX

UTILIZING SOLAR ENERGY

Experiments with Burning Mirrors. — The "Hot-box" Principle. —
Mouchot, Pifre, and Ericsson. — Eneas' Solar Engines. — Proper-
ties of Glass. — Solar Heaters and Reservoirs. — Low Temperature
Solar Engines. — Solar Cooking Appliances. — Solar Metallurgy. —
Resume*. — • Quantity of Solar Energy Available. — Thermo-
dynamic Efficiency. — Reflecting Powers of Mirror Surfaces.

AT present the manufacturing and commerce of the
world is mainly carried on by aid of coal or internal
combustion engines, which derive their fuel from the
decomposed products of prehistoric masses of vege-
tation in which were stored a small fraction of the
solar energy of those bygone times. The modern
great development of water power, electrically util-
ized, also depends on the sun; for by solar heating
water is evaporated from oceans, lakes, rivers, and the
soil, is transmitted inland and precipitated by the
atmospheric circulation which the sun's heat main-
tains, and comes in use when it flows down in the
rivers. Another immense source of water power, not
as yet much utilized, resides in the ocean waves and
tides, which also depend in a high degree upon the
sun. It is not necessary to discuss further these well-
known sources of power, and we shall pass to the vari-
ous means which have been proposed for using the
energy of the solar rays more directly.

UTILIZING SOLAR ENERGY

EXPERIMENTS WITH BURNING MIRRORS

It is said that during a siege of Syracuse in the year
214 B. c. the renowned philosopher Archimedes
burned or scattered the Roman fleet under Marcellus
by concentrating sun rays upon the ships by means of
mirrors erected on the shore. Whatever may be the
truth of this story, which has been doubted, such means
of warfare are not likely to be revived in our day.

Buffon, the French naturalist (1707-1788), tested
the possibility of the circumstance just described. In
1747 he made many experiments with a burning mir-
ror constructed by mounting 360 plane glass mirrors,
each 16 X 22 centimeters, on a frame in such a man-
ner that each could be adjusted separately, so that
all could concentrate their reflected rays to a focus at
any desired distance. Corresponding to the angular
diameter of the sun, the focus was about 44 centi-
meters in diameter at 50 meters, and proportionately
less at shorter focal distances. He found it possible
to set fire to wood at 68 meters. With 45 mirrors he
melted 3 kilograms of tin in a pot, at 6.5 meters, and
with 117 mirrors melted silver at the same distance.
By these experiments he showed the possibility of the
feat of war attributed to Archimedes.

In 1755 Hoesen, a mechanician of Dresden, began
to build up mirrors of paraboloidal curvature. One
of these was over 3 meters in diameter, and so well
made as to concentrate the sun's rays to a focus 1.3
centimeters in diameter. With one of Hoesen's mir-

363

THE SUN

rors of half this diameter Wolf reduced many metallic
ores, and melted coins almost instantly.

THE " HOT-BOX" PRINCIPLE
De Saussure (1740-1799), the Swiss naturalist,
made five half cubes of glass of such sizes as to go one
within the other with some air space between. These
rested inverted on a blackened table non-conductive
of heat. Thermometers were placed between the
vessels and in the air outside. The one between the
fourth and fifth vessels showed the highest tempera-
ture, 87.5° C. In later experiments with glass-covered
vessels he protected the sides and back of the vessel
from cooling by wrapping it with non-conducting
material. When the vessel was exposed to the sun
perpendicularly he observed on one occasion a tem-
perature of 110° C. within. In one experiment he
heated the surrounding medium, keeping its temper-
ature just below the inside temperature, and thereby
practically prevented loss of heat, except through
the front. In this manner he obtained a temperature
within of 160° C. His experiments convinced him
that two, or at most three, sheets of glass over such a
hot box are better than more. He made some essays
at cooking with such devices.

Sir John Herschel describes the following experi-
ments made during his sojourn at the Cape of Good
Hope, 1834-1838.1

1 "Results of Astronomical Observations ... at the Cape of
Good Hope," etc., by Sir John F. W. Hersrhol, Bart,, published 1847.
Appendix C.

364

UTILIZING SOLAR ENERGY

" (439) When, the heat communicated from the sun
is confined and prevented from escape, and so forced
to accumulate, very high temperatures are attained.
Thus, in a small mahogany box blackened inside,
covered with window glass fitted to size, but without
putty, and simply exposed perpendicularly to the
sun's rays, an inclosed thermometer marked, on Nov.
23, 1837, 149° F. ; on Nov. 24, 146°, 150°, 152°, etc., etc.
When sand was heaped round the box, to cut off the
contact of the cold air, the temperature rose on Dec.
3, 1837, to 177°. And when the same box, with its
enclosed thermometer, was established under an
external frame of wood well sanded up at the sides,
and protected by a sheet of window glass (in addition
to that of the box within), the temperatures attained
on Dec. 3, 1837, were at Ih 30m P.M. (Appar. T.)
207.0°; at Ih 50m, 217.5°; and at 2h 44 m, 218°,
and that with a steady breeze sweeping over the point
of exposure. Again on Dec. 5, under a similar form
of exposure, temperatures were observed at Oh 19m,
of 224°; Oh 29m, 230°; at Ih 15m, 239°; at Ih 57m,
248°; and at 2h 57m, 240.5°. As those temperatures
far surpass that of boiling water, some amusing ex-
periments were made by exposing eggs, fruit, meat,
etc., in the same manner (Dec. 21, 1837, et seq.),
all of which, after a moderate length of exposure,
were found perfectly cooked — -the eggs being ren-
dered hard and powdery to the center; and on one
occasion a very respectable stew of meat and vege-
tables was prepared, and eaten with no small relish

365

THE SUN

by the entertained bystanders. I doubt not, that
multiplying the inclosing vessels, constructing them
of blackened copper inside, insulating them from con-
tact with each other by charcoal supports, surround-
ing the exterior one with cotton, and burying it so
surrounded in dry sand, a temperature approaching
ignition might readily be commanded without the
use of lenses."

MOUCHOT, PlFRE, AND ERICSSON

August Mouchot, of Tours, France, was the great-
est pioneer in the utilization of solar heat. He began
his experiments prior to 1860 and continued them
for about twenty years with aid from the French gov-
ernment. He constructed solar cooking appliances,
and later large machines for pumping water which he
installed in Algeria. Mouchot published in 1869 a
work entitled "La Chaleur Solaire et ses Applica-
tions Industrielles. " A second edition appeared in
1879. He gives a history of the art, describes many
applications of solar heat, and summarizes his own
work, including illustrated descriptions of his great
solar engines, and a report of his mission to Algeria to
install for the Government solar pumping plants in
the desert regions.

Solar heaters after the general form of Mouchot's,
that is to say, with a conical or paraboloidal reflector,
and glass-encased tubular boiler, were also con-
structed after the designs of M. Pifre. One of these
was exhibited at the Tuileries Garden in Paris in

366

UTILIZING SOLAR ENERGY

1882, in combination with a steam engine and print-
ing press, and many copies of a paper called the
"Soleil Journal" were printed by solar power.

In America Captain John Ericsson, the inventor of
the famous " Monitor" type of naval vessels, devised
several solar engines, 1868 to 1886. He used a cylin-
dric mirror of parabolic cross section to concentrate
the rays upon a tube. A two-and-a-half horse-power
engine actuated by one of his solar heaters was ex-
hibited in New York at the American Institute Fairs
for several years.

ENEAS SOLAR ENGINES

Fig. 68 shows the solar machine of A. G. Eneas
(U. S. Patents No. 670,916 and 670,917 of March 26,

FIG. 68. — ENEAS' SOLAR ENGINE.

1901). One of his solar generators was in use for a
time at the Cawston Ostrich Farm near Pasadena,

367

THE SUN

California, and others in Arizona, for pumping water.
The mirror is composed of facets of silvered glass ar-
ranged upon the inner surface of a
hollow truncated cone, whose sides
make an angle of 45° to the axis.
The larger diameter of the cone is
stated as preferably as large as
thirty-two feet, and in several in-
stances was actually thirty-six feet.
Decided advantage is claimed in
leaving the lower end of the mirror
open, as it greatly diminishes the
wind pressure, and the part of the
cone omitted is not very useful for
gathering heat. The mounting shown
in the first patent is neither equa-
torial nor alt-azimuth, but this fea-
ture was improved in the second
by substitution of the equatorial
form. A canvas shield was provided
to protect the instrument from rain.
An interesting feature is the form of
construction of the boiler shown in
Fig. 69. The solar rays are focussed
upon the tube F, and the enlarged
parts,/1 and/2, are respectively above
and below the focal region. The up-
per enlargement is a steam and water drum, the lower
a settling chamber for extracting foreign matter from
the water. Two concentric copper tubes,/ and/8,

368

FIG. 69. — BOILER
OF ENEAS' EN-
GINE.

UTILIZING SOLAR ENERGY

connect the two enlarged chambers, so that the water
falls in / and rises in /8, the latter being of course
the hotter. The tube /8 is enclosed by one or more
glass tubes, fn, f12, whose purpose is to retard the
escape of heat from /8 while admitting the solar rays.

Mr. Eneas has been good enough to furnish me the
following details as to the construction of his ma-
chines and their efficiency in actual operation :

"Feb. 14, 1901. Pasadena, California, llh 30m
to 12h 30 m. Cross-sectional area of sunshine inter-
cepted 642 square feet. Air temperature 61° F.
Steam pressure per square inch 145-151 pounds.
Steam condensed 123 pounds.

"Oct. 3, 1903. Mesa, Arizona. Cross sectional
area of sunshine intercepted 700 square feet. Air
temperature 74° F. Average steam pressure per
square inch 141 pounds. Steam condensed per hour
133 pounds. Time about midday.

"Oct. 9, 1904. Wilcox, Arizona. Time 11 A.M. to
12 M. Cross sectional area of sunshine intercepted
700 square feet. Steam pressure per square inch
148-156 pounds. Steam condensed 144.5 pounds.

"The engines used were of the fore and aft com-
pound condensing marine type, complete with direct
connected air and feed pump. Size 224" X 6" X
41/2" and operated at 460 to 520 revolutions per min-
ute, with about ~ cut off and 25" to 26" vacuum.
ID

The steam used in the engine was superheated about
40° F. in the later machines.

369

THE SUN

"I find 3.71 British Thermal Units per square foot
per minute given as the greatest amount of heat ob-
tainable during the trial runs. The machines re-
sembled in design Patent No. 670,917 with equatorial
mounting. In the 1904 model the greatest and least
diameters of the mirror were 36 and 19 feet, and the
angle of inclination between its axis and its sides 45°.
The mirrors were made of white glass similar to what

Chance Brothers of London make, and were about —

inch thick, 18 inches wide, and 24 inches long, and
were sprung to the curvature of the frame. White
glass was used to reduce the loss from absorption.
The area of sunshine intercepted is the net area after
deducting for shadows caused by the tension rods
and frame work. In the later machines built, the
mirrors were set so as to concentrate the reflected
rays on two parts of the boiler instead of its entire
length as in the Pasadena machine. This change
gave better results" (perhaps because of the better
protection of the remaining parts of the boiler by
non-conducting wrapping instead of glass tubes).
"The total cost of the machine complete with engine
and pump was $2,160.

"An average day's run at Wilcox gave results
about as shown in table on following page. Date
October 14, 1904.

370

UTILIZING SOLAR ENERGY

TIME HOURS
7 A.M. FOCUSSED

Steam Pressure
in Pounds

Inches on
Water Gauge

120

125

136

140

26^

152

30 +

1 P.M.

146

30+

141

30 +

126

" Vacuum 23. Gallons of water pumped, 146,780.
Total lift plus friction 39.4 feet. "

(This test would indicate an average horse power
for the whole day of about 2-g-. From data to be
given below it has been computed that this means the
transformation of about four per cent of the solar
radiation intercepted by the mirror into mechanical
work. From coal the best engines transform from
twelve to fifteen per cent of the heat of combustion
into mechanical work, but probably not in so small a
plant as this. The result of course depends on the
efficiency of the steam engine used, as well as on that
of the boiler.)

Mr. Eneas continues:

" As a result of my experience with about nine dif-
ferent types of large reflectors, I believe: (1) That
with similar mirrors perfected in details about 3.90
British Thermal Units per square foot per minute
would be the greatest amount of heat obtainable at
noontime in Arizona and other cloudless regions of

371

THE SUN

similar latitude. (2) That better progress in utiliz-
ing solar heat commercially for power can be made
along lines described in the Engineering News of May
13, 1909. But the actual obtaining of any great
amount of power from solar rays is still an unsolved
problem. "

If we take the number of calories per square centi-
meter per minute available as 1.4, we find from Mr.
Eneas, figure of 3.71 British Thermal Units per square
foot per minute as the " greatest amount of heat ob-
tainable during the trial runs" that about seventy-
two per cent of the solar radiation was turned into
heat in steam. His estimated maximum possible
number (3.90 B. T. U.) corresponds to seventy-six
per cent. This is really a very satisfactory result.
The maximum steam pressures recorded correspond
to a temperature of about 185° C.

PROPERTIES OF GLASS

The use of one or more glass casings as an adjunct
to the boiler of the Eneas solar engine is quite analo-
gous to the use of glass by de Saussure, Herschel, and
Mouchot, and also to its common use by gardeners
over their hotbeds. Glass transmits radiation very
freely between wave lengths 0.37//, in the ultra-violet
and 2.5/4 in the infra-red. This range, as indicated
by Fig. 26, includes nearly all the solar radiation.
The interposition of a single thin glass plate in a
beam of sunlight diminishes the intensity about fif-
teen per cent. This decrease is owing principally to

UTILIZING SOLAR ENERGY

reflection. The rays emitted by the outside of the
boiler, if we estimate its temperature at 500° abso-
lute Centigrade, would have their wave length of
maximum intensity at about Q/JU and would be almost
wholly prevented from directly escaping as radiation
by the glass. A large fraction would suffer " metallic
reflection" by the glass back to the boiler tube, and
the remainder, being absorbed in the glass itself,
would tend to raise its temperature and that of the
air space, and so to diminish convection from the
boiler to the glass. Furthermore, the glass also pre-
vents wind from blowing on the boiler, and cuts off
all direct convection of heat to the outside air, which
is fully as valuable a function as the restraint of out-
ward radiation. Thus, the employment of the glass
greatly promotes the efficiency of the device, for it
raises decidedly the temperature of the boiler. We
shall notice below the connection between tempera-
ture and the possible thermodynamic efficiency of the
engine.

We have already given the interesting story told
by Sir John Herschel of the dinner he cooked under
glass by solar heat. The late Secretary S. P. Lang-
ley was greatly interested by this story and had more
than one "hot box" constructed on similar princi-
ples. The writer designed one of them. It con-
sisted of two round shallow wooden boxes, the inner
one 50 centimeters in diameter, the outer 60 centi-
meters, placed concentrically one within the other
and each covered by a tightly fitting glass plate. The

373

THE SUN

boxes were further protected by a layer of feathers
about 10 centimeters thick all around the sides and
back of the outer box. The inner one had a black-
ened metal sheet near its bottom, and suspended a
little above this a blackened thermometer. The
whole device was mounted equatorially and kept
toward the sun. On November 4, 1897, at Washing-
ton, operating with three glass plates, the thermom-
eter reached 118° C. while the outside temperature
was 16° C.

The question might be asked whether much higher
temperatures are not practicable to attain in such a
manner without mirrors or lenses to concentrate the
heat. Perhaps with better construction it might be
possible even to reach 200° C. with such contrivances.
The limiting temperature is reached when the solar
heat introduced is balanced by the escape of heat by
conduction through the glasses and through the in-
sulating material at the back. The effective losses
diminish with increasing thickness of insulating ma-
terial, increasing area of the "hot box, " and increas-
ing numbers of glass plates. But, unfortunately, the
increase of the number of glass plates diminishes the
quantity of solar radiation reaching the inner cham-
ber, so that, as found by de Saussure, two or three
glasses give best results. The writer has tested with
the following results the effect of introducing in a
beam of 'sunlight at normal and also 45° incidence
successive plates of the common glass 1.5 to 2.0 milli-
meters thick used for 8X 10 photographic plates, and

UTILIZING SOLAR ENERGY

of plates 8 to 10 millimeters thick used for instrument
covers :

Percentage transmission of glass plates.

Normal

45°

NUMBER OF PLATES

Thin glasses

86.5

74.5

63.5

53.3

85.0

71.8

60.0

49.0

Thick glasses

79.

34.

50.

39.

SOLAR HEATERS AND RESERVOIRS

U. S, Patent No. 230,323 of July 20, 1880, was is-
sued to Messrs. Molera and Cebrian, who proposed to
omit the costly and intricate optical devices for con-
centrating the solar heat as used by Mouchot, Erics-
son, and others, and even the mechanical devices for
presenting the heater broadside toward the sun. They
proposed a horizontal boiler composed either of a
large number of blackened tubes laid side by side,
or a pair of plates enclosing a thin stratum of liquid,
and communicating in either case with a suitable
engine designed for working at low temperatures.
These inventors make no mention of a glass cover for
their boiler, but its introduction would undoubtedly
have increased the efficiency of their apparatus very
greatly.

The erection upon the roof of a building of a series
of water tanks protected by a non-conducting mate-
rial at the back, and by a glass cover above, and com-
municating with the water system of the bath, is
26 375

THE SUN

much used in Southern California, and doubtless
elsewhere, as a means of providing warm water. Such
devices ordinarily furnish a considerable supply of
•water too hot to bear the naked hand in, and save the
discomfort of fire in warm weather.

In all countries the sun is obscured more or less of
the time by clouds and during the night hours, so
that various inventors have proposed the combina-
tion of a device for gathering solar heat and a large
heat reservoir, which usually takes the form of a tank
of water having non-conducting walls, and is situated
above the level of the heater, to which it communi-
cates by pipes. U. S. Patent No. 784,005 of Feb. 28,
1905, to E. C. Ketchum recognizes such features in
combination with a vaporizing chamber situated
within the reservoir, and containing some vaporizable
liquid suitable for running a low temperature engine.
In the event of a very prolonged cloudy spell the in-
ventor proposes also a furnace for heating the reser-
voir independently of the sun.

Low TEMPERATURE SOLAR ENGINES

Within the last ten years, at least two serious at-
tempts have been made to devise commercially eco-
nomical means of employing the hot-box principle for
power. Both series of experiments are described in
the Engineering News for May 13, 1909, referred to
by Mr. Eneas. The inventors are Mr. F. Shuman, of
Philadelphia, and Messrs. H. E. Willsie and J. Boyle,
Jr., of Cranford, N. J. The Shuman heat absorber is a

376

UTILIZING SOLAR ENERGY

level hard rolled plot of ground rendered waterproof
by covering it with asphaltum and enclosed by plank
partitions rising a few inches above the bottom. In
this tank water is filled to a level of about three inches
and over it a thin layer of paraffin, which of course
melts in the sun, and hinders evaporation and radia-
tion from the water surface, while transmitting the
solar rays to the water and asphaltum. The whole
tank is tightly covered with a single layer of glass set
in oiled cotton packing. Wind screens are erected
to protect the tank from convection losses. The cost
of such construction is said not to exceed twenty-five
cents per square foot, and it is expected to produce a
horse power for each 160 square feet (it is not stated
if this is the average of all conditions or only the re-
sult in the most favorable hours, but almost certainly
it is the latter) . The water flows from the heater to a
steam turbine operated in connection with a vacuum
pump. Assuming an initial temperature of 202° F., the
vacuum causes the explosion of perhaps ten per cent
of water into steam and the reduction of the tem-
perature of the mixed steam and water to about
102° F.

As the maximum possible thermodynamic effi-
ciency under such conditions is fifteen per cent, it is
unlikely that as much as five per cent of the sun's
heat can be converted into mechanical work. A
large storage reservoir, built below ground and well
insulated, is connected with the apparatus in such a
manner that the excess of hot water during the hot-

377

THE SUN

test part of the day goes in at the top of the reservoir,
while water from the bottom of the reservoir is with-
drawn to supply that withdrawn from the heating
tank. During the morning and evening hours, or
under cloudy conditions, the motor can be run from
the reservoir. The plant is still in the experimental
stage, but appears to be well planned, and promises
considerable success.

The apparatus of Messrs. Willsie and Boyle has
been more thoroughly tried, so that Mr. Willsie gives
actual figures as to cost and efficiency. They prefer
to build an entirely wooden basin coated with asphalt,
for they find the sand even of the desert to contain
moisture which injures its quality for a non-conductor
of heat. In order to promote a more rapid circula-
tion of the water, and its consequent higher efficiency
to absorb heat, they incline the basin. In their latest
construction the water runs from a first basin with
one glass cover to a second with two, and from this it
drips over a row of pipes containing sulphur dioxide
gas. They employ a low pressure sulphur dioxide
engine of the type developed 'n Germany by Pro-
fessor Josse. In their experiments they run between
temperatures approximating 200° F. and 100° F., but
at midday their heater sometimes reaches near!y 260°
F. They also combine their apparatus with a large
reservoir for use at night or in cloudy weather. Four
installations have been erected by Willsie and Boyle,
the first at the St. Louis Exposition, the others at
Needles, Arizona, a place that all travelers who have

378

UTILIZING SOLAR ENERGY

been there will agree is well qualified for experiments
with solar heat! Mr. Willsie estimates the cost of
installing a sun-power plant at $164 per horse power,
and the cost of operating 400-horse-power steam-
electric and solar-electric plants in desert regions at
2.08 and 0.61 cents per electric horse-power hour,
respectively.

SOLAR COOKING APPLIANCES

Experiments in solar cooking which attracted con-
siderable public attention were made in 1878 by W.
Adams of Bombay, India. Fig. 70 shows the very
simple apparatus employed by him for cooking pur-
poses. The eight-sided conical concentrator was
made of wood lined with silvered glass. It was
hinged upon a board and adjusted by a wedge and by
rotating the board so as to face the sun. The posi-
tion of the apparatus required to be changed about
once each half hour. The cooking vessel of copper
was enclosed in a glass case and fixed to the back of
the concentrator. Mr. Adams wrote to the Scientific
American,1 that the rations of seven soldiers, consist-
ing of meat and vegetables, were thoroughly cooked
by it in a couple of hours, in January, the coldest
month of the year in Bombay ; and that the men de-
clared the food to be cooked much better than in the
ordinary manner. It was also tried with success by
several people in Bombay and in the Deccan. The

1 June 5, 1878.
379

THE SUN

380

UTILIZING SOLAR ENERGY

dish is stewed or baked, according as the steam is re-
tained or allowed to escape. Adams' reflector was
two feet four inches in diameter.

SOLAR METALLURGY

Besides devices for producing power, for cooking,
and for warming water for domestic purposes, by
solar heat, we may note its proposed application for
metallurgy. U. S. Patent No. 277,884 of May 22,
1883, was granted to John Clark of England for a
" Method of Reducing Metals from their Ores. " The
inventor proposes to use a concave mirror built of
segments of silvered glass or burnished metal,
mounted in a manner convenient to face the sun, and
adapted to focus the solar rays upon a stick of ore,
for instance of the oxide or chloride of aluminum or
magnesium, formed into a convenient shape by com-
pression from the powdered substance. He proposes
either to mix solid reducing agents with the ore or
else, when the ore is heated to a suitable temperature,
to convey a gaseous reducing agent, as hydrogen or
carbon monoxide, to the incandescent material. The
excess of the reducing agent is supposed to prevent
reoxidation of the reduced metal, but this may be
further guarded against by enclosing the whole ap-
paratus in a glass roofed chamber filled with a neu-
tral or reducing gas. The advantage claimed for the
proposed use of solar rather than other sources of
heat is the fact that a very high temperature can thus
be readily obtained.

381

THE SUN

RESUME

In the preceding pages we have noted various
devices which, singly or in combination, have been
employed by numerous inventors for the utilization
of solar heat. They comprise first a large surface
for receiving the sun's rays. This may be fixed in a
horizontal or other preferred position, or progress-
ively inclined by suitable mechanism to suit the posi-
tion of the sun. In the former case the surface is
blackened to promote absorption, and the heat thus
derived is communicated to some liquid for domestic
use or for the running of a low temperature heat en-
gine. More commonly mirrors (or sometimes lenses
or prisms) are provided for concentrating the rays to
an approximate focus. Usually the mirror is com-
posed of a large number of facets of plane silvered
glass or burnished metal arranged upon a frame of
suitable general curvature. The form of the reflect-
ing combination may be a paraboloid, or cone of rev-
olution, or an arc of a cylinder of parabolic cross sec-
tion. At the approximate center of concentration
of the rays is located a heater for the ore to be reduced
or the liquid to be vaporized. Advantage is gained in
this case, and also in the fixed forms of solar heater,
by encasing the heated part with glass in the direc-
tion from which come the solar rays, and protecting
it by non-conductors of heat in other directions. The
means of presenting apparatus to the sun usually em-
ployed by astronomers, such as the English type of

382

UTILIZING SOLAR ENERGY

open fork equatorial mounting, which would seem to
be excellently adapted for the purpose, do not appear
to have appealed to the solar engine inventors, as a
rule. They have generally devised more complex
mechanical movements for their purposes, including
circular tracks, slotted hinged uprights, intermediate
types between the alt-azimuth and equatorial forms
of mounting, etc. Excepting the solar heaters for
bath purposes commonly installed in the roofs of
houses, it does not appear that appliances for utilizing
solar heat are yet introduced with economical success
in practice, for although much work has been done in
this line for centuries, we hardly ever see any of the
machines.

We shall conclude this chapter by a consideration
of some of the data to be used in the design of solar
heat apparatus.

QUANTITY OF SOLAR ENERGY AVAILABLE

We may first inquire how much solar radiation is
available. The following data are computed from
the Smithsonian pyrheliometric observations at
Washington and Mount Wilson. Sun rays may be
received on a surface at right angles to the beam
(" normal incidence"), in which case the surface must
be moved by suitable mechanism to follow the ap-
parent motion of the sun in the heavens. On the
other hand, the rays may be received on a fixed hori-
zontal surface, in which case their intensity will
diminish as the cosine of the sun's zenith distance.

383

THE SUN

In either case there is the decrease of the intensity of
the rays depending on the length of path in the at-

f.5
1.0

71 gives the
mean intensity
of direct sun-
shine in calo-
ries per square
j centimeter per
minute for
Mount Wilson
and Washing-
1 ton. Horizon-
tal distances
give ' ' air mass-
es, " or, in other
words, secants
of the zenith
111 distances of the
sun1. Vertical
distances are
calories. One
pair of curves,
III and IV, is
for the receiv-
s, I and II, for

'X.

5* § CALORIES. £

SEC.Z 234

FIG. 71. — INTENSITY OF SUN RAYS. (Mount
AVilson and Washington.)
I, II. Normal incidence. Ill, IV. On
horizontal surface.

g surface horizontal, the other

1 The secant, of (ho zenith distance ceases to represent closely the
"air mass" for zenith distances above 78^9° where sec. 7 - 5. From
some measurements made at very low sun the data given below are
extended to sun rising and setting.

UTILIZING SOLAR ENERGY

" normal incidence." The curves I and III are for
Mount Wilson.

In Fig. 72 (upper half) is shown for sea, level and
6000 feet elevation, both for horizontal and normal
incidence, the march of the sun's direct radiation

L/T.:

30"

FIG. 72. — INTENSITY OF SOLAR RADIATION.

Sea-level and 6,000-feet elevation. Normal incidence and on hori-
zontal surface.

from noon to sunset on December 22, February 17
(and October 25), March 21 (and September 23),
April 22 (and August 22), June 22, at which times
the sun's declination is— 23J^°— 12°, and 0°, + 12°
-h 233/2°, respectively. The data are computed for
latitude 38° N. Horizontal distances give the hours,

385

THE SUN

and vertical distances the calories per square centi-
meter per minute. Similar computations have been
made for latitudes 20° N., 30° N., and 45° N., but
are not shown in Fig. 72. From these results come
the data represented in the lower half of Fig. 72. The
curves show the number of calories of solar heating
per square centimeter per day falling in cloudless
weather on surfaces at horizontal and normal inci-
dence, at sea-level and 6000 feet altitude respectively
for the given latitudes. In each group the two
upper curves are for normal incidence; the highest
for 6000 feet elevation. In the following table is a
summary of the whole matter expressed in calories
per square centimeter per year, and also in square
feet required on the average per horse-power assum-
ing complete absorption and transformation, and the
sun to shine 261,000 minutes per year.

NORMAL INCIDKNCK

HORIZONTAL SURFACK

Sea-level

6,000 feet

Sea-level

<;,<><)() feet

20°

292,000

362,000

185,000

226,000

Calories per sq.

30°

287,000

355,000

170,000

203,000

cm. per year

38°

271,000

342,000

152,000

185,000

45°

270,000

340,000

137,000

169,000

20°

10.5

8.5

16.6

13.6

Average sq. feet

30°

10.7

8.8

18.1

15.1

per horse-

38°

11.3

9.0

20.2

16.6

power

45°

11.4

9.1

22.4

18.2

It is not difficult to absorb ninety-five per cent of
the solar radiation falling upon a surface. Lamp-

386

UTILIZING SOLAR ENERGY

black is employed as an absorber if the temperature
is low, and platinum black electrolytically deposited
if the temperature is so high as to burn off lampblack.
There are many regions of the earth where the days
are seventy-five to ninety per cent cloudless, or even
more. Hence we may conclude that there are many
regions where for the average daylight hours it is
practicable to absorb on a surface of from one to two
square yards solar heat mechanically equivalent to a
horse-power. But in the production of mechanical
power from solar heat only a small percentage is ac-
tually utilized.

THERMODYNAMIC EFFICIENCY

It is shown in works on Thermodynamics that a
perfect engine taking in heat at the absolute temper-
ature TI, and rejecting it at T2, can transform only

rp rp

the fraction - l * of the heat into mechanical
^i

work. For illustration, suppose the engine taking in
heat at the boiling point of water, 373° C. absolute,
and rejecting it at the freezing point, 273°, the maxi-

100

mum efficiency possible will then be — = 26.8 per

o7o

cent. This thermodynamic law gives the efficiency
of a perfect engine, and it does not matter what its
nature, if its actuating energy is heat. A thermo-
electrical engine or a steam engine are both heat en-
gines, and their efficiency cannot exceed that calcu-
lated by the above rule. In fact, however, no heat

387

THE SUN

engine is perfect, and the best triple expansion con-
densing steam engines, with the best constructed
boilers, hardly ever convert as much as fifteen per
cent of the heat of combustion of the coal they con-
sume into work. If a heat engine works from a very

high temperature to a low one, the fraction

•M

may approach nearly to unity. For instance, sup-

rr^ rr\

pose T! = 1000° and T2 = 300° then —L^~ 2 = 70

per cent. This accounts in part for the high effi-
ciency of internal explosion engines, which develop
high temperatures in their cylinders, and often con-
vert twenty-five per cent of the heat of combustion of
their fuel into work. On the other hand, the losses of
heat by conduction, convection, and radiation in-
crease rapidly with rising temperatures, so that if
engines, are used at very high temperatures the ther-
modynamic gain may be counterbalanced by a prac-
tical loss.

REFLECTING POWER OF MIRROR SURFACES

TABLE XXX. — Percentage reflecting power of various surfaces

WAVE LENGTH

0.35M

0.40M

0.45,*

0. 50M

0.60/u

0.70ft.

0.80M

LOOM

Glass1, -silvered on back.

Sample A

Glass, silvered on back,

Sample B

Glass, mercurv back ....

Silvered on glass (Chem.

Dept.)

Nickel (Electrolyt. Dep.)

Speculum metal

1 The reflecting power of mirrors coated on the back differs greatly
with the character of the glass used. Sample A is ordinary optical

388

UTILIZING SOLAR ENERGY

Taking all things into consideration, glass plates
silvered on the back are probably the best materials
for constructing the mirrors of solar heaters.

One may well question whether the solar engine of
the future will have mirrors and driving mechanism.
The "hot box" of de Saussure and Sir John Herschel
as applied by Willsie and Boyle and Shuman is so
cheap that the low efficiency inseparable from its low
working temperature seems not to bar its use com-
mercially. Some gain in efficiency may be made by
installing the fixed heating surface parallel to the
earth's axis instead of horizontal, but perhaps the
increased cost may offset this gain. The efficiency
of the apparatus depends on the excellence of the
glass protection in front. If one could, in addition,
make a vacuum under the glass economically, the
efficiency would be much higher. This device de-
serves much attention.

It seems highly probable that solar cooking uten-
sils, combined with water heaters and heat reservoirs,
and embodying the "fireless cooker" principle, will
come into extensive use. For it is not hard to see
that very inexpensive apparatus may be designed for
combining these utilities, and that housekeepers will
welcome a relief from the hot kitchen conditions of
summer.

flint glass, 12 millimeters thick, silvered on the back by chemical
deposition. Sample B is ordinary commercial plate glass of a
greenish tinge, about 8 millimeters thick, similarly silvered. The
glass of sample B perhaps has an absorption band in the upper infra-
red spectrum.

389

THE SUN

Good designing, avoiding costly and complicated
construction and devices likely to require frequent
attention, combined with a fuller knowledge of the
properties of materials available, and cleverness in
adapting means to promote efficient results, — these if
supported by a moderate outlay of money for experi-
mental work may perhaps soon make the utilization
of solar energy very extensive.

CHAPTER X

THE SUN AMONG THE STARS

Stellar Distances. — Magnitudes. — Sun's Magnitude and Light
Emission. — Solar Motion. — Star Groups. — Double Stars. — • Stel-
lar Masses and Densities. — Mira Ceti and the Sun. — Stellar
Spectra. — The Classification of Stellar Spectra. — Radiation
Distribution. — Evolution of the Solar System. — Stellar Evolu-
tion.

AT first glance the stars appear to be about as much
like the sun as the fireflies of a summer night. It is
only prolonged investigation which has proved that
the sun is merely a star, and by no means the largest
of them; and that if the sun should be removed a
great distance it would appear like one of the stars.
The Copernican view that the sun is the center
about which the earth and planets revolve seemed
satisfactory enough as far as concerned the solar
system, but was for centuries hard of belief as
regards the stars. For it required the assumption
that these were all so distant that the enormous dis-
placement of the earth in space between summer and
winter produced no measureable changes in their ap-
parent relative positions. If the reader will walk a
hundred paces in any direction within a forest, he will
instantly see that the trees change their relative direc-
tions from him, and if he rides in the cars he perceives
27 391

THE SUN

that the foreground of the .landscape appears to re-
volve. Objects at greater and greater distances from
him are seen to be less and less apparently affected
among themselves by his motion. Accordingly, if the
astronomer of a century ago agreed with Copernicus
he had to believe that since the stars do not sensibly
change in their relative positions during the year,
they are so distant from the earth that a displacement
of between one and two hundred million miles in the
position of the earth as viewed from the nearest star
subtended an angle too small for him to measure. In
other words, as he could observe changes as small as a
couple of seconds of arc, he had to believe that the
stars were at any rate all more than 100,000,000 X
100,000 miles away. In our day we know that this
is so, for the distances of some of them have actually
been measured, but a century ago the astronomers
took it on faith, merely because they accepted the
Copernican system.

The first successful measurements of stellar pa al-
laxes (a star's annual parallax is the angle the radius
of the earth's orbit subtends viewed from the star)
were made by Struve, at Dorpat, on Vega, 1835 to
1838, and by Bessel on the star 61 Cygni, 1837 to
1840. The latter faint star was selected on account
of its large proper motion. Bessel's result was 0."35,
and Struve's about one-quarter second. The former
is nearly correct, the latter about twice too large. It
was a great feat to measure such small angles as these.
In modern practice the efforts to measure parallaxes

THE SUN AMONG THE STARS

absolutely has practically been discontinued, and in-
stead relative parallaxes are determined. That is,
instead of measuring, for instance, the apparent ab-
solute change of polar distance of a certain star due to
the earth's revolution around the sun, astronomers
now for the most part determine how much a given
star appears to shift among very faint neighboring
stars owing to the same cause. For it is now as-
sumed that the very faint stars on the average are so
very distant that they have no sensible parallaxes,
or at most a very minute and approximately known
average parallax, which can be applied as a correc-
tion. Stars are generally selected for individual par-
allax measurements because they have relatively
large " proper motions," or progressive apparent
displacement among the stars. This is usually
a safe criterion of comparative nearness, as ap-
pears from our illustration of the railway train
above. In a survey of ninety-two stars pub-
lished a few years ago by Chase of Yale, there
were found the following numbers of stars between
given limits of parallax:

Number of Stars

Limits of Parallax

0".25to 0".20

0".20to 0".15

0".15to O'MO

O'MO to 0".05

0".05to 0".00

0". 00 to -0". 05

-0". 05 to -O'MO

-O'MO to -0". 15

393

THE SUN

The negative parallaxes are of course illusory, and
hence we may suppose that some positive ones are
also erroneous, so that of this lot no more than two-
thirds have measurable parallaxes, and of the whole
lot three-fourths are more distant than 100,000,000 X
2,000,000 miles. It is customary to express such
enormous distances in terms of " light years. " Light
travels in one year about 6,000,000,000,000 miles.
Accordingly, the number given above is about thirty
light years, which represents approximately the dis-
tance of a star whose parallax is one-tenth second.
The vast majority of stars are many times as distant
as this, and the nearest yet found is d Centauri, whose
distance is about four light years.

STELLAR MAGNITUDES. THE SUN'S MAGNITUDE

The relative brightness of the stars is expressed in
" Magnitudes, " a star of the first magnitude giving
about 2.5 times the light of one of the second. On
this scale Polaris is nearly of the second, Aldebaran
nearly of the first, Vega nearly of zero, Sirius — 1.4,
and the sun — 26.5. A change of five magnitudes
makes a change of one hundredfold in the light, so
that the sun gives the earth over 90,000,000,000 times
the light of Aldebaran. If removed to the distance
of Aldebaran, whose annual parallax is O."ll, the sun
would become a star of the fifth magnitude, and ap-
pear only about as bright as the fainter stars among
the six easily seen in the Pleiades. Accordingly, Al-
debaran emits about forty-five times as much light

,",!) I

THE SUN AMONG THE STARS

as the sun. There are some stars, among them Rigel,
Canopus, and Deneb, which are sensibly of zero par-
allax and yet of first magnitude, or brighter, so that
they must emit many thousands, perhaps hundreds
of thousands of times as much light as the sun. On
the other hand, there are many stars whose light
emission is very much less than the sun's, among
them the rapidly moving star whose parallax was
measured by Bessel, sixty-one Cygni. Its light emis-
sion is one-tenth that of the sun.

SOLAR MOTION AMONG THE STARS

As in a forest walk the trees in front seem to sepa-
rate as we approach, and those behind to crowd to-
gether as we recede, so the stars exhibit a tendency to
move from the approximate direction of the constel-
lation Hercules towards the constellation Argo in the
Southern Hemisphere. In consequence of the great
distance of the stars these displacements, called proper
motions, are very slow, not often exceeding 100" per
century, and usually very much less. Nevertheless,
the observations of star places are so exact that the
foci of the motions have been determined with an un-
certainty of only a few degrees. That in the North-
ern Hemisphere lies in R'ght Ascension 270°, Declin-
ation + 30°, in the constellation Hercules about 10°
southwest of the bright star Vega. The cause of the
phenomenon is the motion of the solar system, rela-
tively to the stars in general, toward the position just
defined, called the solar apex. The rate of motion is

395

THE SUN

determined from the apparent proper motions of the
stars of known parallaxes, or from the general spectro-
scopic survey of the stellar motions in the line of sight.

Professor Campbell has been good enough to give
me the following summary of various determinations :

"Sir William Herschel in 1783 deduced from the
proper motions of thirteen stars (all then available)
that the solar system was travelling approximately
toward the star Lambda Herculis in Right Ascension
262°, Declination -f 26° x.

" Many determinations of the goal of the sun's way
were made in the latter half of the nineteenth century,
as the proper motions of stars became known in
greater numbers. Of those based upon the most ex-
tensive lists of proper motions we mention the follow-
ing:

" Newcomb's coordinates for the apex of the sun's
way, deduced from about 3,100 Bradley stars, are
Right Ascension 275°, Declination + 31°L>.

" From 2,640 Bradley stars,, Kapteyn deduced the
position of the apex Right Ascension 274°, Declina-
ation + 29°.53.

" From the proper motions of 5,413 stars Boss has
computed the apex to be at Right Ascension 270°. 5,
Declination + 34°. 3; and his estimate of the velocity
of the solar motion is 24 km. per second.4

1 Philosophical Transactions, vol. xv, page 405, 1783.

2 The Stars, page 91, 1901.

* Astronomische Xticttn'chtcn, vol. rlvi, page 17, 1901.
4 Astronomical Journal, No. 614, 1910.
396

THE SUN AMONG THE STARS

" Several solutions for the elements of the solar
motion have been based upon the observed radial
velocities of stars.

" In 1900 Campbell, from 280 stellar radial veloci-
ties in the northern three-fifths of the sky, obtained
for the position of the apex Right Ascension 277°. 5,
Declination + 20°; and for the speed 19.9 km. per
second.

"*[n 1909 Hough and Halm based a solution upon
about 500 stellar radial velocities. They obtained
for the position of the apex Right Ascension 271°,
Declination + 25°.6; and for the velocity 20.85 km.
per second.

" In 1910 Campbell deduced the elements of the
solar motion from the observed radial velocities of
1034 stars and thirteen nebulae. His position of the
apex was deduced as Right Ascension 272°, Decli-
nation + 27°. 5; and the velocity as 17.8 km. per
second.

" It should be held in mind that the motion of the
solar system is a purely relative term, and in every
case refers to the particular group of stars used as a
basis for the solution. The computer's aim should
always be to have his observational material as ho-
mogeneous and as representative of the entire sidereal
system as possible.

" It appears that an uncertainty of several degrees
exists as to the direction of the solar motion with
reference to the entire sidereal system, and perhaps of
several kilometers as to the speed of this motion.

397

THE SUN

Perhaps the following values are as probable as any
that we can at present assign :

" Apex at Right Ascension 270°, Decimation
+ 30°. Velocity 18 km. per second.

" It should be said, however, that some astrono-
mers would consider the following values as more
probable :

" Right Ascension 272°, Declination +33°;

" Velocity 20 km. per second.

" Kapteyn and Frost have obtained indications that
the speed of the solar system, with reference to stars
of spectral type B, is considerably greater than with
reference to the system as a whole, but the number of
B-type stars employed in the discussion is perhaps
too small to yield results entirely trustworthy.

" Campbell has found that the velocity of the solar
motion, with reference to stars of spectral types B, A,
and F to F4, inclusive, is in essential agreement with
the velocity deduced from stars of spectral types F5
to G, inclusive, K and M.

" It does not clearly appear that the direction and
speed of the solar motion are functions of the dis-
tances of the stars used as a basis for the solutions."

STAR GROUPS

Whether the sun has companion stars in its course
is not known certainly, but there are known groups of
stars which seem to form well-defined systems moving
with a common trend. Such a group is. the Pleiades,
including, besides the six stars easily visible, a much

398

PLATE XIX

THE PLEIADES. (G. W. Ritchey.)

Photographed with the 2-foot reflector of the Yerkes Observatory, 1901,
October 19. Exposure 3^ hours. Cramer Crown plate.

THE SUN AMONG THE STARS

larger number of telescopic stars. Their connection
to form a common system is shown by at least three
kinds of evidence. First, it is highly improbable that
so many stars of that brightness would fall in so small
a region of the sky, if the distribution was purely at
random. Second, excluding a few stars not regarded
as belonging to the system, the stars mentioned have
equal proper motions in the same direction. The
common proper motion is about 6" per century.
Third, there is a nebula, of filmy cloud patch in the
sky, visibly connected with the several stars of the
group and evidently confirming their common con-
nection (see Plate XIX). The Pleiades group, in-
cluding small stars partaking of the common motion,
measures nearly 100' of arc in average diameter. The
parallaxes of the stars are not certainly measurable,
but their distance has been estimated with some
plausibility to be not less, at any rate, than 200 light
years. Hence, the rad us of the system is not less
than three light years, or 18,000,000,000,000 miles,
which is 6,000 times the radius of Neptune's orbit.
If, indeed, the group is actually as small as this, it
would mean perhaps a hundred good sized stars
nearer together than the sun is to its nearest stellar
neighbor.

The curious connection of nebulosity with the
Pleiades is not without its counterpart in many other
regions of the sky, and even our own solar system
seems not to be devoid of it. There is observable on
dark nights, nearly in the plane of the ecliptic, a light

399

THE SUN

not to be regarded as incipient twilight, called the
Zodiacal light when viewed towards the sun, and
called the Gegenschein in the opposite direction. See-
liger has estimated, on reasonable assumptions, that
it is the matter contained in this ring of nebulosity
which causes the outstanding perturbation in the
orbit of Mercury, not to be accounted for by the at-
traction of known planetary masses. It has been
supposed that stars have their origin in nebulae, and
if so the Pleiades stars would seem to be less advanced
in their course of evolution than the sun, but we shall
recur to this.

Whether it is gravitation which controls the motion
of the sun among the stars, and whether such a vast
system as the Pleiades is, like the planetary systems,
in orderly gravitational movement, are questions
which as yet there are no means of fully solving, but
the affirmative is generally believed It has been
computed by Newcomb, however, that there is not
enough matter in the universe to control the motion
of such runaway stars as 1830 Groombridge and Arc-
turus.

DOUBLE STARS

That gravitation is an universal property seems to
be proved by the existence of well-observed elliptical
orbits in the cases of many pairs of double stars.
Since there are less than ten thousand stars to the
sixth magnitude in the whole heavens, the chances are
almost infinitely small that two of them should be
found within 5" of one another on a random distribu-

THE SUN AMONG THE STARS

tion, for there are over 20,000,000,000 squares of 5"
in the whole sky. But in fact there are many pairs
closer than this among the visible stars, so that a
physical connection in most such cases is practically
certain. In some cases, as in that of the very bright
pair of 0.4 and 1.9 magnitude composing a Centauri,
stars separated by a much greater interval (in this
case averaging 17". 1) are proved to be physically
connected because they are observed to go through a
periodic change of position with respect to one an-
other. The orbit of a Centauri is completed in eighty-
one years. By spectroscopic observations of motion
in the line of sight a great number of stars not tel-
escopically resolvable are proved to be physically
connected doubles because of the variable velocity
observed. In some cases of spectroscopic binaries
the companions are indicated by doubling of the
spectrum lines, but quite often one of the objects is
too faint to give a spectrum, and its existence is noted
only because the periodically variable positions of the
lines in the observed spectrum indicate that the star
observed is affected by orbital motion.

STELLAR MASSES AND DENSITIES

The spectroscopic method gives the projection on
the line of sight of the linear velocity of one or of both
components in their orbits. The telescopic method
gives the projection at right angles to the line of sight
of the angular motion of the components. Both
methods give the period of the revolution. When the

401

THE SUN

parallax of the object is known, as in the case of a
Centauri, the projected linear dimensions of the
orbits are easily found. It is possible in the case of
accurately observed telescopic binaries of known par-
allax, or of pairs whose motions have also been ob-
served spectroscopically, to determine the actual
linear dimensions of the two orbits, and (assuming the
law of gravitation) the relative masses of the two
stars. The combined mass, as compared with the
combined mass of the earth and sun, follows easily
from Kepler's third law. For if we regard the mass
of the earth and sun combined as unit mass, the rad-
ius of the earth's orbit as unit distance, and the year
as unit time; then calling the period, total mass, and
mean radius vector of the binary, P, M, and R, re-
spectively, we have, if matter has the same constant
of gravitation everywhere:

iwr R3
M = pi-

Since R and P are both known for a well-determined
orbit, we thus find M, the ratio of the combined mass
of the binary to the combined mass of earth and sun.
In the case of a Centauri the total mass is twice that
of the sun, and the components being approximately
of equal mass, they are singly about of the same mass
as the sun. Their mean distance apart is 23.6 times
the radius of the earth's orbit.

By such processes the combined masses of various
binary stellar systems have been determined. The
resulting masses are sometimes less, sometimes a few

403

THE SUN AMONG THE STARS

fold greater than the sun's. Had they come out of
another order of magnitude entirely, it would have
seemed doubtful if the constant of gravitation has the
same value in other systems than ours. But as
things are, we seem justified in supposing that gravi-
tation is an universal unchanging property of matter.
There is a method proposed by Pickering for find-
ing a relation between the surface brilliancy and den-
sity of the average star of a binary system whose
period and magnitude, on the scale of brightness, are
known. Without going into an explanation of the
matter, which may be found in works on the stars, we
shall be interested in the conclusion, which is that
stars in general give much more light in proportion to
their masses than does the sun. Astronomers gener-
ally incline to believe that the discrepancy indicates
for the stars generally a smaller density than that of
the sun. In a few instances, another line of argument
regarding star densities is possible. There are some
binary systems whose orbits are of such small dimen-
sions, and lie so nearly in a plane with the earth, that
the components regularly eclipse one another, and
the quantity of light of the binary thereby suffers
periodic variability. In such a system the duration
of the eclipse compared with the period of the orbit,
gives a measure of the relative diameters of the stars
relatively to the diameter of their orbits. Proceeding
in this fashion it was shown by Roberts that the aver-
age densities of these variables (called " Algol vari-
ables" after the name of the famous spectroscopic

403

THE SUN

binary star which is the type of the class) is no more
than one-eighth that of the sun. This general con-
clusion was independently confirmed at the same
time by Russell.

We have spoken incidentally of the Algol type of
eclipsing variable stars. Without going far into the
discussion of stellar variability, it will be of deep in-
terest, in view of the somewhat irregular and very
slight variability of the sun, to speak of another kind
of variable stars of which Omicron Ceti (or Mira
Ceti) is the type. This star is sometimes as bright as
the second magnitude, and sometimes as faint as the
ninth or fainter. Accordingly its range is several
thousandfold in brightness. It goes through its
cycle in an average period of about 331.6 days, but is
sometimes thirty or forty days early or late in coming
to a maximum. Its maxima and minima are not
uniformly bright, for sometimes it attains only the
fifth magnitude at maximum, and sometimes it falls
only to the eighth magnitude at minimum. The
time required to rise from minimum to maximum
brightness is only about two-thirds the time required
to fall to a minimum. The shape of the light curve is
variable too, as the maxima continue longer at some
recurrences than at others.

The spectrum of Mira is of the third type, to which
Antares belongs,1 distinguished by the fluted spectra
found to some extent in sun spots. The spectrum
varies as the star's brightness varies, becoming

1 See below.

101

THE SUN AMONG THE STARS

stronger in the violet, and especially in its bright
violet hydrogen lines, in the maximum phases. The
spectrum indicates a high velocity of recession from
the sun (66 kilometers per second), but there is no
evidence from it that Mira has companions.

In many respects Mira's variability suggests the
solar variability associated with sun spots. True, the
fractional change of solar radiation is perhaps not

more than 773773:7: as great as that of Mira, but in the
100,000

existence of a fixed average period subject to large
individual departures, an unequal intensity of max-
ima, an unsymmetrical and variable light curve,
there is a strong similarity to what the sun spot curve
suggests for the sun. In one respect there is a di-
vergence. Mira increases in brightness faster than
it decreases. The change of temperature of the
earth seems to indicate that the sun's radiation is at
a maximum when sunspots are fewest. But the sun
spots decrease to a minimum slower than they in-
crease to a maximum. Still, with so many features
of similarity there can be little doubt that the dis-
covery of the cause and accompanying phenomena
of the sun spot periodicity will indicate the secret
of the Mira type of variables.

STELLAR SPECTRA

Having taken some note of the distances, motions,
brightnesses, masses and densities of the stars as
compared with that of the sun, and having seen that

405

THE SUN

we owe this information to knowledge gained of the
sun itself, and of the solar system, we may now turn
our attention to the spectra of the stars, and see
wherein and how far the sun is a type in that respect.
We have noted that in the solar spectrum dark lines
of the metals are the prevailing feature. Calcium
and hydrogen lines sometimes give bright reversals in
their centers. Helium seldom produces a dark pho-
tospheric line,but in the spectrum of the chromosphere
its bright lines are conspicuous along with those of
hydrogen and calcium. In the spectra of sun spots
the dark lines of metals are still conspicuous, but are
nearly overshadowed in importance by banded spec-
tra of various compounds, and the violet end of the
spectrum is very weak in them compared with the
red, or with the violet of the ordinary photospheric
spectrum.

These various peculiarities of the solar spectra find
counterparts in the stars. There is a large class of
stars whose spectra are hardly to be distinguished,
line for line, from that of the sun. Among the most
exact duplicates is the spectrum of the principal star
of the brilliant binary system Capella. From this
solar type we can pass either way; in the one direc-
tion to stars on which the red predominates, and
banded spectra overshadow the metallic lines, or in
the other direction to blue stars in which lines of
hydrogen or helium are almost the sole features aside
from the continuous spectrum.

By the kindness of Director Campbell of the Lick
406

THE SUN AMONG THE STARS

Observatory and Director Frost of the Yerkes Ob-
servatory, I give here in Plates XXA and B and
XXI a series of spectra illustrative of the gradation
from the so-called helium or Orion stars to the so-
called carbon stars, which lie at opposite ends of the
scale. While considering these, let us note more
closely the diversities of stellar spectra.

The Classification of Stellar Spectra.

Father Secchi in 1867 divided the spectra of stars
into four great classes. Class I comprises the blue
and white stars. In their spectra dark lines of metal
are few and feeble, but the dark hydrogen lines are
well marked. This class is the most numerous, and
includes among other prominent stars, Sirius, Vega,
and Procyon. Class II comprises the yellow stars
whose spectra are filled with metallic lines. This
class includes the sun, also Capella, Arcturus, and
Aldebaran. Class III comprises orange and red
stars whose spectra show, besides many dark metallic
lines found in the stars of the second type, also nu-
merous dark bands or flutings. These consist, like
the terrestrial oxygen bands, of series starting with
well-marked heads and shading off from these to-
wards the red. These flutings are now recognized
to be caused by oxides of titanium and other metals,
and by hydrides as, for example, of calcium. This
class includes Antares and Betelgeuse. Class IV
comprises some deep red stars, whose spectra also
contain bands or flutings, but with the shadings

28 407

THE SUN

toward the violet. These flu tings are attributed to
carbon or its compounds. The stars of Class IV are
all faint. The two brightest are nineteen Piscium
(5.3 mag.) and 152 Schjellerup (5.5 mag.).

The stellar classification of Secchi is still much
used in general descriptions, although more detailed
systems of classification have been lately adopted.
The accompanying Plates XXA and B and XXI
illustrate some of the differences between Secchi's
types. It is however, practically impossible, with-
out having had personal handling of direct spectrum
photographs, to note at a glance the significant vari-
ations in spectra. The spectral types of Secchi
merge, of course, gradually together, so that in some
cases' one would be doubtful in which of two classes
to assign a star.

There are two principal modifications to be made
to Secchi's classification. First, and most impor-
tant/among the blue or white stars occur many whose
spectra are distinguished by the absorption lines of
helium, more than by those of hydrogen. Lines
of oxygen and silicon also sometimes occur in these
helium star spectra, but most metallic lines are
extremely faint or invisible. Helium stars are nu-
merous in the constellation Orion and in the Milky
Way. Secchi's Class I may then be divided into
two principal sub-classes, the helium or Orion stars,
and the hydrogen or Sirian stars. The helium stars
not infrequently show some bright emission lines in
their spectra besides the dark or absorption Jines.

408

a sa

,>«j

a -813133

& ||||

i iiii

I 1111

ejfi

S3.itf

ssss

THE SUN AMONG THE STARS

Such bright lines are generally of hydrogen, but the
helium line D3 is also bright. Vogel's classification
includes a third division of Class I for such bright
line stars.

But there is a class of stars for which Pickering at
one time proposed to add a Class V to Secchi's sys-
tem, whose spectra have as their main characteristics
bright lines or bands in the yellow and blue, some due
to hydrogen, others of unknown origin. The bright
line, or so-called Wolf-Rayet stars, are situated
mostly in the Milky Way or the Magellanic Clouds,
and, except 7 Vela, are faint stars. Some of the
ultra-violet lines bright in the spectra of Wolf-Rayet
stars are also bright lines in the spectra of certain
nebulae.

There has been adopted at the Harvard College
Observatory a more detailed system of stellar classi-
fication than either Secchi's or Vogel's, and which
includes numbered gradations of the lettered main
divisions, so that a very large number of varieties
of spectra may be indicated. A spectrum marked
B3A, or more briefly B3, is one which is estimated to
be three-tenths the way from a typical B star to a
typical A star, and similarly for other combinations.
The following table gives parallel designations of
typical stars under the classifications of Secchi, Vogel,
and Harvard College Observatory. In the Harvard
classification the type Q is reserved for the "new
stars" which have passed their paroxysm of bright-
ness. Oa is the designation of Wolf-Rayet stars.

409

THE SUN

Class 0 has other subscripts b, c, d, and class M has
subscripts b, c, not included in the table.

TABLE XXXI. — Classification of stellar spectra

STAB

Harvard

Vogel

Sccchi

(q Argus, or Carina)
(7 Argus, or Vela)
(29 Canis Majoris)

oQ.

Ib
Ib

Alcyone
Sirius
Altair
Canopus
Procyon
The Sun

Arcturus
AMebaran

(\ Orionis)
(8 Orionis)
fo Tauri)
(a Canis Majoris)
(a Aquilae)
(a Argus, or Carina)
(a Canis Minoris)
Also Capella (a Auriga?)
(K Geminorum)
(a Bootis)
(a Tauri)

Oe5B

B5A
A

ASF
F
F5G
G
G5K
K
Kf>M

Ib
Ib
Ib

I.S-II.
I.S-H.
Ia3-Ha
11.
Ila-HIa
H.-IH.
Ha-IIIa

I
I
I
I
I
I
II
II
II

Betelgeuse

Mira

(a Orionis)
(o Ceti)
(19 Piseium)

III.

IHa
Hlb

III
III
IV

In some stars the spectrum of hydrogen assumes
form which was, to be sure, predicted from the nu-
merical spectrum series relations, but which has never
been experimentally produced in the laboratory. We
cannot yet tell, therefore, what conditions such stars
typify. The striking analogy between the third type
spectra and those of sun spots, taken in connection
with the proved relatively low temperature of sun
spots noted in Chapter IV, indicate clearly a pro-
gression of temperature from stars of type II to those
of type III as well as of spectrum.

SPECTRAL DISTRIBUTION or RADIATION
Wilsing and Scheiner have lately made a long
series of spectral photometric observations on stars

410

THE SUN AMONG THE STARS

of Vogel's various spectral types, to see how the dis-
tribution of energy in their spectra compares ; and with
a view to estimating the temperatures prevailing in the
stars by comparison with the distribution computed
for " black-body" spectra. Their conclusion is that
the temperatures vary regularly with the type among
the stars investigated, from upwards of 10,000° of
the absolute centigrade scale, to below 3,000°. As
there is some question as to the validity of the tem-
peratures deduced, I give here what seems a more
direct and quite as interesting a summary of their
results, namely, the mean spectral distributions for
four average stars of each of the seven spectral
classes investigated. The spectra have been put
equal at wave length 0.448/4.

TABLE XXXII. — Intensities in stellar spectra. (Wilsing and Scheiner.)

TYPE

STARS

INTENSITY

\0?44S

A0?480

A0?584

A0f638

»' (

/3 Can. Min., 12 Can. Yen.
aDelphini, aPegasi

jiooo

836

579

505

Ia2 {

a Androm. , y Coronae
y Ophiuchi, y Lyrae

} 1000

796

625

525

Ia3-IIa.. j

a Trianguli, £ Geminorum
fi Leonis, 8 Aquilae

} 1000

948

902

845

„,. ..{

y Pegasi, TJ Leonis
p Leonis, £ Pegasi

} 1000

887

578

530

Ha [

TJ Bootis, ft Virginis
ju. Herculis, y Cygni

} 1000

998

993

1005

Ila-IIIa. {

a Arietis, o Tauri
& Cancri, /3 Ophiuchi

} 1000

1205

1766

1897

Ilia |

a Orionis, 5 Virginis
\ Serpentis, & Sagittae

} 1000

1368

3296

4406

According to the energy spectrum data given in
Chapter III the sun's spectrum would fall in their

411

THE SUN

class (Ia3-IIa). The results given indicate that the
order of the series of spectra, given by Vogel from an
inspection of the character of the Fraunhofer lines,
has a strong support from the distribution of inten-
sities of the continuous spectra as well. Furthermore,
the order given is the order proper to a series of
spectra from soifrces of successively lower and lower
temperatures.

EVOLUTION OF THE SOLAR SYSTEM

The inquiring mind is ever stimulated by the
query: What means the order of the heavens, and
can we not draw from it a reasonable view of the
evolution of the universe, including the solar system?
The famous Laplace, in 1796, crystallized and ampli-
fied the conceptions which earlier philosophers
had foreshadowed into his famous Nebular Hypoth-
esis of the formation of the solar system. As mod-
ified a half century later by the discovery of the con-
servation of energy, it presumes a gaseous nebula,
larger than Neptune's orbit, in primitive rotation.
By virtue of its immense extent it contains potential
energy of position which is transformed into heat as
the nebula condenses, and thus is supplied the energy
of radiation. The gravitation of the nebula in con-
junction with the occasional collisions of its mole-
cules tended, it is supposed, to produce condensation.
At certain critical times the revolving mass separated
rings, and these by condensation produced the plan-
ets. The planets in condensation likewise threw off

412

THE SUN AMONG THE STARS

rings which formed the moons. In Saturn's case
rings still persist. The view accounts for the pre-
vailing tendency of the planets, their satellites, and
the sun, to rotate in the same direction, and for the
approximately common plane of their orbits and ro-
tations. The exceptions of retrograde motion were
not known in 1796, nor were they discussed, so far
as is known, by Laplace, in his later revisions of his
theory.

According to Chamberlin and Moulton the La-
placian hypothesis, even as modified and clarified by
the work of Helmholtz, Roche, Darwin, and others,
fails conspicuously to account for a number of things.
Principal among these are: A. The considerable
eccentricities of some of the planetary orbits and the
inclinations of their planes among themselves, and
with respect to the sun's equator. B. The neg-
ative rotation of some of the satellites, and the small
periods of revolution of some of them as compared
with the periods of rotation of their primaries. C.
The difficulty of understanding how rings could be
left off in the shrinking of the nebula, whether it
were gaseous or meteoric in structure, and the still
greater difficulty of understanding how a ring, if left
off, could condense into a planet. D. The difficulty
of accounting for the enormous discrepancy between
the present moment of momentum of the system
and that which must apparently have formerly pre-
vailed.

Chamberlin and Moulton have proposed the
413

THE SUN

"Planetesimal Hypothesis" of the evolution of the
solar system. They might start with a spiral nebula.
Since there are millions of these objects in the sky
this basis is justified. But the authors have gone
even further, and suggested that in the course of
ages two stars may approach so near together that
they will mutually raise enormous tides. Tides
occur in pairs at opposite ends of a diameter. Such
tremendous disturbances as thus supposed, together
with the eruptive tendencies due to intense heat,
would perhaps combine to cause many masses of
matter, varying greatly as to quantity, to be pro-
jected from each tidal region. The relative motion
and gravitation of the two stars would tend to change
the motion of projection of the masses into motion
of revolution in orbits about the primaries. When
the action first occurred, the disturbing star being far
off, and the attraction of the erupting star acting
preponderatingly, the orbits of the erupted masses
would be small, and their periods of rotation short.
At closest approach of the disturbing star the con-
trary would prevail. The outcome would be a two-
branched spiral (see Plate XXV), containing many
masses of all sizes revolving in orbits about the parent
star (our sun) . As the inner orbits are of less period
than the outer, the spiral form will become more and
more coiled, and at length cease to present a spiral
appearance.

Mutual attraction and collisions among the numer-
ous masses would lead to the concentration of the

414

THE SUN AMONG THE STARS

lesser masses and particles on the larger ones, or
their revolution about them as satellites. The au-
thors show that collisions tend on the whole to de-
crease the ellipticity of the supposed orbits, so that
the larger planets, on which most collisions have
occurred, would have the most nearly circular orbits.
The lighter gases would be early lost as atmospheres
from the smaller planets and satellites owing to the
consequences of the kinetic theory of gases. But
with increasing size caused by the accretion of par-
ticles in collision, the occluded gases would be forced
out of the interior by growing pressure, and so after
a time atmospheres would be supplied again to plan-
ets of medium size. The large planets would retain
the gases from the start as atmospheres.

The numerous fragments called the asteroids re-
mained almost unaltered from lack of large masses in
their neighborhood to capture them. Their eccen-
tric orbits and high inclinations are evidence of the
comparative rarity of collisions among them. The
retrograde motions and relatively high velocities
occurring among the satellites seem to present no
difficulty in the view of the authors.

The plane of the sun's rotation they believe to
have been modified by the falling back of much
ejected material not forced into clear orbits. Prob-
ably the original plane of rotation was at consider-
able angle to the present, but has been brought
nearer the average plane of the planetary orbits by
such collisions,

415

THE SUN

Perhaps the greatest difficulty of the hypothesis
• is to account for the supply of solar radiation during
the immense period of time that the earth, as shown
by geological evidence, has retained practically its
present dimensions and form, and its present tem-
perature. Moulton assumes, as the only explanation
available, and one which he thinks is also required
by the Laplacian theory, that probably the contrac-
tion theory of the sun's heat accounts for only a
small part of the solar energy. It would seem as if
the Laplacian theory had a great advantage here, for
it presupposes the general extension of the nebula
beyond the orbit of the earth, when the earth began
to form. Hence there was an immense store of en-
ergy to be gained by contraction. On the contrary,
the spiral nebula of Chamberlin and Moulton appar-
ently had no such general extension, but retained
nearly all of its matter from and after the catastrophe
in the center of things. Furthermore, the general
extension of the nebula of Laplace enables us to sup-
pose that the earth was for a very long time receiving
radiation from a large portion of a hemisphere, or
even (by reflection within vestiges of the nebula)
from a sphere, so that we need not suppose that the
intensity of this radiation was great, and therefore
we can assign a very much longer life to the contrac-
tion source of energy than we could if we were obliged
to think of the solar radiation as always requiring
to be at its present intensity, during geologic time,
in order to maintain terrestrial temperature.

in;

THE SUN AMONG THE STARS

Prof. T. J. J. See has just published (after the above
resume of nebular hypotheses was written) his vol-
ume on this subject. In his view the sufficiently
close approach of two stars to form a spiral nebula,
as assumed by Chamberlin and Moulton, is too in-
frequent to deserve consideration. He would assume
the spiral nebulae to be formed by the close approach
of two nebulous streams, and the curling of them
together by mutual gravitation, or by the curling up
of a single nebulous stream owing to its own grav-
itation, but he does not show that such phenomena
are apt to happen more frequently than that sug-
gested by Chamberlin.

Such a spiral nebula is, according to him, the parent
of the solar system, but unlike Chamberlin and
Moulton, his nebula would not have its central con-
densation, the sun, mainly formed before the planets
began to form, but all would be forming at the same
time, by capture of particles by larger masses in the
exercise of mutual gravitation, and in the vicissitudes
of mutual encounter between the larger and smaller
bodies of the nebula. There seems to be much in
common between this " capture theory" and Moulton
and Chamberlin's accretion theories. See finds that
the orbits of the planets will be rounded up by the
resistance (that is, the continual encounter with par-
ticles) which they find in the nebulous medium. Here
he is in close accord with Chamberlin and Moulton,
who have found, as stated above, "that collisions tend
on the whole to decrease the ellipticity of the supposed

417

THE SUN

orbits, so that the larger planets, on which most col-
lisions have occurred, would have the most nearly
circular orbits." But See believes further that the
present orbits of the planets are very far within the
orbits they had when they were principally formed.
It would seem on the whole, that, excepting in the
method of forming his nebula, Professor See's views
follow the general lines laid down by Chamberliri
and Moulton, but with, this difference that they
allow the sun to be forming at the same tune as the
planets, and in a very extended space, so that the
problem of supplying energy by contraction for con-
tinuing the solar radiation in ample measure through-
out geological time is easier for See than for Chamber-
lin and Moulton. See's conception also permits us
to suppose the solar part of the nebula was so much
expanded as to shine upon the earth from a large
angle in the earlier geological epochs, as was required
for the foundation of what we have termed " Hypo-
thesis (B)" in Chapters VI and VII.

STELLAR EVOLUTION

We will now consider a little more closely the gen-
eral view that nebulae are stars in the making, and
that the stars progress through a series of tempera-
tures, and at length, like the earth and moon, reach a
cold final condition. Plates XXII to XXVI give a
series1 of nebulous forms ranging from the chaotic

1 It is very questionable if we should interpret this scries of forms
as implying a series in order of development. I am greatly indebted
to my friend, Mr. G. W. Ritchey, for this fine group of photographs.

418

PLATE XXII.

THE GREAT NEBULA IN ORION. (G. W. Ritchey.)

Photographed with the 2-foot reflector of the Yerkes' Observatory,
1901, October 19. Exposure 1 hour. Cramer Crown plate.

THE SUN AMONG THE STARS

nebulae in Orion and Cygnus to trhe well-developed
spiral and ring forms of Andromeda and Lyra.
The number of nebulae observable with the Mount
Wilson reflector probably reaches into the millions.
In Plate XIX we saw that the Pleiades stars are
plainly wrapped in nebulosity, and seem as if still in
process of condensation. This peculiarity is shared
by other star groups, notably by some in Orion.
The spectra of the Orion stars have a simplicity far
more in common with the simple spectra of gaseous
nebulae than with the lined and banded spectra of
the solar and Antarian stars. Stars of the Orion
type have in many instances nebulous appendages,
and besides seem to be of extremely small density,
according to the tests we have noted above. Hence,
it is supposed that the first evolutionary step is the
passage from a nebula to a helium star. Nevertheless,
it is found that the great spiral Andromeda nebula
gives at its center an essentially solar type of spec-
trum.

But even admitting the connection of cloudlike
nebulae and helium stars, why should we believe that
the nebula is the first and not the last end of the
chain, in point of time, or that the other types of
spectrum have the same order in their secular de-
velopment as they do in our arrangement of them
according to their physical appearance? As to the
first branch of the question, we know that gravita-
tion tends to condense matter, whether by capture as
of the meteors by the earth, by the opportunities

419

THE SUN

offered in molecular collisions, as required by La-
place's Nebular Hypothesis, or by the collision of
meteors in orbits, as proposed by Chamberlin and by
See. In any of these cases the centrally directed
force of gravitation inevitably seizes its opportunity
to draw in the retarded particle. Excepting the dis-
ruptive tendency of the close approach of two stars
invoked by Chamberlin and Moulton, and the escape
of gases by molecular activity according to John-
stone Stoney, there is not known any cause for the
separation of the constituents of a star into a nebula.
This would require an enormous expenditure of en-
ergy, whose possible source, except as just indicated,
it is hard to conceive. The probability of the close
approach of two stars would seem at first sight to be
very small, for Newcomb has computed that on the
average a sphere of radius 412,500 times the radius
of the earth's orbit contains but one visible star. On
the other hand, there may be enormous numbers of
invisible bodies, and even the number of stars in
space is so large that such near collisions may actually
occur rather frequently, measuring time by centuries.
We shall recur to the question of the order of events
in stellar evolution.

Admitting the view that nebulae generally tend to
condense, not to expand, their rise of temperature
with condensation, if gaseous, was proved by Lane in
1876. If we adopt the usual view that yellow stars
are more advanced than the blue ones, how are we
to explain the circumstance that the blue stars, which

12Q

PLATE XXIII.

NEBULA N. G. C. 6992 CYGNI. (G. W. Ritchey.)
Photographed with the 2-foot reflector of the Yerkes' Observatory,'
1901, October 5. Exposure 3 hours. Cramer Crown plate.

THE SUN AMONG THE STARS

seem to be nearest the nebulae in type, are by Wilsing
and Schemer's observations, and by an appeal to
ordinary experience, apparently hotter than the
yellow ones? As a reply to this objection we must
note that whether with most astronomers we accept
a photosphere, or assume a purely gaseous sun as
discussed in Chapter VI, the inner parts of the sun,
or of a star, are not visible to the observer. The inner
parts may in fact be hotter for yellow than for blue
stars, without in any way altering the succession of
apparent surface temperatures found by Wilsing and
Scheiner, for different type stars.

But it is by no means clear that a yellow star is
necessarily older than a blue star in actual time, and
indeed it does not seem necessary to admit that
every star of the helium or hydrogen type of spec-
trum will necessarily, with lapse of time, become a
solar or Antarian star. The similarity of spectrum
lines proves that certain elements found in the earth
exist in the sun and in the stars. When stars fail to
exhibit any of the spectral lines of an element we can-
not know that this element exists in those particular
stars, for we are not fully justified in supposing that
it does so on the assumption that conditions do not
favor the production of its spectrum. It may pos-
sibly be, then, that Sirius, for instance, will never
show a solar type of spectrum, however cold it may
grow superficially.

I develop this line of thought for consideration in
connection with the discussion and catalogue of

421

THE SUN

spectroscopic binaries recently published by Prof.
W. W. Campbell.1 The observed spectroscopic
binary stars range in period from less than a day to
more than a year, and visual binaries carry the range
of periodic times up to thousands of years at least.
Campbell, in summarizing the existing observations
of spectroscopic binaries, draws attention to certain
relations between the periods of orbital revolution,
eccentricities of orbits, and types of spectrum. The
following table shows these results:

TABLE XXXIII. — Spectroscopic binaries. Spectral types, periods, and

eccentricities*

Periods •*

" short "
8

short

Od-6d
16

2*4

(10)0.04

5d-10d
10

6'|9
(5)0.10

10d

73d2

(11)0.34

yean
1

1.9
(1)0.0

"long"
long

Oand BTyjkM
Mean Period

Mean Eccentricity

A Types
Mean Period.

short

2(.1r>5

1
0*2

12
42d2

2
26.45

Mean Eccentricity

(5)0.04

(1)0.50

(S) 0. f).-,

(1)0.59

F Types
Mean Period

G
3d!

( l) o.o.-)

5dO

(1)0.01

4
145dl
(3)0.15

11.1
(3)0.44

1
long

Mean Eccentricity

G to M Types
Mean Period

3
104d8

9
24.3

13
long

Moan Eccentricity

(2)0.06

(8)0.38

Total
Mean Period

12
short

2d-.!»

6d90

7 3d.->

15
20.5

14
long

Mean Eccentricity

(19)0.04

(7)0.14

(24) 0 . 30

(13)0.38

* From Lick Observatory Bulletin No. 181.

This summary shows clearly that the "earlier"
types of spectra are associated in spectroscopic binary

Observatory Bulletin No. 181. Also "Pub. Astr. Soc.
Pacific," April, 1910.

422

PLATE XXIV.

THE GREAT NEBULA IN ANDROMEDA. (G. W. Ritchey.)

Photographed with the 2-foot reflector of the Yerkes' Observatory. 1901,

September 18. Exposure 4 hours. Cramer Crown plate.

THE SUN AMONG THE STARS

stars, as a rule, with shorter periods arid smaller ec-
centricities than are the " later" types of spectra.
Campbell also gives a table of fifty telescopic binaries
arranged in order of their periods in five groups of ten
each. The periods range from 5.7 years to 194.0
years, and not one of the stars named (which gener-
ally is the principal star of the pair) has a spectrum
of the 0 or B type, while many specimens of types
A, F, and G, and some of K, are found. As for the
eccentricities, these are all large, averaging 0.461,
0.453, 0.495, 0.531, and 0.483 in the five groups.
The general average period is seventy-two years, and
average eccentricity of orbit 0.49.

In summary for the telescopic binaries:

SPECTRAL
TYPE

No. of
Stars

No. of

Stars

Mean
period

Mean
eccentricity

O-B. .

Short periods. . . .

32.8

0.48

G-K..

18
14

Long periods ....

108.1

0.51

M-N
Unknown. . .

0
9

In the words of Campbell: " Visual double stars
clearly abhor the O and B types, and visual double
stars of relatively short periods clearly abhor M and
N types.

"What, " says Campbell, "is the significance of
these facts? Let us recall that Darwin and Poincare
studied the origin of binary stars from theoretical
considerations, and came to the conclusion that a

29 423

THE SUN

condensing nebulous mass, rotating on its axis con-
stantly faster and faster, to keep pace with loss of
heat by radiation, should eventually separate into
two nebulous masses revolving around their mutual
center of mass. These two masses would, in the
beginning, be revolving in contact in orbits essen-
tially circular. With advancing time, tidal disturb-
ances should cause the two bodies to draw apart rap-
idly at first, and less rapidly later. In the spectro-
scopic binary systems described — have we not a tol-
erably complete sequence of orbits illustrative of the
Darwin-Poincare hypothesis? The short-period or-
bits should be circular or nearly so, and should
appertain preferentially to stars of early spectral
types ; the longer periods should, in general, attach
to the more eccentric orbits and the older spectral
types; and these are the facts. established by actual
observation of binary systems. ... It will be noted
that in these widely separated (telescopic binary) sys-
tems there is not a single 0 or B type, representing the
early stages of binary existence. There are a few A
types, but the major number are of the advanced F
type and G and K types. I suspect the K, M, and N
types are not more fully represented for the reason
that in these old-age systems the two components
are in general so far apart that the periods of revolu-
tion are many hundreds or thousands of years."

Campbell considers also the relative masses of the
two components in the binary systems for which this
is known. In seventeen cases where the components

PLATE XXV

SPIRAL NEBULA M. 51 CANUM VENATICORUM. (G. W. Ritchey.)

Photographed with the 5-foot reflector of the Mount Wilson Solar

Observatory. Exposure 10% hours. Seed 23 plate.

THE SUN AMONG THE STARS

are of unequal masses he finds that with one excep-
tion the lesser component is of an " earlier type of
spectrum, or bluer, than the more massive one."
Here, at first sight, is a most surprising thing. Of two
stars admittedly of equal age, the one of greater mass
is in general the more advanced. Campbell says:
"An hypothesis of Huggins, suggested at first
rather casually, and later discussed more seriously,
appears to me to be of great merit, especially when
Schuster's extension of the hypothesis is applied.
Huggins' suggestion is as follows: 'Another way of
looking at the problem is perhaps possible. May it
be that the effect of the great mass on surface density,
together with the working of Lane's law, by which the
temperature of a condensing gaseous mass so long as
it is subject to the laws of a purely gaseous body will
continue to rise, will favor in such stars the coming
in of a solar type of spectrum at a somewhat rela-
tively earlier time?' Schuster's extension suggests
in effect that the lighter gases — hydrogen, helium,
and so on — which surround a star in its early age, will
be pulled down on a star of small mass but lightly,
and a long period will be required for the absorption
of these gases. Such a star would remain effectively
young, as judged by its spectral type, longer than its
more massive primary. In the latter, the greater
gravitational power would lead to more rapid absorp-
tion of the lighter surrounding gases, and the pre-
dominant influence of the metallic absorption would
enter earlier. It seems reasonable to suppose that

425

THE SUN

the greater internal gravitation of the more massive
primary will generate heat more rapidly, and cause it
to live its life more rapidly than in the case of the less
massive secondary."

Should not this explanation of the prevailing ten-
dency to earlier types of spectra for the lesser com-
ponents of binary stars take more explicitly into con-
sideration a difficulty suggested by ordinary experi-
ence of cooling bodies, Planck's law of spectral energy
distribution, and Wilsing and Schemer's results on
the relative temperatures of the stars? For it is the
blue stars which we should suppose to be, and which
Wilsing and Scheiner find to be, superficially hottest,
and as the bluer stars are generally supposed to be
also of less density than the yellow ones, their sur-
faces are also greater in proportion to their masses.
Hence, if their radiating coefficients are equal to those
of yellow stars, they should radiate more rapidly and
advance more rapidly in spectral type thereby, if, as
is often assumed, advance in spectral type is a mere
function of radiation and consequent condensation.

I venture to suggest that if the view of Campbell
as to relative masses and types of spectra is well
founded,1 its significance in this respect may be the
following. As we do not know that the two compon-
ents of a binary are of similar constitution, may it
not be that the bluer component has a smaller coeffi-

1 Not all aslronomsrs are agreed that it is the general rule for the
smaller component of a binary to be less advanced in spectral type,
but Campbell's review of the evidence seems very convincing.

420

PLATE XXVI.

FIG. 1. NEBULA H. V. 24. (G. W. Ritchey.)

FIG. 2. RING NEBULA IN LYRA. (G. W. Ritchey.)

Photographed with 5-foot reflector of the Mount Wilson Solar Ob-
servatory. Exposures: Fig. 1, 5 hours, Seed 23 plate. 1910, March
6. Fig. 2, 45m Seed process plate. 1910, July 1.

THE SUN AMONG THE STARS

dent of radiation than the other? By this I mean that
if the two objects were of equal size and temperature,
the blue star would emit less radiation. We have no
evidence whether this state of affairs occurs for bodies
of stellar temperatures, but decided differences of
radiating power were found by Paschen and others
at moderately high temperatures for various solid
substances, some of which might even have been
expected to be approximately "black bodies." As-
suming this explanation, the blue stars might radiate
slower, even though of decidedly higher temperatures
and larger surfaces in proportion to their masses than
their yellower neighbors.

As to the assumed difference of constitution of the
two components, Campbell has suggested that they
were originally one object, which separated owing to
too rapid rotation. In such a case might not the
smaller object usually carry with it a preponderance
of the lighter elements which composed the original
star or nebula? We have seen in Chapter VI that
the lightest elements lie furthest out in the sun, and
it seems reasonable to suppose that the same holds
in the case of a just separating binary, so that per-
haps they might tend to accumulate in the bulging-
out component of smaller mass. If this is so, then
the presence of a chromospheric type rather than a
photospheric type of spectrum or, in other words, the
assuming of the spectrum of early stellar type, should
naturally be associated with the lesser component,
because it has preponderatingly the light elements,

427

THE SUN

hydrogen, helium, etc., rather than the heavy metals
whose lines throng the solar spectrum. I do not
mean by this to say that the lesser star has none of
the heavy elements, and the greater star all of them.
Rather that the lesser star has so large a supply of the
lighter elements that they effectually screen by their
scattering of light l the radiation of the heavier ones
lower down, just as it is probable that the elements
of the platinum group are obscured in the sun as ex-
plained in Chapter VI. In the greater star the ele-
ments hydrogen, helium, etc., while present, I sup-
pose to be less plentiful, so that the heavier metals
occupy practically a surface position, and, hence,
give their typical spectra.

But it will be urged that this view implies too
much, and does not take into account the progressive
change of spectral type shown to occur with increas-
ing age of binaries. In other words, that it would
imply that, once a blue star, never a solar star.
Before answering this objection let us examine Table
I of this book, which shows that the four outer, and
according to the Laplacian hypothesis probably old-
est, planets of the solar system are all of low density,
even lower (notwithstanding their probably low tem-
peratures) than that of the enormously hot sun, and
four times as low as the densities of the four inner
planets. May not some support of the view just
advanced be gained from this circumstance? Were
not these planets constructed from the solar nebula,

1 See Chapter VI.
428

THE SUN AMONG THE STARS

accumulating with them a preponderance of the
lighter solar constituents? If so, did not their forma-
tion tend to advance the type of the solar spectrum?
May not other stars, binary stars not excluded,
whether primaries or secondaries, also give rise to
relatively small planets and satellites, thereby losing
their lighter surface materials, and hence advancing
themselves in spectral type? Such planets, if of
small masses relatively to their primaries, could not
be observed, and hence may, for all we know, exist.

Dr. Johnstone Stoney showed many years ago that
the lighter elements gradually escape from atmos-
pheres according to the kinetic theory of gases, and
the more rapidly the higher the prevailing tempera-
tures. He explained by this means the absence of
water vapor from Mars, -of all gases from the moon,
and of helium and hydrogen in marked quantity from
the earth. May not this process, favored as it must
be by the high temperatures of the blue stars, aid in
course of ages to divest them of hydrogen, helium,
etc., and thereby tend to advance their type?

Without wholly accepting until we have fuller
evidence the relation pointed out by Campbell in
regard to spectral types and masses of binaries, the
suggestion just made as to a possible path of stellar
evolution is, of course, not limited in its application to
the cases of binary stars. It may be that all the blue
stars are of early spectral type because their elements
of high atomic weight are obscured by the lighter
gases, hydrogen, helium, etc., and that with the es-

429

THE SUN

cape of these gases to space or the formation of satel-
lites, the spectral type will be advanced to the solar
stage, and from that by cooling to the Antarian.1

Returning from these perhaps too presumptuous
digressions, I shall finally call attention to some data
noted by Prof. Kapteyn as perhaps " valuable in the
classification of the stars in the order of their evolu-
tion. "2I He remarks first the progressive increase of
" peculiar" stellar velocities3 for stars of the advan-
cing spectral types. In the following little table he
sums up the results thus far available.

TABLE XXXIV. — Spectral types and velocities in space

Type of spectrum or object

Peculiar radial
velocity per second

Number

B to B 9

km
6 5

A to A 5

12.6(11.2)

Fto F..

14 5

G to G 5

12 6

KtoKS

15.4

19.3

Planetary nebula

''6 S

Orion nebula

0 1

13 1

:') 7

1 Consult, in this connection, T. J. J. See's "Researches," vol. ii,
p. 589.

1 Contributions, Mount Wilson Solar Observatory, No. 45.

* By this is meant "the velocity freed from that part which is due
to the motion of the solar system through space."

430

THE SUN AMONG THE STARS

He then discusses the " peculiar" proper motions
of over 2,000 stars and finds the ratio of their average
magnitudes to the solar motion as follows :

Ratio Number

Type I (B, A) 1 .02 1144

Typell(FtoK) 1.46 1093

Type unknown 1 . 45 381

He concludes that the ratio

Average linear velocity of the F, G, K stars

Average linear velocity of the A stars
cannot be smaller than 1.3.

It is greatly to be regretted that there are not more
radial velocities for nebulae of the Orion type known
as yet, but the evidence seem's to indicate that such a
nebula is to be regarded as practically stationary, and
that the stars of advancing spectral types are affected
by progressively greater and greater motions in space,
but that the planetary nebulae are not to be classed
with the nebulae of Orion type, but at the other end
of the chain.

Kapteyn says:1 "The phenomenon of the increase
of velocity with the evolutional stage of the stars
must give rise to speculation as to its cause. The
observational results contained in our table naturally
lead us to conclude that the matter from which the
stars originate must have little or no velocity. How
is this possible under the influence of the combined
attraction of the rest of the system? Is it not as if2

1 Contributions of the Mount Wilson Solar Observatory No. 45.

2 " We need not necessarily make the hypothesis that really primor-
dial matter is not subject to gravitation. If, for instance, as was
suggested to me by a friend, the tenuity of this matter were such that

431

THK SUN

gravitation had no effect on the cosmical matter in its
primordial state? If this be so — as soon as matter
changes from this state to another in which gravity
begins to act, or to act freely, motion will arise, and it
is evident that, as a rule, the motion must be acceler-
ated, at least during immense periods, so that the
longer the period elapsed since the birth of the stars
the greater must be their average velocity."

After calling attention to the argument from binary
systems, which we have already considered, Kapteyn
mentions the two great star streams which, according
to his researches, embrace the stars, and notes that,
as shown by Dyson, the stars of type I " diverge less
from the general drift of the two streams than the
other stars. " Such a result harmonizes with the view
that the Orion stars are relatively young. But, says
Kapteyn: "Not only this. Observation shows fur-
ther that for the Orion stars the stream velocity is
small ... as compared with . . . the rest of the stars.
Apart from the advantages that we may derive from
this result for the classification of the stars in the
order of their evolution, it has, I think, a great im-
portance in its bearing upon the question of the gen-
it were very materially hindered in its motion by the matter which we
must assume as filling the universe in order to explain the phenom-
enon of selective absorption of light recently found, the velocity
of this matter could not exceed the value for which the resistance is
equal to the total attraction. . . . Other suppositions may probably
be made of forces which, in the primordial state of matter, counter-
art gravity. But it is evident that in such cases where gravity is
just counterbalanced by another force, things happen r/.s /'/ there wen-
no force at all." (May not light pressure be such a force?)

432

THE SUN AMONG THE STARS

eration of the star streams themselves. For it proves
that the streaming motion, too, is not an initial
motion, but one generated at an epoch which, for the
stars of any one type, must be placed at a time rel-
atively but little preceding the time when they
passed through the Orion-type stage."

The results of Kapteyn aid greatly to convince us
that the progress of evolution is from the Orion type
of nebula at the beginning, to the fourth type star at
the end, and not the opposite, in the lapse of time.
For we shall see from them that there is a real prog-
ress from one stage to another, marked by the gradual
march of velocities. It only remains to show that
the march is in the supposed direction and not its
opposite, and for this purpose the study of one part
of the course is as good as another. Now we know
that the stars of the second type resemble the sun's
photosphere, and those of the third type the sun
spots in their spectra, and that this difference is
brought about in the sun by the mere reduction of
temperature. A reduction of temperature, however,
must finally occur when a star exhausts its sources of
energy. Hence, the third type stars must probably
be a later stage of evolution than the second, and the
progress of evolution is therefore from the Orion
nebula to the fourth type star. This conclusion is
supported also by Campbell's discussion of binary
stars. »

Various considerations, then, recommend the view
that the stars are formed from nebulae, take first the

433

THE SUN

Orion type, and pass on with age to the solar class,
and thence, with cooling, to the Antarian stage anal-
ogous with sun spots in spectrum. We may suppose
that a still more advanced, and usually final stage, is
the cold one of which the earth and moon are types.
It is speculating far from the sure ground of observa-
tion to say it, but do not these conclusions, and es-
pecially Kapteyn's discussion of the velocities of
nebulae and stars, indicate that the entire stellar
system arose from a sort of formless, relatively
motionless, chaos, and will at length reach a dark
and unknown end?

CONCLUSION

In our study of the sun and its relations with the
earth and stars, the discoveries of two types of inves-
tigators have come prominently before us. To one
class belong those geniuses whose roving minds in-
cline them to try this and that new thing, and whose
acute perceptions enable them to turn even their most
random observations into glorious discoveries. In-
vestigation to them is like happy exciting play to a
child. These have their place and their ever-present
reward. To another class belong the patient ob-
servers and philosophers who, from a love of science
and a sense of duty to their age and to posterity,
have gradually enlarged by tedious observation and
laborious analysis that precious store of exact knowl-
edge whose value time cannot impair but can only
enhance. The men of both classes are deserving of
admiration, the former for their brilliancy, the latter
for their perseverance. As those of the former class
are continually receiving their meed of praise from
their contemporaries, it will not be amiss to offer our
tribute to the others, and recall to mind the work of
Newton, whose immortal "Principia" he suffered to
remain unknown until by the importunity and finan-
cial means of his friend Halley it came at last to pub-
lication; Laplace, Gauss, Hansen, Newcomb, and

435

THE SUN

many more who erected the wonderful edifice of
mathematical astronomy on the foundation of New-
ton's law of gravitation; the long series of observers
from Galileo down, whose sun spot records were com-
bined by Wolf with such rare skill, after the patient
work of twenty years by Schwabe had indicated the
sun-spot cycle; Carrington, Spoerer, and the others
whose numerous observations revealed the law of
rotation of the sun; Kirchhoff, A. Angstrom, and our
own wonderful Rowland, whose spectrum researches
are the foundation of solar physics; the unremem-
bered army of meteorological observers whose plod-
ding records are sometimes scoffed at by the more
brilliant, but which nevertheless share in the en-
hancement of value produced by generous Time;
Bradley, the father of exact stellar observation, whose
thousands of accurate star places are priceless to
modern astronomy; Argelander, whose enormous
work, the " Durchmusterung " of the northern stars
is yet in the prime of its usefulness; Huggins, whose
pioneer, yet long sustained, investigations in as-
tronomical spectroscopy laid the foundation of the
study of stellar evolution.

These men and many more who were actuated by
the same motives have passed on, but their work still
lives. There still remains, and ever will remain in
solar and stellar investigation, room for such work;
and on the thorough doing of it in our time the won-
derful flowers of future discovery, whose beauty our
eyes cannot see, or our imaginations picture, must

436

CONCLUSION

largely depend. If we now had such long, unbroken,
and accurate series of meteorological records of nu-
merous stations in all parts of the world, on land and
sea and in the air, as posterity must depend on us to
supply; if we now had those long-kept, numerous,
and accurate observations of stellar parallaxes,
brightness, forms of spectra, velocities, and other
data which Kapteyn longs for, but in vain; if we now
had accurate measurements of the solar constant of
radiation going back centuries ; in short, if we could,
as we find the need of it, consult the records of the
Past to verify the surmises of the Present, then solar
and stellar knowledge would advance with such leaps
and bounds that we could soon see the great pano-
rama of the universal evolution unroll before us.

The child is said to long to grasp the moon. Who,
in his maturer years, has never wished that he might
stand upon the moon, and watch the earth at full, a
glorious planet of the night, four times as far from
rim to rim, and twice as bright in every part as is the
moon herself! Who, thinking more gravely, has not
wished sometimes he had been born in later years,
when he could share the fuller understanding yet to
come? Shall we not live in hope that if we worthily
contribute to that happy end, we, too, may join with
that great company whose patient and sound labors
have given us what we know, and in a future life with
them may see unrolled the wider view which here we
long to see in vain?

INDEX

Abbot, C. G., 108, 190.
pyrheliometer, 79.
solar theory, 236-279.
Mrs. C. G., 132.
Aberration of light, 22.
Absorption, and radiation, G4.

atmospheric, 291, 308.
Adams, W., 379.

W. S., 25, 100-104, 124-127,
180-182, 205-209, 213,
255-258, 268.
Air mass, 286-290.
Anderson, 233.
Angstrom, A, 436.

K., pyrheliometer, 77.
Apex of solar motion, 395.
Arago, 142.
Archimedes, 363.
Arctowski, 190, 317.
Argelander, 436.
Armstrong electrical machine,

213.
Assimilation of carbon by

plants, 335-344.
Atmosphere, earth's, extinction,

74, 294-297.
height, 74.
spectrum, 88.

Atomic weights and spectrum
intensities, 91, 93, 169,
253, 317.
Aurora and sun spots, 189, 220.

Ball, R., 222.

Balmer series, 170, 173.

Becker, G. F., 273, 278.

L., 132, 133.
Belopolsky, 135.
Bessel, 392.

Bigelow, 111, 135, 190, 317.
Body, "black," 64, 65.
Bolograph, 83, 292.
Bolometer, 81, 84, 86.
Boss, 396.

Bouguer formula, 74, 289, 293.
Boussingault, 358.
Boyle, Willsie and, 376-379.
Boys, C. V., 28.
Bradley, 396, 436.
Brown, 358.

and Escombe, 339.

E. W., on Moon's motion,

7, 16.

Buffon, 363.

Buisson, 96, 99, 104, 256.
Bunsen, 38, 39, 87, 98, 349.
Burning mirrors, 363.

Calcium, circulation in sun,
101.

hydride, 209.

level, 97, 119, 126, 127.

lines, wave length, 102

spectroheliograms, 118.
Calorie, 68.

439

INDEX

Calvert, 132.

Campbell, 135, 169, 180, 203,

395-398, 406, 422-429.
Carbon, in sun, 93.

dioxide, absorption and radi-
ation, 281, 291.
and plants, 335-344.
Carrington, 32, 122-126, 153,

192, 198, 436.
Cavendish, 28.
Cebrian, Molera and, 375.
Chamberlin, 274, 276, 330, 413-

418.

Chase, 393.
Chemistry, of stars, 405.

of sun, 87, 91.
Chevalier, 124, 126.
Christie, 195.
Chromosphere, flash spectrum,

168, 177-179.
heights, 171, 172.
in daylight, 137, 142, 149.
spectrum, 88, 143, 145, 180,

233.

in daylight, 180, 181.
Clark, 381.

Coast survey, U. S., 18.
Coelostat, 36.

Comparator measurements, 61.
Convection currents in sun, 100,

267.

versus radiation, 103.
Cook, Captain, 14.
Copernicus, 1, 392.
Cornu, 21.
Corona, solar, 131-136, 263-

265.

Coronium, 134.
Cortie, 210.
Cranks, 8.
Cyanogen in the sun, 94.

Darwin, G., 413, 424.

I)e Candolle, .'555.

De Saussure, 364, 374.

Dcslandres, 117, 135, 167.

Diffraction, 51, 52.

Dispersion, anomalous and reg-
ular, 233.

Dixon, Mason and, 14.

Doppler effect, 23, 41, 42, 88,
100, 124, 257, 397, 401,
422, 430.

Dorsey, 21.

Double reversal, 147, 148.

Duner, 124, 126.

Dyson, 135, 169-171, 432.

Earth, dimensions and mass,

18, 27.
temperature and radiation,

308-316.

and sun's variability, 317.
Ebermeyer, 340.
Eclipse, solar, 128-131.
of 1868, 137.
of 1900, 132.
of 1905, 132, 173-179.
of 1908, 133, 135.
Efficiency, thermodynamic, 387.
Elements, chemical, boiling

points, 238.
in sun, 87, 89-91, 252.
represented in flash-spec-
trum, 169, 170, 177,
179.

Elkin, 12.

Ellerman, 118, 120, 167.
Emission, selective, 71.
Encke, 15.

Eneas, solar machine, 367-372.
Energy, available quantity, so-
lar, 383-387.

440

INDEX

Energy, spectrum, solar, 68, 82,

292.

over sun's disk, 107, 109,
Engelmann, 342.
Enhanced lines, 100, 104, 105,

207.

Ericsson, 366.
Eros, 12, 16, 17.
Escombe, Brown and, 339.
Etiolation, 345.
Evershed, 99, 115, 166-169, 177,

209, 210, 213, 244, 255.
Exner, 300-304.
Eye, as optical instrument,

84.
Eye-pieces, solar, 33, 34.

Fabricius, 183.
Fabry, 96, 99, 104, 256.
Faculae, position and appear-
ance, 85, 202.
Faye, 123.
Fenyi, 152.
Fizeau, 19.
Flacculi, 118, 121.

motion, 122.
Flash-spectrum, 168, 170.

of 1905, 173-180.
Fluorescent spectrum, 264.
Fluorite, transparency, 56.
Foucault, 20.
Fowle, 108, 190, 317.
Fowler, 209.
Fox, 100, 122, 124, 126, 198,

270.
Fraunhofer, 39.

lines, 82, 87, 98.

principal solar, 90.
Fresnel rhomb, 44, 211.
Frost, 169, 172, 204, 205, 398,
407.

Gale, 208, 209.

Galileo, 1, 183, 436.

Gas, levels in stars, 426-429.

spectrum, 71, 249.
Gauss, 435.
Gegenschein, 400.
Geography, plant, 346-354.
George III., 13.
Gill, 11.

Glacial periods, 273-277, 322.
Glass, transparency, 56, 375.
Gould, 317.
Grant, 142.

Granulation, solar, 202.
Grating, action, 51.

concave, 54.

plane, 55.

ruling, 52.
Gravitation, 2.

constant of, 28, 29.

is universal, 400.

Hale, 59, 61, 117, 118, 124-126,
181, 182, 208, 209, 211,
212, 268.
Halm, 24, 124, 126, 185, 256,

271, 397.

solar theory, 225-228.
Hann, 312.
Hansen, 15, 435.
Hartmann, 234.

Harvard, classification of stel-
lar spectra, 409.
college observatory, 23.
Hastings, 221, 226.
Heat and radiation, 64.
Heliomicrometer, 60.
Heliotropism, 354.
Helium, in chromosphere, 170,
prominences, 139.
stars, 409.

441

INDEX

Helmholtz, 226, 248, 277, 278,

413.

Hercules, constellation, xx, 296.
Herschel, 137, 141.

J., 33, 364.

W., 396.
Hinks, 13.
Hipparchus, 5.
Hoesen, 363.
Holden, 135.

Hot-box, 364, 376-379, 389.
Hough, 397.

Huggins, 141, 142, 425, 436.
Humphreys, 43, 99.
Hydrogen, in stars, 409.

in sun-spots, 206.

level in sun, 97, 119, 127.

spectroheliograms, 118.

Insolation, march of, 285.
Intensities, spectrum, and at-
omic weights, 89, 91, 93.
Interferometer, 96.
Iron, level in sun, 97.

Janssen, 137, 138, 142, 168, 203.
Jewell, 43, 99, 115, 169, 171-

173, 255.
Josse, 278.

Julius, xxiii, 182, 259, 260, 262.
Jupiter and sun-spots, 188.
satellites and sun's distance,

23.

Kapteyn, xxiv, 396, 430-433,

437.

Kayser, 93, 254.
Keeler, 232.
Kepler, 4-6, 27, 402.
Ketchum, 376-378.
,', 208.

Kirchoff, xxiii, 38, 39, 64, 87,

98, 4:;ti.

law, (if,, 71, 208.
Knicp, and Minder, 344.
Knopf, 232.
Koppen, 190, 317.
Krigar-Menzel, 29.
Kustner, 24.

Lambert, formula, 74.
Lampblack absorption, 386.
Lane, law, 224, 255.
Langley, 81, 130, 200-206, 226,

295, 311, 373,
Laplace, 6, 412, 413, 435.
Layer, reversing, 88, 170, 178.
Levels of spectrum formation,

97, 102, 118-120, 126,

127, 171.
Leverrier, 16.
Lewis, 135, 136, 263.
Light, and plant growth, 342-

346.
direction and plant growth,

354-357.

velocity, 19-21, 46.
year, 294.

Limb spectrum, 105.
Littrow spectroscope, 58.
Lockyer, 90, 137, 139, 142,

147, 168-171, 185, 195,

196, 222, 224, 255.
Lord, 169.

Magnesium hydride, 209.
Magnetism, effect on spectrum,

43.

in sun-spots, 211-213.
terrestrial and sun-spots,

187-191.
Manson, 273.

442

INDEX

Mars, and solar parallax, 11.
Mascari, 195.
Maskelyne, 14, 27.
Mason and Dixon, 14.
Maunder, 126, 189, 198.
Measurement, spectrum, 61.
Michell, 28.
Michelson, A. A., 21, 96.

W. A., 79, 260.
Minder, Kniep and, 344.
Mira Ceti, variability, 404.
Mirrors, burning, 363.

reflecting power, 388.
Mitchell, S. A., 168-172, 180,
181, 255.

flash-spectrum of 1905, 173-

180.

Mohler, 43, 99.
Moissan, 237, 239.
Molera and Cebrian, 375.
Moore, 133.
Motion iu line of sight, 40, 127,

213, 397, 422.
Mouchot, 366.

Moulton, 271, 330, 413-418.
Mount Whitney, 69, 296-298.
Mount Wilson, xxiii, 57, 69, 89,
118, 121, 210, 296-298.

Solar Observatory, 57, 60.

Natanson, xxiii, 305, 343.
Naval Observatory, U. S., 132,

173.

Nebulae, and stars, 418-421.
solar, 276, 326, 412-418.
velocities, 430.
Neptune, distance, xviii, 3.

sun's attraction, 2, 3.
Newcomb, 8, 15, 16, 21, 27, 185,
190, 193, 236, 278, 317,
396, 400, 420, 435.

Newton, xviii, 6, 28, 86, 435.

Nicol prism, 44, 211.

Nitrogen required by plants,

334.
Nordmann, 190, 317.

Olmsted, 209.

Oltmans, 356.

Ore reduction, 381.

Orion stars, 408, 419, 430-432.

Oxygen in sun, 92.

Parallax, solar, 9, 12-16, 22-25.

stellar, 392.
Paschen, 427.
Permian glaciation, 273-277,

322-329.
Perot, 96.

Perrine, 135, 169, 263.
Perrotin, 21.
Perry, 130, 152, 198.
Perseus, new star in, xix.
Pfeffer, 358.
Photography of sun, 34, 86.

spectrum, 57.

Photosphere, spectrum, ele-
ments, 91.
extent, 86.
sun's, 85, 217, 220.
Pickering, E. C., 226, 228, 403,

409.

Pifre, 366.

Planck, law, 66, 251, 426.
Planetesimal hypothesis, 414.
Planets, minor, 12, 13, 17.

principal data, 3.
Plants and light direction, 354-

357.

and the sun, 331-361.
as energy accumulators, 357-
361.

443

INDEX

Plants, chemical requirements,
331-335.

geography, 346-354.

light requirements, 349-354.

rest periods, 348.
Plateau, 31.

Platinum black absorption, 387.
Pleiades, 398.
Pogson, 137.
Poincare", 423.
Polarization of corona, 134.

of light, 44.
Potassium in sun, 92.
Pouillet, 73, 74.
Pressure, effect on spectrum,
43, 104.

in sun, 98, 213.
Pringoheim, 261-264.
Prism action, 47-49.
Procyon, 40.

Prominences and sun-spots, 149,
153, 155.

classification, 156.

detached, 161.

eruptive, 157, 162, 163.
rapid change, 164, 165.
spectrum, 162.

in full daylight, 138-142,
149.

magnitude, 155.

number and distribution, 152,
153.

quiescent, 159-161.

solar, 121, 213.

specj;roheliograms, 166.

spectrum, 138-143.
Pyrheliometer, errors, 78.

Michelson's, W. A., 79.

Pouillet's, 73.

silver disk, 75, 76.

waterflow, 79.

Quartz, transparency, 56.

Radiation and plant growth,
342-346.

and temperature, 41, 64, 70,
308-316.

measurement, 283, 284.

nature, 63, 286.

over sun's disk, 107, 110.

solar constant of, xx, 75, 295,
298.

unit, 68.

versus convection in sun, 103.
Radiator, perfect, 64-66.
Radium in sun, 94, 272.
Rayet, 137, 138.
Rayleigh, xxiii, 241, 304.
Reed, 148.

Reflecting powers, 388.
Refraction, anomalous and reg-
ular, 233.

circular, 229.

indices, 46.

law, 48.
Reinke, 342.
Respighi, 155.
Reversal, double, 147, 148.
Reversing layer, 88.

pressures, 99.

thickness, 171.
Ricco, 166, 195.
Rice-grains, 85.
Richarz, 29.
Ritchey, 418.
Roberts, 403.
Roche, 413.

Rock-salt, transparency, 56.
Romer, 23.
Rosa, 21.
Roscoe, 302, 349.
Rosse, 311.

444

INDEX

Rotation, solar, 122-127.
Rowland, 52, 62, 89, 90, 169,

175, 212, 254, 268, 436.
" Preliminary Table," 89.
wave lengths, corrections, 95,

96.

Runge, 93, 176, 254.
Russell, 404.

St. John, 89, 101, 121, 127,

213, 269.
Salisbury, 274.
Salt, rock-, dispersion, 233.

transparency, 56.
Sainson, 23, 271.
Saros, 129.

Scattering of light by gases, 241.
Schaberle, 265.
Schehallien, 28.

Scheiner, discovery of sun-
• spots, 183.
J., 113, 247, 410, 421, 426.
Schmidt, xxiii, 244.
solar theory, 228.
Schramm, 303.
Schultz, 237.
Schuster, xxiii, 188, 236, 242,

271, 301.

Schwabe, 184, 436.
Schwartzchild, 103, 108, 203,

204, 237, 247.
Schwendener, 338.
Secchi, 142, 152, 156, 161, 166,

196, 198, 236, 407-409.
See, T. J. J., 103, 226, 247, 417,

430.

Shackelton, 168, 169.
Shearman, 190.
Shuman, 376.
Sidgreaves, 152, 198.
Silver, reflecting power, 56.

Sky, light of, 299-307, 349-354.
Slocum, 166.

Smithsonian Institution obser-
vations, 69, 86, 203, 284,
298, 321, 384, 387.
Smithsonian Observatory, Mt.
Wilson, 37.

pyrheliometer, 75.
Smoke, absorption, 386.
Snow telescope, 60.
Sodium, anomalous dispersion,

233.

Spectroheliograms, 118-120.
Spectroheliograph, 59, 117.

and prominences, 166.
Spectroscope, 45.

d vitesse, 117.

grating, 53, 54.

prismatic, 50.
Spectrum, analysis, 39.

appearance, 38.

corona, 134.

limb, 105.

lines, broadening in sun-
spots, 211.
telluric, 43.

meaning, 39-41.

measurement, 61.

negative elements, 95.

prominence, 138, 145-147.

solar, extent, 86.

sun-spot, 203-210.
Spoerer, 122, 123, 126, 193,

194, 220, 436.
Stark, 261.

Stars and nebulae, 399, 418-
421.

density, 403.

distance, xix, 392.

double, 400-404.

evolution, 418-434.

445

INDEX

Stars, fission of, 429-433.
groups, 398.
magnitudes, 394.
mass, 402, 424.
Orion, 408, 419.
solar, 408, 419.
solar and Antarian, 433.
spectra, 405-412.
and velocities, 430.
classification, 407-409, 422.
energy, distribution, 410.
spectroscopic, binary, 401,

422-429.
streams, 432.
Stefan saw, 67, 110, 250.
Sterner, 7.
Stomata, 338.
Stone, 317.
Stoney, 222, 420.
Stralonoff, 123, 126.
Struve, 392.

Sun, among the stars, 391-434.
and plants, 331-361.
axis, 122.
brightness, distribution, 105-

107.

compared to Mira Ceti, 404.
cooking, 379.
corona, 263.
density, 30.

dimensions and mass, 26, 27.
distance, geometrical meth-
ods, 9, 12, 13, 15.
gravitational methods, 15.
summary, 25, 26.
velocity of light methods,

22, 23.
energy supply, 271, 277, 383-

387.

gaseous, xxiii, 30, 217, 229,
237.

Sun, general features, xviii.
heaters and reservoirs, 375-

380.

machines, 366-372, 376-379.
motion among stars, 395.

in space, xx.
nature of, Abbot, 236-279.

Halm, 225-228.

Julius, 233-236.

Schmidt, 228-233.

Young, 215-225.
ore reduction, 381.
origin, 412-418.
photosphere, edge darkening,
249.

levels, 250.

nature, 246.

" rice-grains," 247.

spectrum, 249.

temperature variation, 250.

thickness, 244.
prominence spectrum, 259.

velocities, 259.
radiation, constant, 75.

dependence on air mass,
286.

measurement, 283, 284.
rotation, 122-127.
sharp boundary, 241.
spectrum and chemical ele-
ments, 252.

bright line, 259.

center and edge, 255.

dark line, 251.
stellar magnitude, 394.
temperature, 70, 109-116,

265.
utilization of energy, 362-

390.

variability, 317.
Sunflower, economy, 361,

446

INDEX

Sun-spots and associated phe- Temperature, attained in " hot-

noinena, 189, 190.
and llocculi, 121, 122.
and magnetism, 187, 191,

211.
and prominences, 149, 153,

155.

and solar rotation, 122.
and terrestrial temperature,

190, 191, 317.
coolness of, 207-210.
darkness of, 203, 204.
discovery, 183.
distribution on sun, 193-19G.
drift, 192.
formation and life history,

19G-199.

length of period, 185.
level, 199.

motion within, 213.
nature of, 2G7.
periodicity, 184-188, 192, 225.
pressure within, 213.
relative numbers, 184.
size, 183.

spectrum, 203-210.
titanium oxide in, 94.
System, solar, dimensions, 1, 3.
evolution, 412-418.
relative distances, 5.

Tacchini, 152, 158, 166.

Tatnall, 254.

Telescope for solar work, 31,

58.

snow, 60.
tower, 58.
Temperature and plant growth,

341.

and radiation, 41, 64, 70, 308-
316.

box," 305, 374.
earth's, at high altitude,

280-283.

geologic, 273-277, 322.
importance in sun, 265.
of stars, 411.

progressive, 429, 433.
of sun, 109-116.
terrestrial, 311-316.
Tennant, 137, 141.
Thermodynamic efficiency, 387.
Thomson, 213.
Tisserand, 123.
Titanium oxide in sun, 94,

209.
Transmission, atmospheric, 294-

297.
Transparency of optical media,

56.

Trouvelot, 156.
Turner, 25, 132, 133.
Tyndall, pyrheliometer, 75.

Uranium in sun, 94.

Veeder, 220.

Venus, transits, 13, 15.

Very, 311.

Villager, 108, 203, 204.

Vogel, 228, 409-412.

Von Gothard, 152.

Vortices, solar, 120.

Water, required by plants, 332.

vapor, radiation and absorp-
tion, 281, 291.
Wave lengths, accuracy, 62.

and levels, 98.

and plant growth, 342.

and pressures, 99.

447

INDEX

Wave lengths, principal solar,

90.

range, H. 2s:?.
Rowland's 89.

corrected, 02, 9(>.
\\ hcatstone, bridge, 81.
Wiedemann, E., 255.
Wien, displacement law, 07, 70,

100, 119, 250.

Wien-Planck law, 00, 111, 251.
Wiesner, 302, 349-354.
Wilczynski, 271.
Willsie and Boyle, 370-379.
Wilsing, 113, 123, 271, 410,421,

420.
Wilson, A., 199, 200.

W. E., 204, 205.
Witt, Eros parallax, 10.

Wolf, 184, 185, 430.

sun-spot numbers, 1M4, ISC,

194, 1 !>:..
\\oltVr, 185.
Wood, R. W., 233, 204.

Yerkes Observatory, 132, 107,

407.
Young, 5, 10, 134, 137, 144,

100, 101, 105, 108, 193,

190, 200.
views of sun's nature, 215-

225.

Zeeman, 43, 211.
Zodaical light, 400.
Zollner, 142.

(1)

THE END

UNIVERSITY OF TORONTO
LIBRARY

Do not

re move

the card

from this

Acme Library Card Pocket

Under Pat. " Ref. Index File."
Made by LIBRARY BUREAU

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "The sun"

See other formats