"? *~~ I Lf JUL 1 2 '33 +*y {/ UNIVERSITY OF CALIFORNIA AT LOS ANGELES \L. \ THE SUN JUL 1 2 '33 BY C. A. YOUNG, PH. D., LL. D. PROFESSOR OF ASTRONOMY IN PRINCETON UNIVERSITY WITH NUMEROUS ILLUSTRATIONS NEW AND REVISED EDITION NEW YORK D. APPLETON AND COA^PANY 1896 DEPARTMENT OF ASTRONOMY UNIVERSITY OF CALIFORNIA AT LOS ANGELES COPYRIGHT, 1881, 1886, 1895, BY D. APPLETON AND COMPANY. a PKEFACE TO THE KETISED EDITION. SINCE the original publication of this book in 1881 great advances have been made in our knowledge of the sun, and in the four or five editions which have subse- quently appeared the attempt has been made to keep the book measurably up to date by the addition of appendices and notes. The time has come, however, when such expedients are no longer adequate, and the author has therefore thoroughly revised the work, rewriting portions, em- bodying notes in the text, and adding whatever seemed necessary to make the book fairly representative of the solar science of to-day. The progress of discovery with respect to helium has been so continuous and rapid during the revision and printing of the work, that I have found it neces- sary to append a supplementary note upon the subject. Special thanks are due to Prof. Hale for several of the finest of the twenty new illustrations, and to Ginn and Co. for the use of one or two cuts from my Gen- eral* Astronomy. November, 1S05. v 204830 FROM PREFACE TO THE FIRST EDITION. Iris my purpose in this little book to present a gen- eral view of what is known and believed about the sun, in language and manner as unprofessional as is con- sistent with precision. I write neither for scientific readers as such, nor, on the other hand, for the masses, but for that large class in the community who, without being themselves engaged in scientific pursuits, yet have sufficient education and intelligence to be inter- ested in scientific subjects when presented in an un- technical manner ; who desire, and are perfectly com- petent, not only to know the results obtained, but to understand the principles and methods on which they depend, without caring to master all the details of the investigation. I have tried to keep distinct the line between the certain and the conjectural, and to indicate as far as possible the degree of confidence to be placed in data and conclusions. It is hardly necessary to say that the work has small claims to originality. I have made use of material suited to my purpose from all accessible sources ; possi- FROM PREFACE TO THE FIRST EDITION. v ii blj in some cases (though I hope not) without giving sufficient credit to the original authority. I have been specially indebted to Secchi, Lockyer, Proctor, Ran- yard, Vogel, Schellen, and Langley. . . . PRINCETON, August 1, 1SS1. CONTENTS. INTRODUCTION. PAGE The Sun's Relation to Life and Activity upon the Earth. Brief State- ment of the Principal Facts relating to the Sun, and of the Ac- cepted Views as to its Constitution I CHAPTER I. DISTANCE AND DIMENSIONS OF THE RUN. Importance of the Problem. Definition of Parallax. Aristarchus's Determination of the Parallax. Different Available Methods. Observations of Mars and of the nearer Asteroids. Transits of Venus. Observations of Contacts, Heliometric Measures, and Photographic Work. Determination of Solar Parallax by means of the Velocity of Light ; by Lunar and Planetary Perturbations. Illustrations of the Immensity of the Sun's Distance. Diameter of the Sun. The Sun's Mass and Density CHAPTER II. METHODS AND APPARATUS FOR STUDYING THE SUR- FACE OF THE SUN. Projection of Solar Image upon a Screen. Carrington's Method of determining the Position of Objects on Sun's Surface. Solar Photography. Photoheliographs. Cornu's Methods. Telescope with Silvered Object-Glass. Herschel's Solar Eyepiece. The Polarizing Eyepiece ix x CONTENTS. CHAPTER III. THE SPECTROSCOPE AND THE SOLAR SPECTRUM. PAGE The Spectrum and Fraunhofer's Lines. The Prismatic Spectroscope ; Description of Various Forms and Explanation of its Operation. The Ditt'raetion Spectroscope. The Concave Grating. Analyzing and Integrating Spectroscopes. The Telespectroscope and its Adjustment. The Spectrograph. Explanation of Lines in the Spectrum. Kirchhoff's Researches and Laws. The Sun's Ab- sorbing Atmosphere and Reversing Layer. Elements present in the Sun. Lockyer's Researches and Hypothesis. Basic Lines. Dr. H. Draper's Investigations as to the Presence of Oxygen in the Sun. Schuster's Observations. Efi'ect of Motion upon Wave- Length of Rays and Spectroscopic Determinations of Motion in Line of Sight 59 CHAPTER IV. SUN-SPOTS AND THE SOLAR SURFACE. Granulation of Solar Surface. Views of Langley, Nasmyth, Secchi, and others. Faculse. Nature of the Photosphere. Janssen's Photographs of Solar Surface the Roseau Photospherique. Dis- covery of Sun-Spots. General Appearance and Structure of a Spot Its Formationand Disappearance. Duration of Sun-Spots. Remarkable Phenomena observed by Carrington and Hodgson. Observations of Peters. Dimensions of Spots. Proof that Spots are Cavities. Sun-Spot Spectrum. " Veiled Spots." Rotation of Sun. Equatorial Acceleration. Explanations of the Accelera- tion. Position of Sun's Axis and Sccchi's Table for its Position Angle at Different Times in the Year. Proper Motions of Spots. Distribution of Spots 102 CHAPTER V. PERIODICITY OF SUN-SPOTS; THEIR EFFECTS UPON THE EARTH, AND THEORIES AS TO THEIR CAUSE AND NA- TURE. Observations of Schwabe. Wolfs Numbers. Proposed Explanations of Periodicity. Connection between Sun-Spots and Terrestrial Magnetism. Remarkable Solar Disturbances and Magnetic CONTENTS. x i PAGE Storms. Effect of Sun-Spots on Temperature. Sun-Spots, Cy- clones, and Rainfall. Researches of Symons and Mcldrum. Sun- Spots and Commercial Crises. Galileo's Theory of Spots. Her- schel's Theory. Secchi's First Theory. Zollner's. Faye's. Secchi's Later Opinions. Theories of Loekyer, Schaeberle, Oppol- zer, and others 151 CHAPTER VI. THE CHROMOSPHERE AND THE PROMINENCES. Early Observations of Chromosphere and Prominences. The Eclipses of 1842, 1851, and 1860. The Eclipse of 1868. Discovery of Jans- sen and Loekyer. Arrangement of Spectroscope for Observations upon Chromosphere. Spectrum of Chromosphere. Lines always present. Lines often reversed. Spectrographs of Hale and Des- landres. Motion Forms. Double Reversal of Lines. Distribution of Prominences. Magnitude. Classification of Prominences as quiescent, and eruptive or metallic. Isolated Clouds. Violence of Motion. Observations of August 5, 1872. Theories as to the Formation and Causes of the Prominences . 192 CHAPTER VII. THE CORONA. General Appearance of the Phenomenon. Various Representations. Eclipses of 1857, 1860, 1867, 1868, 1869, 1871, and 1878. Proof that the Corona is mainly a Solar Phenomenon. Brightness of the Corona. Connection with Sun-Spot Period. Spectrum of the Corona. Application of the Analyzing and Integrating Spectro- scopes. Polarization. Evidence of the Slitless Spectroscope as to the Constitution of the Corona. Changes and Motions in the Corona. Its Form and Constitution, and Theories as to its Nature and Origin 237 CHAPTER VII L THE SUN^S LIGHT AND HEAT. Sunlight expressed in Candle-Power. Method of Measurement Brightness of the Sun's Surface. Langley's Experiment. Dimin- x ji CONTENTS. PAGE ution of Brightness at Edge of the Sun's Disk. Hastings's View as to Nature of the Absorbing Envelope. Total Amount of Ab- sorption by Sun's Atmosphere. Thermal, Luminous, and Actinic Rays : their Fundamental Identity and Differences. Measurement of the Sun's Radiation. Herschel's Method. Expressions for the Amount of Sun's Heat. Fouillet's Pyrheliometer. Crova's. Violle's Actinometer. Langley's Researches. Absorption of Heat by Earth's Atmosphere : by the Sun's. Question as to Differences of Temperature on Different Portions of Sun's Disk. Question as to Variation of Sun's Radiation with Sun-Spot Period. The Sun's Temperature Actual Effective. Views of Secchi, Ericsson, Pou- illet,Vicaire, Rosetti, Le Chatelier, and Wilson and Gray. Evidence from the Burning-Glass. Langley's Experiment with the Besse- mer " Converter." Permanency of Solar Heat for last Two Thou- sand Years. Meteoric Theory of Sun's Heat. Helmholtz's Con- traction Theory. Possible Past and Future Duration of the Sun's Supply of Heat. Siemens's Untenable Theory .... 276 CHAPTER IX. SUMMARY OF FACTS, AND DISCUSSION OF THE CONSTI- TUTION OF THE SUN. Table of Numerical Data. Constitution of Sun's Nucleus. Peculiar Properties of Gases under High Temperature and Pressure. Characteristic Differences between a Liquid and a Gas. Consti- tution of the Photosphere and Higher Regions of the Sun's At- mosphere. Professor Hastings's Theory. Pending Problems of Solar Physics 323 NOTE ON HELICM 344 INDEX .351 THE SUN INTRODUCTION. THE SUN'S RELATION JO LIFE AND ACTIVITY UPON THE EARTH. Brief Statement of the Principal Facts relating to the Sun, and of the Accepted Views as to its Constitution. IT is true that from the highest point of view the sun is only one of ti multitude a single star among millions thousands of which, most likely, exceed him in brightness, magnitude, and power. He is only a pri- vate in the host of heaven. But he alone, among the countless myriads, is near enough to affect terrestrial affairs in any sensible degree ; and his influence upon them is such that it is hard to find the word to name it ; it is more than mere control and dominance. He does not, like the moon, simply modify and determine certain more or less important activities upon the surface of the earth, but he is almost absolutely, in a material sense, the prime mover of the whole. To him we can trace directly nearly all the energy involved in all phenomena, mechanical, chemi- cal, or vital. Cut off his rays for even a single month, and the earth would die ; all life upon its surface would cease. There always has been a more or less distinct recog- nition of this fact. The first man's experience of the 2 INTRODUCTION. first sunset ever witnessed by human eyes must have made it tremendously obvious, when he saw the sun descend below the horizon, and the darkness close in upon the earth, and felt the chill of night, and fell asleep not knowing of a sunrise to come unless, per- haps, some divine revelation took pity on the hopeless terror he must otherwise have suffered, or unless he may have been, like a little child, slow to notice and unable to comprehend what would frighten a more in- telligent being. But while the material supremacy of the sun has always been recognized by thoughtful minds, and has even been made the foundation of religious systems, as with the Persians, it has been reserved for more mod- ern times, and to our own century, to show clearly just how, in what sense, and how far the sunbeams are the life of the earth, and the sun himself the symbol and vicegerent of the Deity. The two doctrines of the corre- lation of forces and the conservation of energy, having once been distinctly apprehended and formulated, it has been comparatively easy to confirm them by experi- ment and observation, and then to trace, one by one, to their solar origin, the different classes of energy which present themselves in terrestrial phenomena to show, for instance, how the power of waterfalls is only a trans- formation of the sun's heat ; and that the same thing is true, a little more remotely but just as certainly, of the power of steam, of electricity, and even of animals. The idea is now so familiar that it is hardly necessary to dwell upon it, and yet, for some of our readers at least, it may be worth while to examine it a little more closely. Whenever work is done, it is by the undoing of some previous work. When a clock moves, it is the unwind- INTRODUCTION. 3 ing of a spring or the falling of a weight which keeps it going, and some one must have wound it up to begin with. If the water of a river falls year after year over a cataract, and is intercepted to drive our mill-wheels, the river continues to run because some power is con- tinually raising and returning to the hill-tops the water which has flowed into the sea a process precisely equivalent to the daily rewinding of the clock. If the powder in a rifle explodes and drives out the bullet, its explosive energy depends upon the fact that some power has placed the component molecules in such relations that, when the trigger is pulled, and the exciting spark has, so to speak, cut the bonds which hold them apart, they rush together just as suspended weights would fall if freed. Before the same substance, which once was a charge of gunpowder, but now is dust and gas, can again do the same work, the products of the ex- plosion must by some power be decomposed, and the atoms replaced in the same relations as before the firing of the gun ; and this process is mechanically analogous to the lifting of fallen weights and placing them upon elevated shelves, or hanging them from hooks, ready to drop again when the occasion may require. Precisely the same thing is true of the heat pro- duced by the combustion of ordinary fuel : it is due to the collapse of molecules, for the most part of oxygen on one side, and carbon and hydrogen on the other, which have been separated and built up into structures by the action of some laboring power. The same can be said of animal power, for all inves- tigation goes to show that in a mechanical sense the body of an animal is only a very ingenious and effective machine, by means of which the living inhabitant which controls it can utilize the energy derived from the food 4 INTRODUCTION. taken into the stomach. The body, regarded as a mech- anism, is only a food-engine in which the stomach and lungs stand for the furnace and boiler of a steam-engine, the nervous system for the valve-gear, and the muscles for the cylinder. How the personality within, which wills and acts, is put into relation with this valve-gear, BO as to determine the movements of the body it re- sides in, is the inscrutable mystery of life ; the facts in the case, however, being no less facts because inex- plicable. And now, when we come to inquire for the source of the energy which lifts the water from the sea to the mountain-top, which decomposes the carbonic acid of the atmosphere, and plant-foods of the soil, and builds up the hydrocarbons and other fuels of animal and vegetable tissue, we find it always mainly in the solar rays. I say mainly because, of course, the light and heat of the stars, the impact of meteors, and the prob- able slow contraction of the earth, are all real sources of energy, and contribute their quota. But, as compared with the energy derived from the sun, their total amount is probably something like the ratio of starlight to sunlight ; * so small that it is quite clear, as we said * Pouillet, about 1838, came to a conclusion entirely inconsistent with the statement of the text. From his actinometric observations, he deduced a temperature of 224 F. (142 C.) for the "tem- perature of space," which is 236 (131 C.) above the absolute zero. To maintain this temperature of 224, he calculated that the stars and space in general must furnish to the earth about 85 per cent, as much heat as the sun supplies. His calculations, however, rest upon assump- tions as to the laws of cooling and radiation which are not at present re- ceived as accurate, and he fails to take proper account of the influence of water-vapor in the air an influence, the magnitude of which was first brought out more than twenty years later by the researches of Tyndall and Magnus. It is now generally admitted, therefore, that his result can not be accepted. INTRODUCTION. 5 before, that a month's deprivation of the solar rays would involve the utter destruction of all activity upon the earth. It is natural, therefore, that modern science should make much of the sun, and that the study of solar phe- nomena and relations should be pursued with the great- est interest. For the last fifty years this has been especially the case : Schwabe's discovery of the perio- dicity of the sun-spots in 1851 ; the development of spectroscopic analysis between 1854 and 1870; the eclipse observations since 1860 ; the researches of Car- rington, Huggins, De La Rue, Lockyer, Janssen, Secchi, Vogel, Langley, Hale, and others ; the establishment of the observatories at Potsdam and Meudon these are all evidences of the ardor with which astronomers have devoted themselves to the problems of solar science, and of their rich rewards. It may be well, before entering upon the more extended discussion of our subject, to summarize here a few of the more important and obvious facts re- lating to the sun, with a brief statement of the views at present generally held in regard to its constitu- tion. To the few unaided eyes which are able to bear its brilliance without flinching, the sun presents the appearance of a round, white disk, a little more than half a degree in diameter i. e., a row of seven hundred suns, side by side, would just about fill up the circle of the horizon. Usually, without a telescope, the surface appears simply uniform, except that there is a slight darkening at the edge, and that once in a while black spots are seen upon the disk. There is nothing in the sun's appearance to indicate his real distance, and, until that is known, of course no conclusion can be arrived at (} INTRODUCTION. as to his true dimensions ; but the heat of his rays is obvious, and, long before the days of telescopes and thermometers, led to the conclusion that he is nothing more or less than an enormous ball of fire. If we watch him from day to day through the year, beginning about the 21st of March, we shall find that at noon he daily rises higher in the heavens, until about the 22d of June ; at this time he ascends to the same height each noon for several successive days, and then slides slowly south, passing on September 22d the ele- vation he had at starting, and keeping on until, on De- cember 21st, he attains his farthest southing ; thence he returns, till he reaches the place of beginning, and night and day again are equal. If, at the same time, one has noticed the stars each night, he will find the constellations to have shifted with the months, in such a way that it is clear that the sun has been traveling eastward among them through the sky^ as well as swinging north and south ; moving, in fact, yearly around the heavens in a path which is a great circle of the sphere, inclined some 23^ to the equator, and called the ecliptic, because it is only when the moon is near this line at new or full that eclipses happen. There is nothing in this motion which of itself can inform us whether its cause is a real movement of the sun around the earth, or of the earth around the sun. At present, of course, every one knows that the earth is really the moving body. A careful watching shows that her path is not quite circular, or, at least, that the sun is not exactly in the center, since it is one hundred and eighty-six days through the summer from the ver- nal to the autumnal equinox, and only one hundred and seventy-nine from the autumnal to the vernal. INTRODUCTION. f This much was known to the ancients, and the one additional fact that the sun's distance is many times greater than that of the moon ; it is all that could pos- sibly be learned without the use of the telescope and instruments of precision. Modern astronomy has gone much further. We now know that the sun's average distance from the earth is about 93,000,000 miles, and consequently that his diameter is about 865,000 miles. The sun has been weighed against the earth and found to contain a quantity of matter nearly 330,000 times as great, and comparing this with his enormous bulk, it appears that his mean density is only about one fourth that of the earth, or one and a quarter times that of \vater in other w r ords, the mass of the sun is about one fourth greater than that of a globe of water of the same size. The visible surface of the sun has been named the photosphere, and by watching the spots, which occa- sionally appear upon it, we have ascertained that it revolves upon its axis once in about twenty-five and a quarter days. At times of total eclipse, when the moon hides from us the body of the sun, we are enabled to see certain outlying phenomena at other times invisible. We find close around the luminous surface a rose-col- ored stratum of gaseous matter to which Frankland and Lockyer some years ago assigned the name of chromo- sphere. Here and there great masses of this chromo- spheric matter rise high above the general level like clouds of flames, and are then known as prominences or protuberances. Outside of the chromosphere is the mysterious co- rona, an irregular halo of faint, pearly light, composed for the most part of radial filaments and streamers, g INTRODUCTION. which extend outward from the sun to an enormous distance ; often more than a million of miles. The spectroscope informs us that, in great part at least, the elements, which exist in the lower regions of the solar atmosphere in the state of vapor, are metals we are familiar with upon the earth; while it shows the chromosphere and prominences to consist mainly of hydrogen and helium, and makes it possible to ob- serve them even when the sun is not hidden by the moon. The secret of the corona it fails to unlock as yet, though it informs us of the presence in it of an unknown gas of inconceivable tenuity. The pyrheliometer and actinoineter measure for us the outflow of solar heat, and show us that the blaze is at least seven or eight times as intense as that of any furnace known to art. The quantity of heat emitted is enough to melt a shell of ice more than a foot thick over the whole surface of the sun every second of time : this is equivalent to the consumption of a layer of the best anthracite coal over five inches thick every single minute. Combining the facts just stated, astronomers are for the most part agreed upon the following conclusions as to the constitution of the sun : 1. The central portion is probably for the most part a mass of intensely heated gases. 2. The photosphere is a shell of luminous clouds, formed by the cooling and condensation of the conden- sible vapors at the surface, where exposed to the cold of outer space. 3. The chromosphere is composed mainly of uncon- densible gases (conspicuously hydrogen) left behind by the formation of the photospheric clouds, and bearing something the same relation to them that the oxygen, INTRODUCTION. 9 and nitrogen of our own atmosphere do to our own clouds. 4o The corona as yet has received no explanation which commands universal assent. It is certainly truly solar to some extent, and very possibly may be also to some extent meteoric. CHAPTER I. DISTANCE AND DIMENSIONS OF THE SUN. Importance of the Problem. Definition of Parallax. Aristarchus's Deter- mination of the Parallax. Different Available Methods. Observa- tions of Mars and of the nearer Asteroids. Transits of Venus. Observations of Contacts and Photographic Work. Determination of Solar Parallax by means of the Velocity of Light ; by Lunar and Planetary Perturbations. Illustrations of the Immensity of the Sun's Distance. Diameter of the Sun. The Sun's Mass and Density. THE problem of finding the distance of the sun is one of the most important and difficult presented by astronomy. It6 importance lies in this, that this dis- tance the radius of the earth's orbit is the base-line by means of which we measure every other celestial distance, excepting only that of the moon ; so that error in this base propagates itself in all directions through all space, affecting with a corresponding proportion of falsehood every measured line the distance of every star, the radius of every orbit, the diameter of every planet. Our estimates of the masses of the heavenly bodies also depend upon a knowledge of the sun's distance from the earth. The quantity of matter in a star or planet is determined by calculations whose fundamental data include the distance between the investigated body and some other body whose motion is controlled or modified by it ; and this distance generally enters into the computation by its cube, so that any error in it in- DISTANCE AND DIMENSIONS OF THE SUN. H volves a more than threefold error in the resulting mass. An uncertainty of one per cent, in the sun's distance implies an uncertainty of more than three per cent, in every celestial mass and every cosmical force. Error in this fundamental element propagates itself in time also, as well as in space and mass. That is to say, our calculations of the mutual effects of the planets upon each other's motions depend upon an accurate knowledge of their masses and distances. By these calculations, were our data perfect, we could predict for all futurity, or reproduce for any given epoch of the past, the configurations of the planets and the con- ditions of their orbits, and many interesting problems in geology and natural history seem to require for their solution just such determinations of the form and po- sition of the earth's orbit in by-gone ages. Now, the slightest inaccuracy in the data, though hardly affecting the result for epochs near the present, leads to error which accumulates with the lapse of time ; so that even the present uncertainty of the sun's distance, small as it is, renders precarious all conclu- sions from such computations when the period is ex- tended more than a few hundred thousand years. If, for instance, we should find as the result of calcula- tion with the received data, that two millions of years ago the eccentricity of the earth's orbit was at a maxi- mum, and the perihelion so placed that the sun was nearest during the northern winter (a condition of affairs which it is thought would produce a glacial epoch in the southern hemisphere), it might easily happen that our results would be exactly contrary to the truth, and that the state of affairs indicated did not occur within ten thousand years of the specified date and all because in our calculation the sun's dis- 12 THE SUN. tance, or the solar parallax by which it is measured, was assumed half of one per cent, too great or too small. In fact, this solar parallax enters into almost every kind of astronomical computations, from those which deal with stellar systems and the constitution of the universe, to those which have for their object noth- ing higher than the prediction of the moon's place as a means of finding the longitude at sea. Of course, it hardly need be said that its determina- tion is the first step to any knowledge of the dimensions and constitution of the sun itself. This " parallax " of the sun is simply the angular semi-diameter of the earth as seen from the sun ; or, it may be defined in another way as the angle between the direction of the sun ideally observed from the center of the earth, and its actual direction as seen from a sta- tion where it is just rising above the horizon. We know with great accuracy the dimensions of the earth. Its mean equatorial radius, according to Hark- ness's latest determination (agreeing, however, very closely with previous ones), is 3963-124 English miles [6377-972 kilometres], and the error can hardly amount to more than - a g \ of the whole perhaps, 800 feet one way or the other. Accordingly, if we know how large the earth looks from any point, or, to speak tech- nically, if we know the parallax of the point, its dis- tance can at once be found by a very easy calculation : it equals simply [206,265 * X the radius of the earth] -H [the parallax in seconds of arc]. * This number 206,265 is the length of the radius of a circle ex- pressed in seconds of its circumference. A ball one foot in actual diam- eter would have an apparent diameter of one second at a distance of 206,265 feet, or a little more than 39 miles. If its apparent diameter were 10*, its distance would, of course, be only ^ as great. DISTANCE AND DIMENSIONS OF THE SUN. 13 Now, in the case of the sun it is very difficult to find the parallax with sufficient precision on account of its smallness it is less than 9", almost certainly between 8-75" and 8- 85". But this tenth of a second of doubtful- ness is more than y^ of the whole, although it is no more than the angle subtended by a single hair at a dis- tance of nearly 800 feet. If we call the parallax 8*80", which is probably very near the truth, the distance of the sun will come out 92,892,000 miles, while a varia- tion of -g^ of a second either way will change it about half a million of miles. When a surveyor has to find the distance of an in- accessible object, he lays off a convenient base-line, and from its extremities observes the directions of the ob- ject, considering himself very unfortunate if he can not get a base whose length is at least ^ of the dis- tance to be measured. But the whole diameter of the earth is less than j-j-g-^ of the distance of the sun, and the astronomer is in the predicament of a sur- veyor who, having to measure the distance of an ob- ject ten miles off, finds himself restricted to a base of less than five feet ; and herein lies the difficulty of the problem. Of course, it would be hopeless to attempt this prob- lem by direct observations, such as answer perfectly in the case of the moon, whose distance is only thirty times the earth's diameter. In her case, observations taken from stations widely separated in latitude, like Berlin and the Cape of Good Hope, or Washington and Santiago, determine her parallax and distance with very satisfactory precision ; but if observations of the same accuracy could be made upon the sun (which is not the case, since its heat disturbs the adjustments of an instru- ment), they would only show the parallax to be some- 14 THE SUN. where between 8" and 10"; its distance between 126,- 000,000 and 82,000,000 miles. Astronomers, therefore, have been driven to employ indirect methods based on various principles : some on observations of the nearer planets, some on calculations founded upon the irregularities the so-called pertur- bations of lunar and planetary movements, and some upon observations of the velocity of light. Indeed, before the Christian era, Aristarchus of Samos had de- vised a method so ingenious and pretty in theory that it really deserved success, and would have attained it were the necessary observations susceptible of sufficient accuracy. His idea was to observe carefully the number of hours between new moon and the first quarter, and also between the quarter and the full. The first interval should be shorter than the second, and the difference would determine how many times the distance of the sun from the earth exceeds that of the moon, as will be clear from the accompanying figure. The moon reaches its quarter, or appears as a half-moon, when it arrives at the point Q, where the lines drawn from it to the sun and earth are perpendicular to each other. Since the angle II E Q = E S Q, it will follow that H Q is the same fraction of H E as Q E is of E S ; so that, if H Q can be found, we shall at once have the ratio of Q E and E S. Aristarchus thought he had as- certained that the first quarter of the month (from N to DISTANCE AND DIMENSIONS OF THE SUN. 15 Q) was about 12 hours shorter than the second, from which lie computed the sun to be about 19 times as dis- tant as the moon. The difficulty lies mainly in the impossibility of determining the instant when the disk of the moon is exactly bisected, and depends partly upon the fact that the lunar surface is very rough, and partly upon the fact that the sun's diameter is nearly twice that of the orbit of the moon, instead of being a mere point, as in the figure. The boundary between light and darkness the terminator, as it is called is both irregular and ill-defined. The real difference be- tween the two quarters is not quite 36 minutes, so that the sun's distance is about 400 times the moon's. For more than 1,500 years, however, the result of Aristar- chus stood unquestioned, having been accepted by Hip- parchus and Ptolemy. The different methods upon which our present knowledge of the sun's distance depends may be classi- fied as follows : 1. Observations upon the planet Mars near opposition, in two dis- tinct ways : (a) Observations of the planet's declination made from sta- tions widely separated in latitude. (5) Observations from a single station of the planet's right ascension when near the eastern and western horizons known as Flamsteed's or Bond's method. 2. Observations of Venus at or near inferior conjunctions: (a) Observations of her distance from small stars measured at stations widely different in latitude. (6)' Observations of the transits of the planet : 1. By noting the duration of the transit at widely-separated sta- tions ; 2. By noting the true Greenwich time of con- tact of the planet with the sun's limb ; 3. By measur- ing the distance of the planet from the sun's limb with suitable micrometric apparatus ; 4. By photographing the transit, and subsequently measuring the pictures. 16 THE SUN. 3. By observing the oppositions of the nearer asteroids in the same manner as those of Mars. 4. By means of the so-called parallactic inequality of the moon. 5. By means of the monthly equation of the sun's motion. 6. By means of the perturbations of the planets, which furnish us the means of computing the ratios between the masses of the planets and the sun, and consequently their distances- known as Leverrier'a method. 7. By measuring the velocity of light, and combining the result (a) with " equation of light " between the earth and sun, or (&) with " the constant of aberration." Our scope and limits do not, of course, require or allow any exhaustive discussion of these different meth- ods and their results, but some of them will repay a few moments' consideration : The first three methods, known as the trigono- metrical methods, are all based upon the same general idea, that of finding the actual distance of one of the nearer planets by observing its displacement in the sky as seen from remote points on the earth. The rela- tive distances of the planets are easily found in sev- eral different ways,* and are known with very great * One method of determining the relative distances of a planet and the sun from each other and from the earth is the following, known since the days of Eipparchus : First, observe the date when the planet comes M to its opposition i. e., when sun, earth, and planet are in line, as in the figure, where the planet and earth are represented by M and E. Next, after a known number of days, say one hundred, when the planet has ad- DISTANCE AND DIMENSIONS OF THE SUN. 17 accuracy the possible error hardly reaching the ten- thousandth in even the most unfavorable cases. In other words, we are able to draw for any moment an exceedingly accurate map of the solar system the only question being as to the scale. Of course, the determi- nation of any line in the map will fix this scale ; and for this purpose one line is as good as another, so that the measurement of the distance from the earth to the planet Mars, for instance, will settle all the dimensions of the system. vanced 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 M S M' moved over by the planet in one hundred days, and also E S E', described in the same time by the earth. The difference is M' S E, often called the synodic angle. We have, therefore, in the triangle M'SE',the angle at E' measured, and the angle M S E' known as stated above, and hence by the ordinary processes of trigonometry we can find the relative values of its three sides. 18 THE SUN. Fig. 3 illustrates the method of observation. Sup- pose two observers, situated one near the north pole of the earth, the other near the south. Looking 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, l>, 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 w r lien 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 displacement which would be produced in the plan- et's position by a transference of the observer from Washington to Santiago, or vice versa. The first modern attempt to determine the sun's parallax was made by this method in 1670, when the French Academy of Sciences sent Richer to Cayenne to observe the opposition of Mars, while Cassini (who pro- posed the expedition), Roemer, and Picard observed it from different stations in France. When the results came to be compared, however, it was found that the planet's displacement was imperceptible by their exist- ing means of observation : from this they inferred that the planet's parallax could not exceed half a minute of arc, and that the sun's could not be more than 10". In 1752 Lacaille at the Cape of Good Hope made similar observations, and their comparison with cor- responding observations in Europe snowed that instru- DISTANCE AND DIMENSIONS OF THE SUN. 19 merits had so far improved as to make the displacement quite sensible. He fixed the sun's parallax at 10", cor- responding to a distance of 82,000,000 miles. In more recent times the method has been frequently applied. It can be used to the best advantage, of course, when, at the time of " opposition," the planet is near its perihelion and the earth near its aphelion, for then the distance between Mars and the earth is the least possi- ble. These favorable oppositions occur in the late sum- mer or early autumn about once in fifteen years, as in 1847, 1862, 1877, and 1892. The meridian observations which furnish the ma- terial of method 1#, and were mainly relied on until re- cently, seem for some reason, perhaps connected with the planet's red color, to be untrustworthy ; at any rate, they generally give values of the parallax persistently about one per cent, larger than the other methods, and rather discordant among themselves. Flamsteed's method, on the other hand, stands very high, especially when so modified as to utilize the co- operation of numerous observers in different countries. Though first suggested long ago, it would have amounted to very little with the instruments then available, and it had been practically lost sight of until the expedition of Dr. Gill to Ascension Island, in 1877, brought out its real value. His instrument was a " heliometer," loaned by Lord Lindsay for the occasion. It consists essentially of a telescope having its object-glass divided into two semi- circular pieces which can slide by each other. Each half of the lens makes its own image of the object under examination, so that by properly setting the semi-lenses the images of two neighboring stars can be made to coincide ; and if we know the displacement of the two 20 THE SUN. lenses, which can be measured by an accurate scale, the angular distance between the stars can be determined with a precision unattainable by any other known p^ cess. The instrument is delicate, complicated, and diffi- cult to use, but in the hands of an adept it is thoroughly reliable. It was with the heliometer that Bessel, in 1838, first sounded interstellar space by measuring the annual parallax* and distance of 61 Cygui. Mr. Gill's observations consisted in measurements of the apparent distance between the planet and the stars lying near its path, and of the distances between the stars themselves ; the principal observatories also co-operated in the work by determining with the utmost precision the absolute places of the stars. It would take too much space to explain fully how from such observations the solar parallax can be accurately worked out ; but any one can easily see that when the planet is rising the effect of parallax (which always makes a body appear lower in the heavens than it otherwise would) is to shift it apparently toward the east / when Mars is in the west the apparent shift, on the other hand, is westward ; and by comparing the measurements made at all hours of the night for several consecutive weeks the planet's regular orbital motion and the amount of this daily parallactic shift can be separately determined with mi- nute exactness. As the final result of the whole operation, Dr. Gill obtained 8'780* 0-020* for the sun's parallax. Several of the minor planets or asteroids which have very eccentric orbits at times come so near us at oppo- * The " annual," or " heliocentric," " parallax " of a star is not the same as its horizontal parallax, or angular semi-diameter of the earth as seen from the star; it is the semi-diameter of the earth's orbit viewed from the star, and is nearly twelve thousand times greater than the other. DISTANCE AND DIMENSIONS OF THE SUN. 21 sition that tliey can be advantageously observed in the same way. They never approach quite as close as Mars does, but per contra they are so much smaller that they look just like stars, and can be observed with the heliometer much more accurately than a planet which presents a disk. Very recently, in 1889 and 1890, a concerted system of observations was made upon Vic- toria, Iris, and Sappho by Dr. Gill, now the Astronomer Koyal at the Cape of Good Hope ; Dr. Eikin, of the Yale College Observatory (which possesses a fine heli- ometer, an exact mate of Dr. Gill's, and the only one in the United States), and two or three German ob- servers with smaller instruments. The results are very satisfactory, ranging from 8'796 /l ' to 8'825 /|f , the mean being 8'807*, with a probable error of only 0'006". So far as can be judged from the details thus far published, this determination must be conceded the pre- cedence over all others in respect to its probable freedom from constant and systematic errors, and from theoret- ical difficulties. In observations of this sort upon Mars or the aster- oids, the position and displacement of the planet, as seen from different stations, are determined by com- paring it with neighboring stars. When Venus, how- ever, is nearest us, she can be observed only by day, so that in her case star comparisons are as a general thing out of the question. But occasionally at her inferior conjunction she passes directly across the disk of the sun, the phenomenon being known as a "transit," These transits are very rare, coming (at present) in pairs, the two transits which constitute a pair being separated by an interval of eight years, while between the pairs themselves there is an interval of either one hundred and thirty or one hundred and thirteen years. 22 THE SUN. They occur either in June or December, and thus far there have been six since the invention of the telescope ; viz., in December, 1631 and 1639 ; in June, 1761 and 1769 ; and in December, 1874 and 1882, the next pair being due in June, 2004 and 2012. On these occasions the parallactic displacement of the planet as seen from different stations can be deter- mined by making any such observations as will enable the computer to ascertain accurately her apparent dis- tance and direction from the sun's center at some given moment. Gregory in 1663 first pointed out the utility of such observations for ascertaining the parallax, but it was not until some fifteen years later that the subject was fairly brought to the attention of astronomers by Hal- ley, who discussed the matter thoroughly, and showed how the problem might be solved with accuracy by observations such as were practicable even with the instruments and knowledge then at command. From that time for fully two hundred years it was the almost universal opinion of astronomers that no other method could rival this as a means of determining the distance of the sun. The transits of 1761 and 1769 were observed in all accessible quarters of the globe by expeditions sent out by the different governments. From different sets of these observations variously combined by different com- puters, values of the solar parallax were obtained ranging all the way from 7'5* to 9'2". A general discussion of all the material afforded by the two transits was first made by Encke in 1822, and he obtained, as the most probable result, the value 8-5776", which from that time for more than thirty years was accepted by all astrono- mers as the best attainable approximation to the truth. DISTANCE AND DIMENSIONS OF THE SUN. In 1854 Hansen, in publishing some of his results respecting the motion of the moon, announced that Encke's value of the solar parallax could not be recon- ciled with his investigations ; within the next six or seven years several independent researches by other astrono- mers confirmed his conclusions ; and the more iecent recomputations by Powalky, Stone, Faye, and others, show that the errors of observation were so considerable in 1769 that nothing more can be fairly deduced from that transit than that the solar parallax is probably somewhere between 8"7" and 8'9". The method of observation then used consisted sim- ply in noting the moment when the limb of the planet carne in contact with that of the sun an observation which is attended with much more difficulty and uncer- tainty than would at first be supposed. The difficul- ties depend in part upon the imperfections of optical instruments and the human eye, partly upon the essen- tial nature of light, leading to what is known as diffrac- tion, and partly upon the action of the planet's atmos- phere. The two first-named causes produce what is called irradiation, and operate to make the apparent diameter of the planet, as seen on the solar disk, small- 24 THE SUN. er than it really is smaller, too, by an amount which varies with the size of the telescope, the perfection of its lenses, and the tint and brightness of the sun's image. The edge of the planet's image is also rendered slightly hazy and indistinct. The planet's atmosphere also causes its disk to be surrounded by a narrow ring of light, which becomes visible long before the planet touches the sun, and at the moment of internal contact produces an appearance of which the accompanying figure is intended to give an idea, though on an exaggerated scale. The planet moves so slowly as to occupy more than twenty minutes in crossing the sun's limb ; so that, even if the planet's edge were perfectly sharp and definite, and the sun's limb undistorted, it would be very difficult to deter- mine the precise second at which contact occurs ; but as things are, observers with precisely similar telescopes, and side by side, often differ from each other five or six seconds ; and where the telescopes are not similar the differences and uncertainties are much greater. The extent of the difficulty can be judged of by the simple fact that, from the whole mass of contact obser- vations obtained in 1874 by the different British parties which observed the transit, three different values of the solar parallax have been deduced by different computers, viz., the official value 8'76" by Airy, 8'81" by Tupman, and 8*88" by Stone. These differences depend mainly upon the different interpretations given to the de- scription of phenomena noted by the observers in the field. In 1882 things were perhaps a little better, as many of the observers had the benefit of experience in 1874. Professor Newcomb deduces from all the observations of internal contact in the two transits a solar parallax DISTANCE AND DIMENSIONS OF TUB SUN. 25 of 8'776* 0'023". But many of the several hundred observations were seriously discordant. The difficulties of the case were fully realized at the time when preparations were making for the observa- tion of the transit of 1874, and astronomers were dis- posed to put more reliance upon micrometric and pho- tographic methods, which are free from these peculiar difficulties, though of course beset with others, which, however, it was hoped would prove less formidable. All the numerous expeditions, therefore, which were sent out by the various governments to observe the transits of 1874 and 1882 were equipped to use one or both these methods. All the eight German parties, two or three of the Russian parties, one English, and one Belgian party were provided with heliometers, and busied themselves during the transit with measuring the distance of the planet from the edge of the sun's disk. The results of the German observations have been fully worked out and published. From the four hundred and forty-six different measures Auwers deduces a solar parallax of 8'878* 0-040" ; the value is surprisingly large, but the magnitude of its probable error indicates that the obser- vations did not agree very closely. The Americans and French placed their main reli- ance upon the photographic method, while the English and Germans also provided for its use to a certain extent. The great advantage of this method is that it makes it possible to perform the necessary measure- ments, upon whose accuracy everything depends, at leisure after the transit, without hurry, and with all possible precautions. The field-work consists merely in obtaining as many and as good pictures as possible. A principal objection to the method lies in the difficulty 26 THE SUN. of obtaining good pictures i. e., pictures free from dis- tortion, and so distinct and sharp as to bear high mag- nifying power in the microscopic apparatus used for their measurement. A most serious difficulty, more- over, is involved in the accurate determination of the scale of the picture that is, of the number of seconds of arc corresponding to a linear inch upon the plate. Besides this, we must know the exact Greenwich time at which each picture is taken, and it is also extremely desirable that the orientation of the picture should be accurately determined that is, the north and south, east and west points of the solar image on the finished plate. There has been a good deal of anxiety lest the image, however accurate and sharp when first produced, should alter in course of time through the contraction of the collodion or gelatine film on the glass plate, but the ex- periments of Eutherfurd, Huggins, and Paschen seem to show that this danger is imaginary. The uncertainty of our present knowledge of the sun's parallax is, however, so small that we can hope to improve it only by means of photographs that are almost absolutely perfect. Unless the picture is so dis- tinct and free from distortion that the relative positions of Yenus and the sun's center can be determined from it on the four-inch disk within 3-^-5- of an inch the plate is practically worthless. But it is to be noted that any mere enlargement or diminution of the diameter of sun or planet will do no harm, provided it is alike all around the circumference of the disk, since the measurement is not from the edge of Yenus to the edge of the sun, but between their cen- ters. Photographic determinations of contact, on the contrary (such as Janssen and some of the English par- ties attempted by a peculiar and complicated apparatus), DISTANCE AXD DIMENSIONS OF THE SUN. 27 are affected with all the uncertainties of the old-fash- ioned observations of the eye alone, and with others in addition ; so that, astronomically considered, they are entirely worthless, although interesting from a chemical and physical point of view. In 1874 two essentially different lines of proceeding were adopted in the photographic observations. The English and Germans attached a camera to the eye- end of an ordinary telescope, which was pointed directly at the sun ; the image formed at the focus of the tele- scope was enlarged to the proper size by a combination of lenses in the camera ; and a small plate of glass ruled with squares was placed at the focus of the telescope and photographed with the sun's image, furnishing a set of reference-lines, which give the means of detecting and allowing for any distortion caused by the enlarging lenses." The Americans and French, on the other hand, pre- ferred to make the picture of full size, without the in- tervention of any enlarging lens : as this requires an object-glass with a focal length of thirty or forty feet, which could not be easily pointed at the sun, a plan proposed first by M. Laussedat, but also independently by our own Professor Winlock, was adopted. The tele- scope is placed horizontal, and the rays are reflected into the object-glass by a plane mirror suitably mounted. The French used mirrors of silvered glass, and took their pictures (about two and a half inches in diame- ter) by the old daguerreotype process on silvered plates of copper, in order to avoid the risk of collodion-con- traction. With the silvered mirror the J;ime of expo- sure is so short that no clock-work is required. The Americans used unsilvered mirrors, to obviate any dis- torting action of the sun's rays upon the form of the 2$ THE SUN. mirror. This, of course, made the light feebler and the time of exposure longer, so that a clock-work move- ment of the mirror was needed to keep the image from FIG. 5. AMERICAN APPARATUS FOR PHOTOGRAPHING TIIK TRANSIT OF VENUS. changing its place on the plate during the exposure, which, however, never exceeded half a second. Fig. 5, taken from the author's "General Astronomy" by permission of the publishers, gives an idea of the ar- rangement. The pier that carries the plate was in a darkened room, into which the rays from the mirror were admitted by a sliding shutter. In 1874 the American pictures were taken by the ordinary wet process on glass, and were about four inches in diameter. In 1882 a gelatine emulsion pro- cess was used. Just in front of the sensitive plate, at a distance of about one eighth of an inch, was placed a reticle, or a plate of glass ruled in squares, and between this and the collodion-plate hung a line silver wire sus- pending a plumb-bob. Thus the finished negative was DISTANCE AND DIMENSIONS OF THE SUN. 29 marked into squares, and also bore the image of the plumb-line, which indicated precisely the direction of the vertical. The Americans also placed the photo- graphic telescope exactly in line with a meridian instru- ment, and so determined, wfth the extremest precision, the direction in which it was pointed. Knowing this, and the time at which any picture was taken, it becomes possible, Avith the help of the plumb-line image, to de- termine precisely the orientation of the picture an ad- vantage possessed by the American pictures alone, and making their value nearly twice as great as otherwise it would have been. The above figure is a representation of one of the American photographs reduced about one half. Fis the image of Venus, which on the actual plate is about one seventh of an inch in diameter ; a a' is the image of the plumb-line. The center of the reticle is marked 30 THE SUN. by the little cross, and the word " China," written on the reticle-plate with a diamond and, of course, copied on the photograph^indicates that it is one of the Peking pictures. Its number in the series is given in the right- hand upper corner. About 90 such pictures were ob- tained at Peking during the transit, and about 350 at all the eight American stations, the work being much interfered with by unfavorable weather at most of them. If we add those obtained by the French, Germans, and English, the total number available reaches nearly 1,200, according to the best estimates. After the pictures are made and safely brought home, they have next to be measured i. e., the dis- tance (and in the American pictures the direction also) between the center of Venus and the center of the sun must be determined in each picture. This is an exceed- ingly delicate and tedious operation, rendered more dif- ficult by the fact that the image of the sun is never truly circular, but, even supposing the instrument to be perfect in all its adjustments, is somewhat distorted by the effect of atmospheric refraction ; so that the true position of the sun's center with reference to the squares of the reticle is determined only by an intricate calcula- tion from measurements made with a microscopic ap- paratus on a great number of points suitably chosen on the circumference of the image. The final result of the measurement comes out something in this form : Peking. No. 32. Time, 14 h 08 m 20'2 8 (Greenwich mean time) ; Venus north of sun's center, 735 < 32" ; east of center, 441-63"; distance from center of sun, 857'75". (The numbers given are only imaginary.) In 1882 less prominence was given to photographic operations by most of the Government expeditions, since the results of. the work in 1874, so far as then published, DISTANCE AND DIMENSIONS OF THE SUN. 31 were not very satisfactory. The American parties, however, adhered to the same apparatus and methods as in 1874, except that the collodion process was replaced by an emulsion. Nearly 1,500 photographs were ob- tained. From the whole system of American photo- graphs Professor Newcomb deduces a solar parallax of 8-857" 0-016". The measures of distance alone give 8-867*, but those of position-angles gave 8-873" in 1874, and 8-772* in 1882. The discordances between the results from different plates, made within a few minutes of each other, show that there is something wrong with the method. The most probable explanation is perhaps to be found in the distortions suffered by the plane mirror of the apparatus under changes of position and temperature. From the 92 French daguerreotypes of 1874 a paral- lax of 8-80" 0-03" was deduced by Obrecht. The English photographs of 1874 proved of little value. They were measured by two different persons, and from the measurements of one (Mr. Burton) a par- allax of 8'25 /l ' was deduced, while from those of the other (Captain Tupman) the result was 8-08". One of the principal difficulties evidently lay in the uncertainty of the scale-value, which was only deduced from the di- ameters of the sun and planet. On the whole, it may be taken as certain that here- after transits of Yenus will not be considered of such supreme importance as in the past. Other less costly operations will give better results for the solar parallax. The methods numbered 4, 5, and 6, on page 16, are usually classed together as " gravitational" since they depend on calculations which are founded on the law of gravitation. One of the best of them is based upon the careful observation of the motions of the 32 THE SUN. moon. The first suspicion as to the correctness of the then received distance of the sun was raised in 1854 by Hansen's announcement that the moon's parallactic in- equality led to a smaller value than that deduced from the transit of Venus a conclusion corroborated by Leverrier four \ears later, from the so-called lunar equation of the sun's motion. It seems at first sight strange, but it is true, as Laplace long since pointed out, that the skillful astronomer, by merely watching the movements of our satellite, and without leaving his ob- servatory, can obtain the solution of problems which, attacked by other methods, require tedious and expen- sive expeditions to remote corners of the earth. Our scope and object do not require us to enter into detail respecting this lunar method of finding the sun's paral- lax ; it must suffice to say that the disturbing action of the sun makes the interval from new moon to the first quarter about eight minutes longer than that from the quarter to full ; and this difference depends upon the ratio between the diameter of the moon's orbit and the distance of the sun in such a manner that, if the in- equality is accurately observed, the ratio can be cal- culated. Since we know the distance of the moon, this will give that of the sun. The results obtained in this way, according to the most recent investigations, ap- pear to fix the solar parallax between 8-767" and 8'802". Newcomb assigns 8'794" as the weighted mean. But the method by which ultimately we shall obtain the most accurate determination of the dimensions of our system is that proposed by Leverrier, depending upon the secular perturbations produced by the earth upon her neighboring planets ; especially in causing the motions of their nodes and perihelia. These motions are very slow, but continuous ; and hence, as time goes DISTANCE AND DIMENSIONS OF THE SUN. 33 on, they will become known with ever-increasing accu- racy. If they were known with absolute precision, they would enable us to compute, with absolute precision also, the ratio between the masses of the sun and earth, and from this ratio we can calculate * the distance of the sun by either of two or three different methods. As matters stand at present, the majority of astrono- mers would probably consider that these secular pertur- bations are not yet known with an exactness sufficient to render this method superior to the others that have been named perhaps as yet not even their rival. Le- verrier, on the other hand, himself put such confidence in it that he declined to sanction or co-operate in the operations for observing the recent transit of Yenus, considering all labor and expense in that direction as merely so much waste. But, however the case may be now, there is no question that as time goes on, and our knowledge of the planetary motions becomes more minutely precise, this method will become continually and cumulatively more exact, until finally, and not many centuries hence, it will supersede all the others that have been described. * One method of proceeding is as follows : Let M be the mass of the sun and earth united, and m that of the earth and moon ; let R be the distance of the sun from the earth, and r that of the moon ; finally, let T be the number of days in a sidereal year, and t the number in a side- real month. Then, by elementary astronomy T?3 3 / T2 M : m = : - ; whence R 3 = r*[ - or, in words, the cube of the sun's distance equals the cube of the moon's distance, multiplied by the square of the number of sidereal months in a year, and by the ratio between the masses of the sun and earth. It is to be noted, however, that T and t are the periods of the earth and moon, as they would be if wholly undisturbed in their motions, and hence differ slightly from the periods actually observed the differences are small, but somewhat troublesome to calculate with precision. 4 34 THE SUN. The parallax of the sun determined by Leverrier in this method, in 1872, came out 8'86*. Professor Newcomb, as the result of his recent ex- haustive researches upon the subject, gets 8'759" O'OIO". The last of the methods mentioned in the synopsis given on pages 15 and 16 is interesting as an example of the manner in which the sciences are mutually connected and dependent. Before the experiments of Fizeau in 1849, and of Foucault a few years later, our knowledge of the velocity of light depended on our knowledge of the dimensions of the earth's orbit. It had been found by astronomical observations upon the eclipses of Jupi- ter's satellites that light occupied a little more than six- teen minutes in crossing the orbit of the earth, or about eight minutes in coming from the sun ; and hence, supposing the sun's distance to be 95,600,000 miles, as was long believed, the velocity of light must be about 192,000 miles per second. Thus optics was indebted to astronomy for this fundamental element. But when Foucault in 1862 announced that, according to his un- questionably accurate experiments, the velocity of light could not.be much more than 186,000 miles per second, the obligation was returned, and the suspicions as to the received value of the sun's parallax, which had been raised by the lunar researches of Hansen and Leverrier, were changed into certainty. The most accurate determinations of the velocity of light have been made in this country by Michelson and Newcomb, between 1879 and 1883, and give as the re- sult 186,327 miles, with a probable error not exceeding twenty miles. From this we can derive the distance of the sun di- rectly by merely multiplying it by the "constant of the DISTANCE AND DIMENSIONS OF THE SUN. 35 equation of light" winch is simply the number of seconds required by light to travel from the sun to the earth. This "constant" is determined by observation upon the eclipses of Jupiter's satellites, and is almost certainly very near 499 seconds, though still doubtful by a frac- tion of a second. This would give 92,977,000 miles for the- sun's distance, corresponding to a parallax of about S'79". During the last twelve or fifteen years continu- ous series of observations have been in progress by new photometric methods both at Cambridge (U. S.) and Paris, and when their result is published we shall un- doubtedly have a much more accurate value of the light- equation. The velocity of light may be utilized in another way to solve the problem, by combining it with the so-called " constant of aberration." This " aberration constant" is deduced from observations upon the fixed stars, and almost certainly lies somewhere between 20-45" and 20'55", corresponding to parallaxes of 8'81" and S'77*. Its determination, however, is somewhat embarrassed by the newly discovered " variation of latitude," and it is expected that new determinations, in which this varia- tion is duly eliminated or taken into account, will give a much more accurate value of the aberration. The only difficulty with these two methods lies in the theoretical question whether we can safely assume that in interplanetary space the velocity of light is identical with that determined by experiments made at the surface of the earth, even after all known corrections for the density of the air, etc., have been applied. Admitting it, there can hardly be a doubt that this "physical method" as it is often called, outranks all others for the present as a means of determining the distance of the sun ; and the reader's attention is called 36 THE SUN. to the fact that it gives directly the distance of the sun, and the parallax only indirectly. It does not depend at all upon our measures of the dimensions or gravita- tional attraction of the earth. Collecting all the evidence at present attainable, it would seem that the solar parallax can not differ much from S'SO", though it may be as much as O'Ol" greater or smaller ; this would correspond, as has already been said, to a distance of 92,892,000 miles, with a probable error of about one eighth of one per cent., or 120,000 miles.* But, though the distance can easily be stated in fig- ures, it is not possible to give any real idea of a space so enormous ; it is quite beyond our power of concep- tion. If one were to try to walk such a distance, sup- posing that he could walk 4 miles an hour, and keep it up for 10 hours every day, it would take 68 years to * The oscillations of scientific opinion as to the value of this constant have been very curious. Early in the century Laplace, in the " Mecanique Celeste," adopted the value 8'81" given by the first discussion of the tran- sits of Venus in 1761-'69 ; but other astronomers, Delambre, for instance, proposed a smaller value. Encke, as has been said before, made a new and thorough discussion of these transits in 1822-'24, and deduced the value 8'58", which held the ground for nearly forty years. About I860 the researches of Hansen, Leverrier, and Stone were thought to have established a value exceeding 8-90", and the " British Nautical Almanac " used 8-95" until the issue for 1882. In 1867 Newcomb published a care- ful investigation, based upon all the data then known, and deduced the value 8-848'. Leverrier, in 1872, found 8'86" from the planetary per- turbations. The "American Ephemeris," "British Nautical Almanac," and the Berlin " Jahrbuch " use Newcomb's value, and the French " Con- naissance de Temps" employs Leverrier's. It appears, however, per- fectly certain, from the work of the last few years, that the figures (8'80") given in the text are much nearer to the truth. Newcomb, in his " Astro- nomical Constants" (January, 1895) gives, as the final value based upon all available data, 8'797" 0'004. Harkness, in his " Solar Parallax and its related Constants," deduces as the result of a most exhaustive dis- cussion 8-809" 0-006. DISTANCE AND DIMENSIONS OF THE SUN. 37 make a single million of miles, and more than 6,300 years to traverse the whole. If some celestial railway could be imagined, the journey to the sun, even if our trains ran 60 miles an hour, day and night and without a stop, would require over ITS years. Sensation, even, would not travel so far in a human lifetime. To borrow the curious illus- tration of Professor Mendenhall, if we could imagine an infant with an arm long enough to enable him to touch the sun and burn himself, he would die of old age before the pain could reach him, since, according to the experiments of Helmholtz and others, a nervous shock is communicated only at the rate of about 100 feet per second, or 1,637 miles a day, and would need more than 150 years to make the journey. Sound would do it in about 14 years if it could be transmitted through celestial space, and a cannon-ball in about 9, if it were to move uniformly with the same speed as when it left the muzzle of the gun. If the earth could be suddenly stopped in her orbit, and allowed to fall unob- structed toward the sun under the accelerating influence of his attraction, she would reach the center in about two months. I have said if she could be stopped, but such is the compass of her orbit that, to make its circuit in a year, she has to move nearly 19 miles a second, or more than fifty times faster than the swiftest rifle-ball ; and in moving 20 miles her path deviates from perfect straightness by less than one eighth of an inch. And yet, over all the circumference of this tremendous orbit, the sun exercises his dominion, and every pulsation of his surface receives its response from the subject earth. By observing the slight changes in the sun's ap- parent diameter, we find that its distance varies some- what at different times of the year, about 3,000,000 204830 38 THE SUN. miles in all ; and minute investigation shows tliat the earth's orbit is almost an exact ellipse, whose nearest point to the sun, or perihelion, is passed by the earth about the 1st of January, at which time she is 91,385,000 miles distant. The distance of the sun being once known, its di- mensions are easily ascertained at least, within certain narrow limits of accuracy. The angular semi-diameter of the sun when at the mean distance is almost exactly 962", the uncertainty not exceeding y^TF of the whole. The result of twelve years' observations at Greenwich (1836 to 1847) gives 961-82", and other determinations oscillate around the value first mentioned, which is that adopted in the " American Nautical Almanac." Taking the distance as 92,885,000 miles, this makes the sun's diameter 866,400 ; and the probable error of this quan- tity, depending as it does both on the error of the meas- ured diameter and of the distance, is some 4,000 or 5,000 miles; in other words, the chances are strong that the actual diameter is between 860,000 and 870,000 miles. Measurements made by the same person, however, and with the same instrument, but at different times, sometimes differ enough to raise a suspicion that the diameter is slightly variable, which would be nothing surprising considering the nature of the solar sur- face. There is no sensible difference between the equa- torial and polar diameters, the rotation of the sun on its axis not being sufficiently rapid to make the polar com- pression (which must, of course, necessarily result from the rotation) marked enough to be perceived by our present means of observation. It is not easy to obtain any real conception of the DISTANCE AND DIMENSIONS OF THE SUN. 39 vastness of this enormous sphere. Its diameter is 109*5 times that of the earth, and its circumference propor- tional ; so that the traveler who could make the circuit of the world in 80 days would need nearly 24 years for his journey around the sun. Since the surfaces of spheres vary as the squares, and bulks as the cubes, of i heir diameters, it follows that the sun's surface is near- ly 12,000 times, and its volume, or bulk, more than 1,300,000 times, greater than that of the earth. If the earth be represented by one of the little three-inch globes common in school apparatus, the sun on the same scale will be more than 27 feet in diameter, and its dis- tance nearly 3,000 feet. Imagine the sun to be hol- lowed out and the earth placed in the center of the shell thus formed, it would be like a sky to us, and the moon would have scope for all her motions far within the inclosing surface ; indeed, since she is only 240,000 miles away, while the sun's radius is more than 430,000, there would be room for a second satellite 190,000 miles beyond her. The mass of the sun, or quantity of matter con- tained in it, can also be computed when we know its distance, and comes out nearly 330,000 times as great as the earth. The calculation may be made either by means of the proportion given in the note to page 33, or by comparing the attracting force of the sun upon the earth, as indicated by the curvature of her orbit (about 0'119 inch per second), with the distance a body at the surface of the earth falls in the same time under the action of gravity, a quantity which has been determined with great accuracy by experiments with the pendulum. Of course, the fact that the sun produces its effect upon the earth at a distance of 93,000,000 miles, while a fall- ing body at the level of the sea is only about 4,000 40 THE SUN. miles from the center of the attraction which produces its motion, must also enter into the reckoning.* This mass, if we express it in pounds or tons, is too enormous to be conceived : it is 2 octillions of tons that is, 2 with 27 ciphers annexed ; it is nearly 750 times as great as the combined masses of all the planets and satellites of the solar system and Jupiter alone is more than 300 times as massive as the earth. The sun's attractive power is such that it dominates all surround- ing space, even to the fixed stars, so that a body at the distance of our nearest stellar neighbor, a Centauri, which is more than 200,000 times remoter than the sun, could free itself from the solar attraction only by dart- ing aw r ay with a velocity of more than 300 feet per sec- ond, or over 200 miles an hour ; unless animated by a greater velocity than this, it would move around the sun in a closed orbit an ellipse of some shape, or a circle with a period of revolution which, in the smallest possible orbit, would be about 31,600,000 years, and if the orbit were circular, would be nearly 90,000,000. We say it would revolve thus that is, of course, unless * The calculation of the sun's mass, from the data given, proceeds as follows : Let M = the sun's mass, and m that of the earth ; R = the dis- tance from the earth to the sun, and r the mean radius of the earth ; T, the length of the sidereal year, reduced to seconds ; and ^ g the distance a body falls in a second at the earth's surface. Now, the distance the earth falls toward the sun in a second, or the curvature of her orbit in a second, is equal to (about 0'119 inch). Hence, by the law of gravita- 2w*R m M tion, | ff : - T> = - : whence, M = m In this formula make v 3-14159; R, 92,900,000 miles; T = 31,- 558,149-3 seconds; r = 3,958-2 miles; and $g = 0-0061035 mile (16'113 feet), and we shall get the result given in the text, viz., M = 330,000 m (nearly). DISTANCE AND DIMENSIONS OF TIIE SUN. 41 intercepted or diverted from its course by the influence of some other sun, as it probably would be. And we may notice here that in many cases certainly, and in most cases probably, the stars are flying through space at a far swifter rate, with velocities of many miles per second. As for the attraction between the sun and earth, it amounts to thirty-six hundred quadrillions of tons : in figures, 30 followed by seventeen ciphers. On this point we borrow an impressive illustration from a care- ful calculation by Mr. C. B. Warring. We may imagine gravitation to cease, and to be replaced by a material bond of some sort, holding the earth to the sun and keeping her in her orbit. If now we suppose this con- nection to consist of a web of steel wires, each as large as the heaviest telegraph-wires used (No. 4), then to replace the sun's attraction these wires would have to cover the whole sunward hemisphere of our globe about as thickly as blades of grass upon a lawn. It would re- quire nine to each square inch. Putting it a little dif- ferently, the attraction between the sun and earth is equal to the breaking strain of a steel rod about 3,000 miles in diameter. If we calculate the force of gravity at the sun's sur- face, which is easily done by dividing its mass, 330,000, by the square of 109 (the number of times the sun's diameter exceeds the earth's), we find it to be 27 times as great as on the earth ; a man who on the earth would weigh 150 pounds, would there weigh nearly two tons ; and, even if the footing were good, would be unable to stir. A body which at the earth falls a little more than 16 feet in a second would there fall 443. A pendulum which here swings once a second would there oscillate more than five times as rapidly, like the balance-wheel of a watch quivering rather than swinging. 42 THE SUN. Since the sun's volume is 1,300,000 times that of the earth, while its mass is only 330,000 times as great, it follows at once that the sun's average density (found by dividing the mass by the volume) is only about one quarter that of the earth. This is a fact of the utmost importance in its bearing upon the constitution of this body. As we shall see hereafter, we know that certain heavy metals, with which we are familiar on the earth, enter largely into the composition of the sun, so that, if the principal portion of the solar mass were either solid or liquid, its mean density ought to be at least as great as the earth's ; especially since the enormous force of solar gravity would tend most powerfully to compress the materials. The low density can only be accounted for on the supposition, which seems fairly to accord also with all other facts, that the sun is mainly a ball of gas, or vapor, powerfully condensed, of course, in the central portion by the superincumbent weight, but pre- vented from liquefaction by an exceedingly high tem- perature. And, on the other hand, it could be safely predicted on physical principles that so huge a ball of fiery vapor, exposed to the cold of space, would present precisely such phenomena as we find by observation of the solar surface and surroundings. CHAPTEE II. METHODS AND APPARATUS FOR STUDYING THE SURFACE OF THE SUN. Projection of Solar Image upon a Screen. Carrington's Method of de- termining the Position of Objects on the Sun's Surface. Solar Pho- tography. Photoheliographs. Janssen's Photographs. Telescope with Silvered Object-Glass. Ilerschel's Solar Eyepiece. The Polar- izing Eyepiece. THE heat and light of the sun are so intense that peculiar instruments and methods are necessary for the observation of his surface. The appliances used in the study of the moon, planets, and stars will not answer at all for solar work. A very excellent method of proceeding where the object is to secure a general view of the sun, without regard to delicate detail, and to determine easily and rapidly the positions of spots and other objects on the sun's disk, is to project his image upon a sheet of card- board by means of a telescope. For this purpose things are arranged as indicated in the figure. The sheet of paper upon which the image is to be thrown is supported in front of the eyepiece by a light framework attached to the telescope. The dis- tance of the screen from the eyepiece depends upon the size of image desired and the power of the eyepiece ; a diameter of from six inches to a foot being generally most convenient. Another screen is usually fitted on 44 THK SUN. the object-glass end of the telescope to balance the first, and shade it from all light except that which has passed through the instrument. If the apparatus is to be used to determine the position of spots on the sun, the sur- face which receives the image must be carefully ad- justed so as to be perpendicular to the optical axis of the telescope. FIG. 7. To determine the position of objects on the sun's disk, Carrington used two lines, ruled at right angles to each other upon the screen, and set at an angle of about 45 with the north and south line or hour-circle. The observations needed to determine the place of a spot on the sun's disk then consist merely in noting with a watch as accurately as possible the four moments at which the edge of the sun's image crosses the two lines (the telescope being, of course, firmly fixed during the whole time), and the two moments when the spot passes METHODS FOR STUDYING THE SURFACE OF THE SUX. 45 them. From these six observations, with the help of the data given in the almanac, the distance and direc- tion of the spot from the sun's center may readily be calculated by formulae which would hardly be suited to these pages, but which may be found in the monthly notices of the Royal Astronomical Society, vol. xiv, page 153. Fig. 8 illustrates this arrangement. Another method, more convenient as involving no calculation, but less accurate, is to use Mr. A. Thom- son's " Charts for Sun-spot Observations," which are given in Sir Robert Ball's "Atlas of Astronomy" (Appletons, New York). Both these methods require, however, the use of a telescope eqnatorially mounted. With an instrument not so mounted, fairly good results may be obtained by drawing upon the screen a circle with a diameter about half that of the field of view, and noting the instants when the edge of the sun becomes tangent to the circle, and when the spot crosses it. With a small telescope thus fitted up, one is in a position to make observations of real value as to the number, position, and motions of the solar spots. Oc- casionally, also, when the air happens to be in good con- dition, a considerable amount of detail can be made out by this method in the spots and upon the solar surface generally. The darkening of the edge of the sun, 4({ THE SUN. caused by the absorption of the solar atmosphere, is very noticeable, and the faculse are conspicuous. One great advantage of the method is, of course, that several persons can thus observe together. A teacher, for in- stance, can in this way exhibit to a class of a dozen all the principal features of the sun's surface, and be sure that they all see the things he desires them to notice. Should any amateur happen to find upon the sun's disk a small, round spot, which he has reason to think is an intra-Mercurial planet, a few observations of the sort indicated above, repeated at intervals of some min- utes, would settle the question immediately, and give a reasonably accurate determination of the rate and direc- tion of movement. If the instrument has an equatorial mounting and clockwork, so that the image remains apparently sta- tionary upon the screen, a very satisfactory tracing can be made upon paper ruled in squares, showing pretty accurately the position and magnitude of all visible spots, in a form suitable to file away for reference. The observations of Carrington's great work upon the eolar spots were for the most part made in this manner. Of late years photography has been extensively util- ized for observations of this sort. The apparatus con- sists of a telescope fitted with a camera-box in place of an eyepiece, and with an arrangement for producing an instantaneous exposure of the sensitive plate to the solar rays. Since, in the ordinary achromatic telescope, the rays which are most effective in photographic action do not come to a focus at the same point as those which most strongly affect the eye, such an instrument, however perfect visually, will not give sharp photographic im- pressions. It is necessary, for the best photographic METHODS FOR STUDYING THE SURFACE OF THE SUN. 47 results, to use object-glasses whose corrections are cal- culated expressly for the purpose. Mr. Rutherfurd, of New York, seems to have been the first to appreciate this, and to construct an instrument specially designed for astronomical photography. To this end, disdaining all compromise, he did not hesitate to sacrifice delib- erately the visual excellence of an exquisite object-glass of thirteen inches diameter, by altering its curves so as to produce the most perfect actinic correction ; and he was rewarded by a success until recently unequaled as regards the perfection of the pictures obtained. Some of his photographs of the sun and moon, obtained about 1866, rival in sharpness and detail the drawings of ac- complished observers. Another and simpler method of obtaining the de- sired corrections, originally tried by Mr. Rutherfurd and rejected as not absolutely the best possible, has been revived and used by Cornu, of Paris. It consists, not in regrinding the two lenses which compose the object-glass, but merely in separating them slightly half an inch or so for an instrument of ten-feet focus. The approximate correction, thus produced, gives excel- lent results, and the instrument is not spoiled for other work, since it requires only a few minutes to restore the glasses to their visual adjustment. In a reflecting telescope there is, of course, no diffi- culty of this sort, since rays of different wave-length and color are not dispersed by reflection as by refraction. Other and still more serious difficulties, however, exist, depending upon the extreme sensitiveness of the reflec- tor to the distorting influence of variations of tempera- ture ; so that, hitherto, reflectors have not equaled re- fractors in the excellence of their photographic work. They have been employed with very good success, how- 48 THE SUN. ever, on several occasions for the photography of solar With telescopes of considerable size the picture is generally formed directly at the focus of the object- glass without further enlargement. This is the case with the pictures made by Mr. Rutherfurd, in which the 'diameter of the sun's image is about If inch. KEV? PlTOTOHELtOGRAPH. Copies of the negatives are afterward made if desired on a larger scale. In smaller instruments, such as the well-known photoheliograph of the Kew Observatory, an enlarging eyepiece is used, so constructed as to dis- tort as little as possible the image formed by the object- glass while magnifying it to a diameter of three or four METHODS FOR STUDYING THE SURFACE OF THE SUN. 49 inches. In this instrument, of which we give a figure, the diameter of the object-glass is only 3 inches, and its focal length 50 inches; the tube, instead of being conical as usual and larger at the object-end, is made pyramidal and larger at the bottom, in order to ac- commodate the plate-holder more conveniently. The whole is mounted equatorially, and driven by clockwork. It was constructed in 1857, under the directions and after the designs of Mr. De La Rue, and proved itself a most efficient arid excellent instrument. A number of other very similar instruments have since been made with slight improvements. Those employed by the English and Russian parties in their photographic opera- tions at the transit of Yenus in 1874 were of this type. So also were those of the German parties, except that they had considerably larger telescopes, with apertures of from six to eight inches. The photoheliograph now used at Greenwich in maintaining the daily record of the sun's surface is one of the Transit of Yenus instru- ments, having a four-inch object-glass, and giving a solar image four inches in diameter. It is mounted equato- rially with clockwork. The instruments at Mauritius and Dehra Dun are similar. More recently Greenwich has come into possession of a nine-inch photoheliograph, which is used in connection with the other to obtain pictures on a larger scale. The sunlight is so powerful that the exposure of the plate has to be made practically instantaneous. The apparatus by which this is effected varies greatly in de- tail in instruments of different types, but in all cases consists essentially of a slide, carrying in it a slit of adjustable width and capable of being shot across in front of the sensitive plate by a strong spring. At the moment of exposure a trigger or telegraphic key is 5 50 THE SUN. touched by the operator, and the slide, previously drawn back and locked by suitable mechanism, is released, and in its flight allows the rays to gleam through the aper- ture for a time, which in different instruments varies from j^-g- to g^oo of a second, according to the size of the instrument, the sensitiveness of the collodion, and the clearness of the atmosphere. We give a figure of Yogel's exposure-slide, which is perhaps as good as any. M is an electro-magnet, which, VOGEL'S EXPOSCHE-SLIDE. on the touch of a telegraph-key in the observer's hand, attracts the armature B, thus releasing the catch C, and allowing the spring S, by the intervention of the cord and pulley, to draw the slide containing the slit A swift- ly across the orifice through which the rays enter the camera. The character of the picture produced depends very greatly upon the proper timing of the exposure. If the intention be to secure an image of the sun with hard, firm edges from which measurements can be made to determine the position of objects on the solar disk METHODS FOR STUDYING THE SURFACE OF THE SUN. 51 as was the case at the transit of Yenus then a relatively long exposure is needed ; but it is to be remembered that the diameter of the sun's image increases very per- ceptibly with lengthening exposure, so that this diam- eter can never be safely used to furnish the scale of measurement. If, on the other hand, what is desired is a picture full of detail, showing the faculse and the structure of the spots, the exposure must be greatly shortened by narrowing the slit or giving the slide a greater velocity ; and it must be added, unfortunately, that the exposure which brings out perfectly the cen- tral portions of the disk is altogether too short for the portions near the limb, where the actinic power is very greatly diminished. This circumstance detracts considerably from the value of the photographic method. The skillful draughtsman can show in the same picture details dif- fering to any extent in intensity, while the photograph is, so to speak, limited to the reproduction of only one certain class of details at a time. Still we can always be sure that, whatever a photograph does show, is an autographic representation of fact, and not a figment of the imagination. This is not the case with drawings ; for it is remarkable how widely two conscientious artists will differ in their representations of the same object, seen by both with the same telescope, and under the same circumstances. As an accurate record of the num- ber, position, and magnitude of the solar spots at any given time, the photograph is, of course, unexception- able. Such a record was maintained by the Kew photo- heliograph for fourteen years from 1858 to 1872 when the work was discontinued. An almost equally important series of photographs was kept up for many 52 THE SUN. years at Wilna, in Kussia, until the burning of the ob- servatory in 1877. Since 1873 the Kew series has been continued at Greenwich, at least two pictures being taken every day when the weather will permit, and more than two if anything of special interest demands it. This Greenwich record is supplemented by the neg- atives taken at Dehra Dun, in India, and at Mauritius. Taken together with the Greenwich plates, they furnish a practically continuous record of the condition of the solar surface. At the same time there are occasional breaks, which might be remedied if we had one or two photoheliographs on the American side of the Atlantic. Of late, Janssen, at the new French physical ob- servatory at Meudon, has carried solar photography to a point far beyond any previous attainment. He has accomplished it mainly by utilizing the fact that there exists in the spectrum, near the Fraunhofer line G, a narrow band of rays which possess a photographic ac- tivity upon the salts of silver much more intense than that of any other portion of the spectrum. It is so intense, indeed, that if the exposure be very short and properly regulated, the effect is practically the same as if the sunlight were monochromatic, consisting of these rays alone : any defect in the color-correction of the object-glass is rendered almost harmless. This makes it possible to use an ordinary achromatic object-glass, roughly corrected for photographic work by merely separating the lenses a trine, according to Cornu's plan. With a five-inch telescope and a suitable enlarging lens, Janssen produces pictures even half a metre in diameter, and of extreme perfection in their delinea- tion of the details of the solar surface. The exposure, ranging from ^fa to y^Vjr of a second, according to the clearness of the air and the altitude of the sun, is effected METHODS FOR STUDYING THE SURFACE OF THE SUN. 53 by a slide closely resembling Yogel's. The impression obtained is very feeble, and requires prolonged and careful development ; but, when at last fairly brought out, is every way admirable. Some very interesting results, which we shall deal with later, have already been deduced from his plates. Photography, however, is not adequate as yet to the study of the most delicate details of the solar surface. For this purpose nothing can take the place of ocular observation by experienced and skillful observers, armed with powerful telescopes and suitable appliances, and on the watch for the few favorable moments when the atmospheric conditions will permit successful work. The instrument must be provided with some form of solar eyepiece expressly adapted to the purpose. The old-fashioned way was to use an ordinary eyepiece, fitted with a dark glass next the eye. If the whole aperture of a telescope of any size is used, the heat at the focus is so great as to endanger the lenses, and ac- cordingly it was customary to " cap down " the object- glass i. e., to put on a cover with a small hole in the center, so as to reduce the aperture to two or three inches. In this way, of course, the heat and light are easily diminished to almost any extent, but the defini- tion is greatly injured- According to well-known opti- cal principles, the image of a luminous point is not a point, even in an absolutely perfect telescope, but, in consequence of the so-called " diffraction " due to the interference of light, becomes a small disk, surrounded by a series of concentric luminous rings ; the smaller the aperture of the telescope, the larger the disk with a given magnifying power. Similarly, the image of a luminous line is not a line, but a stripe of determinate width with fringes on each side. It is easy to see, 54 THE SUN. therefore, that a telescope of small aperture can not possibly be made to show as delicate details as one of larger diameter, and, to get the best results in examin- ing the surface of the sun, we must find some way of diminishing the light and heat without cutting down the diameter of the object-glass (or mirror, if we are using a reflecting telescope). A reflecting telescope whose mirror is of unsilvered glass effects this very beautifully. The unsilvered sur- face reflects only about -fa of the incident light and heat, and although the resulting image is still too bright for the unprotected eye, the heat is not troublesome, and only a very thin shade-glass is needed. Another excellent method is to silver by Liebig's or some analo- gous process the front surface of the object-glass of a refractor. The silver film can be deposited of such a thickness as to allow any desired percentage of the light to pass, while the rest is reflected and not allowed to enter the instrument at all. The image formed in this way is slightly tinged with blue, but is beautifully sharp and steady, there being a great advantage in pre- venting the heating of the air in the telescope-tube, which occurs with every other form of instrument. The telescopes employed by the French parties in the observation of the transit of Yenus in 1874 were pre- pared in this way. With its great advantages, however, the method has on the whole quite as great disadvan- tages, as was evident at Saigon, where clouds were so thick that nothing could be seen through the silver film, and the observer had to rub it off with a cloth before he could do anything. Then, of course, a telescope pre- pared in this way can not be used for any other pur- pose. The common practice, therefore, is, not to adapt the instrument for solar observation by doing anything METHODS FOR STUDYING THE SURFACE OF THE SUN. 55 to its object-glass or mirror, but to accomplish the desired result by some modification or accessory of the eyepiece. One of the best known and most generally useful eyepieces is that devised by Sir John Herschel, and bearing his name. It is represented in Fig. 11, which gives a section of it. The light entering at O encoun- SOLAR EYEPIECE. ters a prism of glass, whose first surface is placed St an angle of 45. The greater part of the light, something over $, passes through the prism, emerging perpen- dicular to its second surface, and goes out through the open end of the tube ; the reflected light, about ^ of the whole, is thrown upward through the eyepiece proper, A B, which is precisely the same as ordinarily used. In this way most of the light and heat are got rid of; too much, however, still passes the lenses for the eye to bear, and it is necessary to use a shade- glass ; but this may be very light. The brightness of the sun varies so much at different altitudes and under different conditions of the atmosphere, that it is de- 56 THE SUN. sirable to have the thickness of the shade-glass adjust- able. This is easily managed by using a long, thin wedge of dark glass, compensated by a corresponding- wedge of ordinary glass, and set in a proper frame, as represented in Fig. 12. The shade-glass should not be r colored, but of neutral tint, so that objects on the sun's surface may be &een of their proper hue. The glass known as " London smoke " very nearly fulfills this condition, and with a shade of this material the appa- ratus is exceedingly satisfactory, and quite sufficient for all ordinary work. Still finer results, however, may be obtained with more complicated and expensive " helioscopes," as they are called, which by means of polarization reduce the light to such a degree that no shade is needed, and, moreover, enable us to graduate the light as we please by merely turning a milled head. There are several forms of the apparatus : we give a figure of one constructed by Merz, slightly modified,* which is perhaps as convenient and effective as any. The light entering at A first en- *The modification consists in substituting the prisms P 1 and P 2 for simple reflectors of black glass, which are very apt to be broken by the heat of the sun's image. METHODS FOR STUDYING THE SURFACE OF THE SUN. 5? counters the surface of a prism, P 1 , set at the polar- izing angle ; about -^f of the light passes through the prism, emerging perpendicular to its rear surface, and, being rejected: about -^ is reflected and polarized by the reflection. The reflected ray next strikes the sur- face of a second prism, P a , and here a considerable por- tion of the remaining light is thrown away. That which MEBZ'S HELIOSCOPE. is left is reflected into the upper portion of the eyepiece parallel to its original direction, through an opening in the top of the circular case in which the two prisms are mounted. The upper case is attached to the lower in such a manner that it can be turned around the line C D as an axis. It contains two plane mirrors of black glass, placed as shown in the figure. With things in 58 THE SUN. the position indicated, a beam of considerable strength would reach the eye at B so strong, in fact, as to be painful ; and the same would be the case if the upper piece were turned 180, bringing the mirrors into the position shown by the dotted lines, with the issuing ray in the prolongation of the incident. But, by turning the upper piece one quarter of a revolution, the issuing ray can be entirely extinguished, and, by turning it less or more than 90, the intensity of the light can be con- trolled at pleasure. As no shade-glass is used, every- thing is seen of its proper tint. Another advantage is, that there is no such disturbance of the orientation of the solar image as happens with every form of diagonal eyepiece. North, south, east, and west fall in their usual and natural places a matter of some importance as regards the convenience of observation. Still other forms of helioscopic eyepiece depending upon polarization have been devised by Secchi, Lang- ley, Christie, Pickering, and others, each with its own peculiar advantages ; our limits, however, forbid more extended treatment of the subject. We add merely that in some cases, as in the study of the internal struc- ture of sun-spots, it is found very advantageous to adopt the device of Dawes, and limit the field of view- by a minute diaphragm made by piercing a card or plate of ivory with a hot needle ; thus excluding the light from any portion of the sun's surface except that under immediate observation. CHAPTER III. THE SPECTROSCOPE AND THE SOLAR SPECTRUM. The Spectrum and Fraunhofer's Lines. The Prismatic Spectroscope ; Description of Various Forms and Explanation of its Operation. The Diffraction Spectroscope. The Concave-Grating. Analyzing and Integrating Spectroscopes. The Telespectroscope and its Ad- justment. The Spectrograph. Explanation of Lines in the Spec- trum. Kirchhoff's Researches and Laws. The Sun's Absorbing Atmosphere and Reversing Layer. Elements present in the Sun. Lockyer's Researches and Hypothesis. Basic Lines. Dr. H. Dra- per's Investigations as to thfi Presence of Oxygen in the Sun. Schuster's Observations. Effect of Motion upon Wave-Length of Rays and Spectroscopic Determinations of Motion in Line of Sight. EVER since the time of Newton it has been known that a beam of white light is decomposable into its con- stituent colors by passing it through a prism, and, under certain circumstances, the result is a rainbow-tinted band or ribbon, which has been called the solar spectrum. In this spectrum Wollaston, in 1802, discovered certain dark shadings, and in 1814 Fraunhofer again and inde- pendently discovered the same thing; and he so im- proved his apparatus and method of observation as to get not merely indefinite shadings, but clear, sharp lines, of which he made a map, assigning designations to many of the principal ones. Indeed, these markings of the solar spectrum bear his name to this day. He, however, could not account for them, further than to show that they did not originate in his instru- ment nor in the earth's atmosphere ; and it was not 60 THE SUN. until the publication of the researches of Kirchhoff and Bunsen, in 1859 and 1860, that the scientific world came to appreciate their meaning and importance. "We speak of the work of Kirchhoff and Bunsen as epoch-making, and such was certainly the case. At the same time the secret of the solar spectrum had been, in part at least, divined before by Stokes, Thomson, and Angstrom ; the latter especially, whose memoir, published in 1853, would certainly have obtained a high celebrity if it had appeared in French, English, or German, instead of Swedish. Swan and Zantedeschi had also given to the spectroscope nearly its present form ; and a number of other investigators, among whom Sir John Herschel, Wheatstone, Foucault, and J. "W. Draper deserve special mention, had each con- tributed something important to the foundations of the new science, for such it has proved to be. The study of spectra has opened a new world of research, and added some such reach to our physics and chemistry as the telescope brought to vision. Of course, any extended discussion of the instru- ments, principles, and methods of spectroscopy would be inconsistent with our limits : we can only treat the subject very briefly. First, then, as to the instrument. It consists usually of three parts : the collimator so called ; the light-ana- lyzing; apparatus, which is sometimes a prism or train of prisms, and sometimes a diffraction grating ; and the view-telescope. Figure 14 shows the construction of a single-prism spectroscope, and the course of the rays of light through it. The collimator is simply a telescope without an eyepiece, and having in the place of the eye- piece a narrow slit. This slit is placed exactly at the focus of the object-glass of the collimator, so that rays THE SPECTROSCOPE. from each point of the slit become parallel beams after passing the lens, and a person looking through the ob- ject-glass, at the slit, sees it precisely as if it were an object in the sky. Optically, the slit of the collimator ABRANGEMEXT OF PRISMATIC SPECTROSCOPE. is thus removed to an infinite distance ; while, mechani- cally, it is still at the fingers' ends, within reach of ma- nipulation and adjustment. The collimator, however, is not essential. Fraunhofer's work was all done with light admitted through a slit in the window-blind at a distance of twenty or thirty feet a much less con- venient arrangement, as is at once evident. The view-telescope, which, however, is no more essen- tial than the collimator, is usually a small telescope with an object-glass of the same size as that of the collimator, and magnifying from five to twenty times. Generally, the collimator and telescope of astronomical spectro- scopes are from three quarters of an inch to two inches and a half in diameter, and from six to forty inches long. The light, after passing the slit and object-glass of the collimator, next strikes the prism or grating, and these two things the slit and the prism or grating 62 THE SUN. are really all that is essential. In the case of a prism (which must be set with its refracting edge parallel to the slit) the rays are bent out of their course, as shown in the figure, and enter the view-telescope, placed at the proper angle to receive them. Suppose, now, for a moment, that the light admitted at the slit is strict- ly homogeneous say red. The eye at the view-tele- scope would then see a red image of the slit, corre- sponding precisely in form and proportions to the slit itself, widening or narrowing as the slit is acted on by its adjusting screw. If instead of a slit the open- ing had some other form, as an arc of a circle, a tri- angle, or a square, the image seen would imitate it, always having the same color as the light admitted. Suppose, again, that the light is not homogeneous, but consists of two kinds mixed together say red and yel- low. Viewing the slit directly, without the spectro- scope, one would only see a single orange-colored im- age ; but with the spectroscope one would see two widely separated images, one of them red, the other yellow. This is because the prism refracts the two kinds of light differently, so that after the rays have passed the prism they strike the object-glass of the view- telescope in different directions, and then make images in different places. If the light is composed not of two kinds only, but many, the images will be numerous, ranged side by side like the pickets of a fence ; and if, as in the case of a candle-flame, the light emitted con- tains an indefinite number of tints, then the slit-images, placed side by side, will coalesce into a continuous band of color. If, in the candle-light, certain kinds of light are specially abundant, then the corresponding slit-images will be more brilliant than their neighbors ; and if, as is usually the case, the slit be narrowed to a THE SPECTROSCOPE. 63 line, these slit-images will become bright lines in the spectrum lines only because the slit is itself a line, which, of course, is the best form to give the light-ad- mitting aperture, in order that the different images may overlap and interfere as little as possible. If any kinds of light be wanting, then the corre- sponding images of the slit will be missing, and the spectrum will be marked by dark bands or lines. SPECTROSCOPE. The cut (Fig. 15) shows the actual appearance of what is known as the chemical spectroscope, ordinarily used in laboratories. Besides the collimator A, and the telescope B, it has a third tube C, which carries a fine scale photographed on glass at the end farthest from the prism. There is a lens in the tube at the end next the prism, so that the observer at the telescope sees this scale running across the field of view at the edge of the spectrum, and thus has the means of. noting accurately 64 THE SUN. the position of any lines he may find. This arrange- ment is due to Bunsen. It is often desirable to obtain a greater separation of the different colors " dispersion," to use the technical term than a single prism would produce. In this case, the rays after passing through the first prism may be transmitted through a second and a third, and so on, until they reach the view-telescope. With prisms as commonly made, it is difficult to use more than six in this manner, but it is possible by reflection properly managed to return the rays through a second prism- train connected with the first, so as to get the virtual effect of from ten to twelve prisms. The instrument figured on pages 78 and 190, and used for observation of the solar prominences, is of this kind. COMPOUND PRISM. DIRECT-VISION PRISM Another way is to use a compound prism, so called, composed of a very obtuse-angled prism, A B E, of some highly dispersive material, usually heavy flint glass, flanked by two prisms of lighter glass with their refracting angles reversed. Prisms of this kind THE SPECTROSCOPE. 65 can be made of much higher dispersive power than simple prisms, and of course a smaller number will answer the same purpose. By properly proportion- ing the angles C A E and E B D, it is possible to make the yellow rays of the spectrum pass through without change of direction, while still retaining a considerable dispersion. An instrument with prisms of this kind is called a " direct-vision " spectroscope, and in some cases is much more convenient than the other forms. Thollon has recently constructed compound prisms having the dense glass prism replaced by a chamber filled with carbon disulphide, which possesses an enor- mous dispersive power ; with a train of these prisms he has obtained views of the spectrum only equaled by the performance of the best diffraction gratings. A dis- persion equal to that of thirty or forty prisms of an ordinary spectroscope is easily reached. The behavior of these disulphide prisms is, however, far from satis- factory for general work, since they are extremely sen- sitive to small changes of temperature, which cause irregular refractions in the liquid, and destroy the defi- nition. We have used the expression, the dispersive power of thirty or forty prisms ; but that is very indefinite, because the dispersive power of a spectroscope depends upon its linear dimensions as well as the kind and num- ber of prisms and is proportional to the dimensions. That is to say, if a given spectroscope has the size of its prisms, and the diameter and focal lengths of its col- limator and telescope doubled, retaining, however, the former slit and eyepiece, its dispersive power will be doubled by the change. Thus a large single-prism in- strument may equal in working power a small one of many prisms. QQ TUB SUN. Lord Rayleigh lias shown that the resolving power of a spectroscope, constructed with prisms of a given substance, depends upon the length of the route pur- sued by the rays of light in traversing them. As has been said, a diffraction grating may replace the prism in a spectroscope. This diffraction grating is merely a system of close, equidistant, parallel lines ruled upon a plate of glass or polished speculum-metal. The closer the lines, the greater the dispersion pro- duced ; the larger the ruled surface, the more light is at the observer's disposal, provided the collimator and view-telescope are large enough to utilize the whole ruling. The greater the total number of lines, the higher the resolving power of the grating, or power of separating close lines in the spectrum. It hardly needs to be said that the making of a sat- isfactory grating is by no means an easy matter. To work a surface optically accurate, and to rule it with perfectly straight equidistant paralled lines, 20,000 to the inch or so, and all of the same width and depth, is one of the most delicate and difficult of mechanical opera- tions. The first that were fit for spectroscopic use were made in this country, about 1S71, by Mr. Eutherfurd, of New York, upon a ruling machine devised and con- structed by him for the purpose : they were first act- ually applied in solar spectroscopy by the writer, in 1873. In 1881, when the first edition of this book ap- peared, some very fine ones had already been produced, with ruled surfaces nearly two inches square, carrying 17.280 lines to the inch. One of these, still in constant use in the Princeton Observatory, is especially excellent, and except in size is hardly inferior to the magnificent specimens now produced at Baltimore by Professor Rowland's wonderful machine, which at present, with THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 67 its newest improvements and refinements, is quite with- out a rival, and almost ideally perfect. Ever since 1882 it has been turning off gratings of admirable excel- lence : the largest have a ruled surface of about 5 by 4 inches, with more than 100,000 lines (20,000 to the inch). These have been widely distributed among sci- entific workers, and it is not too much to say that all the recent important researches upon the solar spec- trum (Thollon's alone excepted) owe their success to Kowland's gratings. An explanation of the way in which diffraction spectra are produced by a grating lies beyond our scope ; for this the reader is referred to any good trea- tise on optics. We merely state, in passing, that diffrac- tion has nothing to do with refraction, but depends upon the fact that the ether waves, of which light con- sists, under certain circumstances " interfere " with each other and produce brilliant effects of color. We say spectra, because, while a prism gives but one spectrum, a grating gives many, and of different degrees of dis- persion, which is often a matter of much convenience. Of course, no one of the spectra is as brilliant as if it were the only one, but with sunlight this is a matter of small consequence. Besides, it is possible, by giving the proper shape to the diamond point which rules the lines and properly regulating the depth of cut, to produce gratings in which most of the light shall be concen- trated in a single one of the spectra at the expense of the rest. A good grating combined with a suitable collimator and telescope constitutes a spectroscope which for most solar work is incomparably superior in power and con- venience to any prismatic instrument of similar dimen- sions, so that as a matter of fact the grating-spectro- 68 IDE SUN. scope lias almost superseded the other in this sort of work. Fig. 17 shows the arrangement of the different parts of such an instrument. The collimator and view-tele- DIFFB ACTION SPECTROSCOPE. scope are placed with their object-glasses close together, and their tubes making as small an angle as possible, consistently with keeping the grating at a manageable distance, since both collimator and telescope must be pointed at the center of the grating. The grating is mounted on a frame with an axis at A, so that it can rotate in the plane of the dispersion, the ruled lines being parallel to this axis. The frame which carries the grating must be so constructed as to support it steadily and firmly, without the slightest strain, for it is essential to its good performance that the surface be strictly plane : an abnormal pressure of a single ounce at one of the corners will sensibly affect its perform- ance, and four ounces would bend the plate sufficiently to ruin the definition. As the different orders of spectra overlap each other (the red end of the second order spectrum over- lapping the blue of the third, etc.), it is sometimes neces- sary to separate them, and this can be done in a man- ner first suggested by Fraunhofer, by interposing between THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 69 the grating and view- telescope a single prism with its plane of dispersion perpendicular to that of the grating, the telescope being then inclined at the proper angle to receive the rays. A direct-vision prism in the eyepiece answers the same purpose, though less satisfactorily. In many cases a suitably colored shade-glass is sufficient. Fig. 18 is from a photograph of an instrument act- ually in use at Princeton for observations upon solar prominences. It is designed to be attached to a nine- inch equatorial, its collimator and view-telescope being each only about thirteen inches long, with a diameter of an inch and a quarter. The prism, P, is used only FIG. 18. PRINCETON SPECTROSCOPE. occasionally and can easily be removed. The telescope, T, is then lowered to the same level as the collimator, so as to be perpendicular to the lines of the grating. There is, however, a form of grating-spectroscope TO THE SUX. invented by Professor Rowland which uses not a flat but a concave grating, and dispenses with both the col- liraator and the telescope. For certain researches, such as the mapping of the spectrum of the sun or of metal- lic vapors, or for comparing together different spectra, CONCAVE GRATING-SPECTROSCOPE. it is the most powerful and effective of all spectroscopic apparatus. The arrangement is that indicated in Fig. 19. The grating, G, is mounted at one end of a stiff bar, C, at the other end of which is placed an eyepiece. THE SPECTROSCOPE AND THE SOLAR SPECTRUM. ft Two pivots, one with its center directly under the -cen- ter of the ruled surface of the grating, and the other at I, distant from the first by the radius of the spherical surface of the grating, connect the rod with two carriages which ride upon the rails R and E/. These are firmly secured exactly at right angles to each other upon the two strong beams A and B, and at the point where the rails would meet is placed the slit S. With this arrangement, if I is made to slide along B either toward the right or left, G will slide along A ; and since the angle at S is a right angle, the three points, G, S, and I, will always be on the circumference of a circle whose diameter is G I. Under these conditions light passing through the slit and striking the grating will form a perfectly focused spectrum at I, which can be viewed with the eyepiece. If desired, a photographic plate can be placed at I and the spectrum photographed. If I is moved toward the right it will run toward the red end of the spec- trum, and toward the violet if moved toward the left. With a six-inch grating the rod G I is usually from fifteen to twenty -five feet long, and the dispersion is tremendous. The apparatus is mounted in a large room perfectly darkened, and the sunlight is brought in by a heliostat mirror, through a suitably protected orifice. It was with an apparatus of this kind that Professor Rowland made his great photographic map of the solar spectrum (page 78), and has studied the spectra of nearly all the chemical elements. The theory of the instrument is quite beyond our range : those who are sufficiently versed in mathematics will find it given in the " Encyclopaedia Britannica," article " Wave-theory of Light," 14. The prismatic and diffraction (or interference) spec- 72 THE SUN. tra differ from each other to a certain extent, not, of course, in the order of colors or of lines, but in their relative distances. In the prismatic spectrum the red and yellow portion is much compressed, while the violet is greatly extended ; with the diffraction spectrum the reverse is the case ; the lines in the violet are crowded together, and those in the red are widely separated. In the diffraction spectrum the lines are almost per- fectly straight ; in the prismatic, generally inore or less curved ; we say generally, because there are forms of high-dispersion spectroscope in which this curvature is corrected. This curvature is caused by the fact that the rays from the top and bottom of the slit do not meet the refracting surface at the same angle as those from the middle of the slit ; they are, therefore, differ- ently refracted ; in consequence, the slit-images of which the spectrum is built up are not straight, but distorted. We may add that the dark lines which often run lengthwise through the spectrum are merely due to roughnesses or particles of dust on the jaws of the slit. It is almost impossible to make and keep the edges of the slit so clean and smooth that lines of this sort will not appear when the opening is very narrow. The spectroscope may be used in two entirely dif- ferent ways: it may simply have its collimator pointed toward the source of light ; or a lens may be interposed between the slit and the luminous object, so as to form 'an image of the latter on the slit. In the first case, the instrument is said to be an in- tegrating spectroscope, because each point in the slit receives light from the whole of the luminous object, so that the spectrum is alike through its whole width, arid represents the average light of the object it lumps the THE SPECTROSCOPE AND THE SOLAR SPECTRUM. ^3 whole, so to speak. In the second case, different parts of the slit are illuminated by light from different parts of the object ; the top of the slit gets the light from one point, the middle of the slit from another, and the bottom from a third. If, then, the lights emitted by the three points differ, their spectra will differ also, and the observer will find that different portions of the width of his spectrum will differ correspondingly the upper portion will be unlike the middle, and the mid- dle will differ from the bottom. An instrument ar- ranged thus is called an analysing spectroscope, because it enables us to determine separately the spectra of vari- ous portions of an object, and thus to analyze its consti- tution ; as, for instance, a sun-spot and its surroundings. For most purposes, especially astronomical, it is much the most satisfactory. Approximately the same end may be reached, in some cases, by placing the slit very near the luminous object, as in flame analysis, but it is usually much more convenient and better to use the lens. In astronomical work the object-glass of a large equatorial telescope is generally employed to form the image of the celestial object, and the spectroscope is attached at the eye-end of the telescope, the eyepiece being removed. The combined instrument is then often called a tele-spectroscope. Fig. 20, on the next page, represents the apparatus long used at the Dartmouth College Observatory. It is usually very important that the slit of the in- strument be precisely in the focal plane of the object- glass of the telescope for the rays especially under ex- amination. On account of the so-called " secondary spectrum" of the achromatic lens, this focal plane is quite different for the different colors, and the spectro- scope requires to be slid in or out, so as to vary the THE SUN. distance of its slit from the great object-glass of the telescope according to circumstances. The same end may be obtained (less satisfactorily, however) by a sec- ond lens between the object-glass and the slit, and pretty near the latter. By moving this lens, the focus FIG. 20. TELK-SpECTEOSCOrE. can be made to fall exactly on the slit. Neglect of this adjustment will make many of the most interesting and important spectroscopic observations quite impossible. As has been already mentioned, photography is often used in connection with the spectroscope, and the name THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 75 spectrograph has been given to the combination instru- ment. Fig. 21 represents the spectrograph which is attached to the 23-inch equatorial of the Halsted Ob- servatory at Princeton. As shown in the figure, it uses a grating, but this can, at pleasure, be exchanged for a train of prisms. Some of the advantages of photography are self- evident as, for instance, the quickness and accuracy with which a map of any portion of the spectrum can be produced, as contrasted with the tediousness of any drafting process. Then, too, since our modern plates admit of any desired length of exposure (instead of dry- ing up in a few minutes, as our old " wet plates " used to), they enable us to get satisfactory negatives of spec- tra far too faint to be visually observed. Finally, there is a long range of ultra-violet spectra composed of rays of too short wave-length and too high a " pitch " to be seen by the human eye, while they are easily photo- graphed. Per contra, the ordinary photographic plate is by no means impartial it is very sensitive to blue and purple, and very obtuse to green, yellow, and red. The new isochromatic and orthochromatic plates are, how- ever, better in this respect, and by using them (com- bined with a reasonable amount of patience) it is now possible to work down even to the red extremity of the visible spectrum. If the collimator of a spectroscope of any form be directed toward an ordinary lamp, or upon the incan- descent lime of a calcium-light, the observer will get simply a continuous spectrum ; a band of color shading gradually from the red to the violet, without markings or lines of any kind. If the instrument be turned to- ward the sun he will obtain something much more in- LAKGK PRINCETON SPECTROSCOI-E (BITTED FOB PHOTOUKAPII THE SPECTROSCOPE AND THE SOLAR SPECTRUM, ff teresting a band of color, as before, but marked by hundreds and thousands of dark lines, some fine and black, like hairs drawn across the spectrum, while others are hazy and indistinct. Most of them retain their appearance and position perfectly from day to day ; some of them, however, are more intense at one time than another, and when the sun is near the horizon certain lines in the red and yellow become extremely conspicuous, in such a way as to make it clear that they, at least, have something to do with our terrestrial atmosphere. Fig. 22 is a reproduction of a portion of Fraunhofer's map of the solar spectrum, showing what one might fairly expect to see (except as to color) with an excellent single-prism spectroscope. Fig. 23 is a drawing of a very small portion of the spectrum in the green, as shown by a very powerful spectroscope. The scale is that of Angstrom's map. The large, heavy lines are known as the little l> group, and are due, as we shall soon see, part of them to the presence of iron and nickel, and part to magnesium, as gases in the solar atmosphere. SPECTRUM MAPS. There are numerous maps of the solar spectrum : the earliest of any scientific value was Kirchhoff's, which 78 THE SUN. appeared in 1861-'62. Its scale was purely arbitrary, and not even self-consistent throughout, so that when Angstrom's map of the " normal spectrum " was pub- 5 GROUP IN SOLAR SPECTRUM. lished in 1868, made with a diffraction grating, and platted on a scale of wave-lengths (one unit of the scale corresponding to one ten-millionth of a millimetre of wave-length), it soon superseded Kirchhoff's, and has not yet ceased to be used for reference. Angstrom's grating was imperfect, however, and at present Kow- land's photographic map, dating from about 1890, is accepted as " the standard." It covers the ultra-violet from about X* 3000, and extends down through the visible spectrum into the red to X 6900, just below the B line. It is unfortunate that it does not" go lower, but Mr. Higgs, of Liverpool, has published a set of photo- graphs of different parts of the spectrum, and two of these complete Eowland's map to the lower limit of * X, is the symbol now universally used for the " wave-length " of a ray of light. " \ 3000 " means that part of the spectrum (invisible in this case) where the wave-length is 3000 t en-mil lionths of a millimetre. THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 79 photographic capability. lu clearness and beauty of execution Mr. Higgs's maps surpass everything that has been done in that line. Thollon's great map cov- ers only the lower half of the spectrum, and is subject to the same objection as that of Kirchhoff an arbitrary scale. Its peculiarity is in presenting the appearance of the spectrum corresponding to different altitudes of the sun. As there are invisible rays beyond the upper or violet end of the visible spectrum, so also below the red there is a still longer range of rays of wave-length too great for the human eye to perceive. Photography carries us a little way into this infra-red region, but not far. For most of our knowledge of it we are dependent upon the " bolometric " work of Professor Langley, to which we shall have to refer more specially in treating of the heat of the sun. He has already succeeded in mapping this " heat-spectrum," as it is sometimes called, sufficiently to show that it is filled with dark bands and lines, and has fixed the position of many of them. If, instead of using the sun or an ordinary flame for the source of light, we examine with the spectroscope an electric spark, or the arc between carbon points, or the light produced by passing the discharge of an induc- tion coil through a rarefied gas, we shall get a spectrum of quite a different sort a spectrum consisting of bright lines upon a dark or faintly luminous background ; and it will be foiTnd that the spectrum developed will al- ways be the same under similar circumstances, depend- ing mainly upon the material of the electrodes (the points between which the discharge passes), and the nature of the intervening gas, but also, to a certain extent, upon its density and the intensity of the elec- tric discharge. So, also, if certain easily vaporized salts 80 THE SUN. are introduced into the blue flame of a Bunsen gas-burn- er, or of a spirit-lamp even, the flame becomes colored, and its spectrum is a spectrum of bright lines, which are perfectly characteristic of the metal whose salt is used. An ordinary candle-flame, indeed, almost always shows one such bright line in the yellow, as had been noticed many years before Swan, in 1857, showed it to be due to the presence of sodium, which in the form of common salt is universally distributed. Fraunhofer, as early as 1814, had discovered that this line (or lines rather, for it is really composed of two, easily separated by a spectroscope of no great power) exactly coincides with the double line which he named D, in the solar spactrum ; and he had found the same line in the spectra of certain stars also ; but he did not know that the line was due to sodium, or in all probability he would have anticipated by nearly half a century the discovery which lies at the foundation of modern spectrum analysis. As has been said before, the principles involved seem to have been more or less distinctly apprehended by several persons Stokes and Angstrom especially years before the publication of Kirchhoif in 1859 ; but it was KirchhofTs work which first bore fruit. It is not necessary to repeat here again the oft-told story how he found that, when sunlight is made to pass through a flame containing sodium-vapor, the D- lines in the spectrum of this sunlight come out with increased intensity ; though, when a screen is interposed between the sun and the flame, the lines are bright, as usual in such a flame. He found, too, that when the incandescent lime-cylinder of the calcium-light is placed behind the sodium-flame, a precisely similar phenomenon occurs, and the bright lines of the flame-spectrum are THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 81 reversed to dark ones.* He found the same thing to hold good also for a flame colored by lithium. The sum of his results may be stated as follows : 1. Solids and liquids, when incandescent, give con- tinuous spectra ; and, as we now know, the same thing is true of gases also at great pressures. 2. Bodies in the gaseous state (and not compressed) give discontinuous spectra consisting of bright lines and bands; and these bright-line spectra are different for different substances, and characteristic, so that a given substance is identifiable by its spectrum. .3. When light from a solid or liquid incandescent body passes through a gas, the gas absorbs precisely those rays of which its own spectrum consists ; so that the result is a spectrum marked by black lines occupy- ing exactly the same positions which would be held by the bright lines in the spectrum of the gas alone. His conclusion, therefore, was that the luminous sur- face of the sun (the " photosphere ") is composed of solid or liquid matter giving by itself a purely contin- uous spectrum, and that the dark lines which mark the spectrum are produced by the transmission of the light through an overlying atmosphere. He believed the photosphere to be a continuous sheet of liquid a mol- ten ocean but numerous facts since learned make it almost certain that it is rather a sheet of " cloud," com- posed of minute drops or dust floating in the lower re- gions of the solar atmosphere. * The blackness of the lines formed in this way is such that it is some- times difficult to believe, what is really the fact, that they are actually brigJder than they were before the lime-cylinder was placed behind the flame, and that their darkness is only apparent, and due to their contrast with the more brilliant background of the continuous spectrum of the incandescent lime. It is very easy, however, to demonstrate the truth by a simple experiment. 7 82 THE SUN. If, then, sodium is present in the solar atmosphere between us and the photosphere, we ought to find in the solar spectrum those lines dark which are bright in the spectrum of sodium-vapor ; and we do. If mag nesium is there, it ought to manifest itself in the same way, and it does ; and similarly for all the substances which spectrum analysis reveals. If this view is correct, it follows also that this atmos- phere, containing in gaseous form the substances whose presence is manifested by the dark lines of the ordi- nary spectrum the sun's reversing layer, as it is now often called would give a spectrum of bright lines if we could isolate its light from that of the photosphere. The observation is possible only under peculiar circum- stances. At a total eclipse of the sun, at the moment when the advancing moon has just covered the sun's disk, the solar atmosphere of course projects somewhat at the point where the last ray of sunlight has disap- peared. If the spectroscope be then adjusted with its slit tangent to the sun's image at the point of contact, a most beautiful phenomenon is seen. As the moon ad- vances, making narrower and narrower the remaining sickle of the solar disk, the dark lines of the spectrum for the most part remain sensibly unchanged, though becoming somewhat more intense. A few, however, begin to fade out, and some even turn palely bright a minute or two before the totality begins. But the mo- ment the sun is hidden, through the whole length of the spectrum, in the red, the green, the violet, the bright lines flash out by hundreds and thousands, almost startlingly ; as suddenly as stars from a bursting rocket- head, and as evanescent, for the whole thing is over within two or three seconds. The layer seems to be only something under a thousand miles in thickness, and the moon's motion covers it'very quickly. THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 83 The phenomenon, though looked for at the first eclipses after solar spectroscopy began to be a science, was missed in 1868 and 1869, as the requisite adjust- ments are delicate, and was first actually observed only in 1870. Since then it has been more or less perfectly seen at every eclipse. Except at an eclipse it has not yet been found possible to observe this bright-line spec- trum, because it is overpowered by the aerial illumina- tion of our own atmosphere. It is not, however, to be understood that the dark lines of the solar spectrum are due entirely or even principally to the stratum of gas which lies close above the surface of the photosphere. Were this so, the dark lines should be much stronger in the spectrum of light from the edges of the disk than in that from the center, which is not the case ; at least, the difference is very slight. The photosphere, as we shall see here- after, is probably composed of separate cloud-like masses floating in an atmosphere containing the vapors by whose condensation they are formed ; the principal ab- sorption, therefore, probably takes place in the inter- stices between the clouds, and below the general level of their upper limit. The beautiful observations of Professor Hastings, of New Haven, in which by an ingenious contrivance he managed to confront and compare directly the spectra of light from the center and edges of the sun's disk, have brought out the facts in the case very finely. Theoretically, then, it is very easy to test the ques- tion of the presence of an element in the sun. It is only necessary to cover one half the length of the spec- troscope-slit with a mirror or prism by which the sun- light is directed into the instrument, while at the same time a flame or electric spark, giving the spectrum of the substance under investigation, is placed directly in THE SUN. 84 front of the other half of the slit. When matters are thus arranged, the observer sees in the instrument two spectra in juxtaposition, each of half the usual width one the solar spectrum, the other that of the element under investigation ; and it is easy to see whether the bright lines of the elementary vapor match exactly with corresponding dark lines in the solar spectrum. ACTION OF THE COMPAEISON- PBISM. COMPARISON PRISM TEOSCOPE. The figures show the usual arrangement of the com- parison-prism, as it is ordinarily called. For the examination of the upper or violet portion of the spectrum, photography is employed with great advantage, the arrangement being precisely the same as that just indicated, except that a sensitized plate takes the place of the human retina, and the impression can be permanently retained for leisurely study. Certain light, too, as every one knows, which is invisible to the eye, strongly affects the photographic plate, so that the comparison can by this means be carried on into the ultra-violet regions of the spectrum. The following full-page illustration is a representa- tion of the arrangement of apparatus used by Mr. Lock- yer in his celebrated researches it is taken from his " Studies in Spectrum Analysis." THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 85 Theoretically, we say, the comparison is easy ; but the practical difficulties are considerable. In the first place, it is not easy to get a spectrum of the body you wish to study, free from lines belonging to other sub- stances the requisite chemical purity is very trouble- some to attain ; and, in the next place, the dark lines of the solar spectrum are so numerous that it requires a very high dispersive power to establish a coincidence with certainty ; a bright line in the spark-spectrum may fall very near a dark line with which it has no connec- tion whatever. When, however, as in the case we have mentioned, the coincidences are not one or two, but numerous, and the lines in question peculiar in their character and appearance, a satisfactory result is soon established. It was in this manner (by comparisons made by the eye and not by photography) that Kirchhoff in 1860 de- termined the presence in the solar atmosphere of the following elements : sodium, iron, calcium, magnesium, nickel, barium, copper, and zinc, the last two rather doubtful at that time. Since then the list has been greatly extended, and in 1891 stood as follows, accord- ing to Professor Rowland, who has been making a most thorough reinvestigation of the whole subject. He has worked with concave-grating apparatus of tlie highest power, and has made his comparisons between the spec- trum of the sun and the spectra of the chemical ele- ments by means of photography. He has used, however, only the electric arc, and not the spark, in producing the spectra, and it is to be hoped that this research may be supplemented by an equally thorough study of spark-spectra. Since the work is not even yet com- plete (1895), the list is to be regarded as provisional only. TUE SPECTROSCOPE AND THE SOLAR SPECTRUM. 87 ELEMENTS IN THE SUN, ARRANGED ACCORDING TO THE NUMBER AND INTENSITY OF THEIE DARK LINES IN THE SOLAR SPECTRUM. Intensity. Number. 1. Calcium. Iron (2,000 or more). 2. Iron Nickel. 3. Hydrogen. Titanium. 4. Sodium. Manganese. 6. Nickel. Chromium. 6. Magnesium. Cobalt. 7. Cobalt. Carbon (200 or more). 8. Silicon.f Vanadium. 9. Aluminium, f Zirconium. 10. Titanium. Cerium. 11. Chromium. Calcium (75 or more). 12. Strontium. Ncodymium. 13. Manganese. Scandium. 14. Vanadium. Lanthanum. 15. Barium. Yttrium. 16. Carbon. f ? Niobium. 17. Scandium. f Molybdenum. 18. Yttrium. Palladium. 19. Zirconium.f Magnesium (20 or more). 20. Molybdenum. -j- Sodium (11). 21. Lanthanum. Silicon. 22. Niobium, f Hydrogen. 23. Palladium.! Strontium. 24. Neodymium.f ? Barium. 25. Copper, t Aluminium (4). 26. Zinc. Cadmium. 27. Cadmium. Rhodium. 28. Cerium. Erbium. 29. Glucinum.t Zinc. 30. Germanium, f Copper (2). 31. Rhodium.f Silver. 32. Silver. Glucinum. 33. Tin. Germanium. 34. Lead. Tin. 35. Erbium. Lead (1). 36. Potassium, f Potassium. 88 THE SUN. DOUBTFUL ELEMENTS. Iridium. Osmium. Platinum. Ruthenium. Tantalum. Thorium. Tungsten. Uranium. ELEMENTS NOT APPEARING IN THE SOLAR SPECTRUM. Antimony. Arsenic. Bismuth. Boron. Caesium. Gold. Indium. Lithium. Phosphorus. Rubidium. Selenium. Mercury. Thallium. Prseseodymium. Nitrogen (vacuum-tube). Sulphur. SUBSTANCES NOT YET TRIED (fiY ROWLAND). Bromine. Chlorine. Fluorine. Iodine. Oxygen. Gallium. Holmium. Tellurium. Terbium. Thulium, etc. Professor Rowland remarks that several of the ele- ments are classified as " not appearing in the solar spec- trum" merely because their arc-spectra show very few strong lines, or none at all, within the limits of the solar spectrum (it might be different with their ^a?^- spec- tra). He adds, what can not too firmly be borne in mind, that the failure to find them " is very little evidence of their absence from the sun itself," and that if the whole earth "were heated to the temperature of the sun, its spectrum would probably resemble that of the sun very closely." Besides these substances which reveal their presence in the sun by dark lines in its spectrum, there are at least two others, helium* and coronium, as they are provisionally named, which show themselves only by bright lines in the spectrum of the chromosphere and the corona, with which we shall deal later. In 1895 helium was at last identified by Ramsay, in connection with his researches upon argon, the new component of our atmosphere. He found the lines of helium in gas * Sec note on page 344. THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 89 disengaged from uraninite and other minerals, where it is associated with the so-called u rare earths." Coro- nium still remains undetermined. All of the above-named elements, except those marked with a f, are represented at times by bright lines in the spectrum of the chromosphere, which will be dis- cussed in another chapter ; and strontium and cerium were observed in that manner by the writer before the coincidence of their lines with dark lines in the ordinary solar spectrum had been satisfactorily made out. As to carbon, the characteristic groups of lines which mark the visible portion of its spectrum are only doubtfully present ; but in the ultra-violet the photo- graphs of Mr. Lockyer have brought out other groups which belong to this element, and the presence of this element has since been abundantly confirmed by Row- land and others. Thus far the most careful observation fails to find, either in the ordinary spectrum or in that of the chro- mosphere, the slightest trace of bromine, chlorine, iodine, or nitrogen, of arsenic, boron, or phosphorus ; of sulphur there are merely doubtful indications in the chromosphere spectrum ; and as regards oxygen, the evidence, on the whole, is against its presence, though the case is peculiar. When we recollect that the non-apparent elements constitute a great portion of the earth's crust, the question at once forces itself, AVhat is the meaning of their seeming absence ? Do they really not exist on the sun, or do they simply fail to show themselves ; and, if so, why ? The answer to the question is not easy, and astronomers are not agreed upon it, though we im- agine that most of them would prefer the latter alterna- tive. Even under the conditions of our terrestrial 90 THE SUN. laboratories we find cases where, when several gases and vapors are mingled at a high temperature, certain ones only of those present appear in the spectrum of the mixture, the others giving no indication of their pres- ence. Then, too, it is now certain that the same sub- stance under diifering conditions may give two or more widely differing spectra ; it is easy to admit, therefore, that we may be unable to reproduce in the electric arc the spectrum of a substance that characterizes it in the sun, and so may fail to identify it. Possibly, in some cases, the very brilliance of the lines of an element may prevent their appearance as dark lines. It is pos- sible, for instance, to make the bright lines of sodium so intense that the light from an incandescent lime- cylinder will not be able to reverse them, and, of course, by making them a little less intense, they may be caused to disappear entirely, being neither brighter nor darker than the continuous spectrum on which they are pro- jected. This may perhaps actually be the case with helium, which gives in the chromosphere spectrum an intensely brilliant yellow line, known as D 3 , because it is very near to the sodium lines, D and D a . At times, and especially in the neighborhood of sun-spots, a very faint dark line marks its place, but usually the spec- trum of the photosphere fails to give the slightest in- dication of its presence. There are, however, in the chromosphere spectrum fifteen or twenty other bright lines (but fainter than D,) which have no dark cor- relatives. Most of these are now known to be also due to helium, and this makes it more likely that the absence of dark lines is accounted for either by the thinness of the helium layer or the intensity of its tem- perature. Nitrogen and hydrogen each have two spectra, one THE SPECTROSCOPE AND THE SOLAR SPECTRUM.- 91 a spectrum mostly composed of shaded bands, while the other consists of sharp, well-defined lines. Oxygen, ac- cording to Schuster's careful researches, has four spec- tra, and carbon is also assigned four by its investigators. There appear to be various possible explanations of these facts. One is, to suppose that the luminous sub- stance, without any change in its own constitution, vi- brates differently and emits different rays under varying circumstances, just as a metal plate emits various notes according to the manner in which it is held and struck. A second assumes that the substance, without losing its chemical identity, undergoes changes of molecular struct- ure (assumes allotropic forms) under the varying cir- cumstances which produce the changes in its spectrum. According to either of these views, although we can safely infer, from the presence of the known lines of an element in the solar spectrum, its presence in the solar atmosphere, we can not legitimately draw any negative conclusion : the substance may be present, but in such a state under the solar conditions as to give a spectrum different from any with which -we are acquainted. Still a different explanation is to suppose, with Mr. Lockyer, that the changes in the spectrum of a body are indications of its decomposition, the spectrum of the original substance being replaced by the superposed spectra of its constituents, so that the absence of the missing substances is real, being due simply to the fact that the solar atmosphere is too hot to permit them to exist in it : they decompose or " dissociate " at a lower temperature. It would be improper to dismiss this hypothesis with a mere passing mention, for during the past fifteen years it has been almost constantly under brisk discussion. Whether true or not, it is certainly not absurd, nor in 92 THE SUN. itself even improbable. The idea that at bottom there is but one material substance, and that all our chemical elements differ only in the way in which their ultimate molecules are built up out of the simple atoms of this " pantogen," is old, and has always been attractive to speculative minds. Grant it, and many otherwise puz- zling facts and relations of the new chemistry become in- telligible. At the same time it has not yet been proved, and so far all attempts to break up the elements into simpler bodies have failed. It seems impossible also to reconcile the hypothesis with the laws which connect the specific heat of bodies with their chemical constitu- tion and atomic weight. It may be added, too, that some of the supposed ob- servational facts upon which Mr. Lockyer relied at first to support his theory have turned out to be mis- takes due to errors of experimentation, or the use of insufficient spectroscopic power. Thus great stress was laid upon the so-called " basic " lines which appear to be common to the spectra of dif- ferent substances. If one runs over Angstrom's map of the solar spectrum he will find about twenty -five lines marked as belonging both to iron and calcium. The same is true of iron and titanium to a still greater extent, and to a considerable degree of several other pairs of substances. This fact might be explained in several ways. The common lines may be due first, to impurities in the materials worked with ; or, second, to some common constituent in the substances (which is Lockyer's view) ; or, third, to some similarity of molec- ular mass or structure which determines an identical vibration-period for the two substances ; or, finally, it may be that the supposed coincidence of the lines is only apparent and approximate not real and exact THE SPECTROSCOPE AN 7 D TUE SOLAR SPECTRUM. 93 in which case a spectroscope of sufficient dispersive power would show the want of coincidence. Now, Mr. Lockyer, by a series of most laborious researches, has proved that many of the coincidences shown on the map are merely due to impurities, and he has been able to point out which of the lines mapped as common to calcium and iron, for inst ce, belonged to each metal. As the iron employed is rendered suc- cessively purer and purer, certain of the common lines become fainter, and such evidently belong to calcium and not to iron. Similarly, when calcium is used, we can point out the lines which are due to the iron con- tamination. But, when all is done, we find that certain of the common lines persist, becoming more and more conspicuous with every added precaution taken to in- sure purity of materials. Moreover, when one of the substances, say the cal- cium, is subjected to continually increasing tempera- tures, its spectrum is continually modified, and these basic-lines, as Mr. Lockyer asserts, are the ones which become increasingly conspicuous, while others disappear. This is just what ought to happen if they are due to some element existing in both the iron and calcium an element liberated in increasing abundance with every rise of temperature. But unfortunately for the theory the application of our present powerful spectroscopes shows that in almost every case these " basic " lines are only instances of close coincidence. The writer in 1880 examined with care the seventy lines given on Angstrom's map as com- mon to two or more elements, and was able to resolve fifty-six of them into doubles or triplets ; and later ob- servers have resolved the rest or shown that they were due to impurities. Professor Rowland remarks that 94 THE SUN. "with the high dispersion" employed by him " Lock- yer's 'basic-lines' are widely broken up and cease to exist." As has been already remarked, the case of oxygen is peculiar. The great A and B bands of the solar spec- trum are certainly due to this gas, as Egoroff first defi- nitely proved in 1883 ; but, as the experiments and ob- servations of Janssen and others have since abundantly demonstrated, it is the oxygen of the eariKs atmosphere and not oxygen in the sun that produces them ; they belong to a low-temperature spectrum of the gas, and not to the spectrum produced by the electric arc or spark. But in 1877 the late Dr. Henry Draper, of New York, announced that he had discovered the presence of oxygen in the sun, and he published photographs which show, in a very plausible manner, the coincidence be- tween the bright lines of this element and certain bright spaces or bands in the solar spectrum. His method of precedure was to form the spectrum of oxygen by means of sparks from a powerful induction-coil, worked by a dynamo-electric machine, itself driven by an en- gine. These sparks passed between iron terminals, in a little chamber wrought out of soapstone, through which a current of pure oxygen was forced at atmospheric pressure nearly ; sometimes, however, air was used in- stead, giving the same results, except that the spectrum of nitrogen was then superadded to that of oxygen. The spectrum of this spark was photographed simultane- ously with that of the sun, the sunlight being brought in through half the slit by a small reflector, and thus a comparison was obtained, free from personal bias, be- tween the solar spectrum and that of the gas. The iron lines, due to the terminals, are a great assistance in test- THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 95 ing the adjustments. The oxygen lines produced in this way at atmospheric pressure are not so well defined as those seen in the spectrum of a Geissler tube, but are rather broad and hazy. In the blue portion of the solar spectrum, which alone is accessible to photography, the Fraunhofer lines are generally very numerous, close, and black ; but here and there is an interval free, or comparatively free, from lines. In a low-dispersion spectroscope such an interval looks like a bright band. Xow, almost every one of the dozen or so bright lines of oxygen, which the photographs display, falls exactly against one of these brighter interspaces, and it seems hardly probable that so many coincidences can be merely due to chance. Dr. Draper afterward repeated the laborious and expensive experiments in a still more elaborate manner, and with results entirely confirmatory. It is, however, extremely difficult to explain how oxygen in the sun's atmosphere can produce such an effect in the ordinary solar spectrum while remaining invisible in the spectrum of the chromosphere ; and the most careful search does not show in it a single one of these bright oxygen-lines. We say of these lines, because Dr. Schuster has shown, with great probability, that a different oxygen spectrum, with only four bright lines in it, has these four all represented by dark lines in the photospheric spectrum, and two of the four in the spec- trum of the chromosphere. With high dispersive powers, the "bright bands" of the solar spectrum entirely lose their prominence, and are even found to be occupied by numerous fine dark lines. Dr. John C. Draper has suggested that these dark lines may be the true representatives of oxygen. Still later photographic comparisons between the 96 THE SUN. solar spectrum and that of oxygen, made with high dis- persive powers, in the Physical Laboratory of Harvard College and at other places, render it certain that no coincidences of solar-lines with the bright lines in the line-spectrum of oxygen exist in the region covered by the photographic plates that were used (\ 3,750 to A, 5,034). Probably most spectroscopists consider this conclusive as to the absence of oxygen from the solar spectrum : at the same time, as Professor Pickering has shrewdly said, " one would scarcely expect to recognize a friend's countenance under the microscope" The discussion can hardly be considered as finally closed. The lines of the solar spectrum not only inform us as to the presence or absence of bodies in the solar atmosphere, but give us, to some extent, indications as to their physical condition. The spectrum of a given body, say hydrogen, varies very much in the relative strength and brightness of its lines, according to the circumstances of its production. If, for instance, the gas be highly rarefied, and the electric spark, which illuminates it, not too strong, the lines will be fine and sharp. Under higher pressure and more intense dis- charges, some of them will become broad and hazy, and new lines, before unseen, will make their appearance. So of other substances ; and this apart from the fact, before stated, that a given element often has several en- tirely different spectra. Changes, such as have been mentioned, go on up to a certain point, and then, sud- denly, an entirely new spectrum appears, not having apparently the slightest connection with the one which preceded it any more than if it came from an entirely different element or mixture of elements ; as, in fact, according to Mr. Lockyer's view, is probably the case. Now, in the solar spectrum, the dark lines character- THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 97 istic of an element are all coincident with bright lines of its gaseous spectrum ; but it is often the case that the relative width and intensity of the solar-lines do not match those of the bright lines in the spectrum ob- tained by artificial means. In the spectrum of calcium, for instance, certain lines, which in our laboratory ex- periments are the most conspicuous, are very faint upon the sun, and others, which are inconspicuous in the spark spectrum, are vastly more important on the solar sur- face. As yet, we are not able with certainty to inter- pret all these variations, but, in a general way, it may be said that they all point to the conclusion that the tem- perature of the solar atmosphere is considerably higher than that of any of our flames or electric arcs. SPECTKOSCOPIC INDICATIONS OF MOTION. At times, also, when the motions of the solar atmos- phere become unusually intense, the spectroscope ap- prises us of the fact, and gives us the means of deter- mining the rate at which the moving masses are advanc- ing toward us or receding from us. If a luminous body is approaching with a velocity at all comparable with that of light, the pitch of the light, if the expres- sion may be allowed its wave-length and number of vibrations per second will be changed arid heightened just as in the case of sound. Most of our readers have probably noticed the curi- ous change in pitch of the bell or whistle of a locomo- tive passing at full speed, especially if we ourselves were on a train moving in the opposite direction. If the velocity is great (about forty miles an hour for each of the trains) the pitch will drop a full third. The explanation, first given by Doppler, of Prague, in 1842, is simply this : If both we and the locomotive 98 THE SUN. carrying the bell were at rest, we should hear the bell's true sound, the pulsations following each other at regu- lar and the real intervals. If, now, we are rapidly ap- proaching the bell, the interval of time between the impact of each pulse upon the ear and the following one will be shortened, because after any pulse has been received we advance part way to meet the next, and so encounter it earlier than if we had remained at rest. Now, this interval of time between successive pulsations is precisely what determines the pitch of the sound : the more pulsations there are in a second the higher the pitch. It is obvious that, if we remain at rest and the bell approaches us, the same effect will be produced, and that, if both are moving, the effects will be added ; and, finally, it is clear that the recession of the hearer from the bell will produce the opposite effect and lower its pitch. Just the same thing holds good of light ; it also con- sists of pulsations, and the refrangibility of a ray and its diffrangibiHty, if we may coin the word, both de- pend upon the number of pulsations per second with which it reaches the diffracting or refracting surface. The more frequent the pulsations the more it will be refracted, and the less it will be diffracted. If, then, we were swiftly approaching a mass, say of incandes- cent hydrogen, we should find the position of each of its characteristic rays in the spectrum slightly altered, and falling farther from the red end of the spectrum (the region of slow vibrations) than if we were at rest. By comparing the positions of these lines with those obtained from a Geissler tube containing hydrogen, we could find how much change was produced, and there- fore how the velocity with which we are approaching the moving mass compares with that of light. Simi- THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 99 larly, if the body were advancing toward us. And, vice versa, if the distance were increasing, the lines would be shifted downward in the spectrum toward the red.* Because the velocity of light is exceedingly great (more than 186,000 miles per second), it is evident that only very swift motions can produce any sensible dis- placement of lines in the spectrum. Since, however, in the neighborhood of sun-spots and in the solar prom- inences, we frequently meet with masses of gas moving from thirty to fifty miles a second, and sometimes as much as three hundred miles a second, it is not unusual, in working with the telespectroscope, to observe the dis- tortion and displacement of portions of a dark line which are produced by these motions, and indicate them. The figure represents the appearance of the C line seen in the spectrum of a sun-spot by the writer on Sept. 22, 1870. The velocities indicated vary from two hundred and thirty to three hundred and twenty miles per second ; the latter is seldom, if ever, exceeded. Results of this sort are so surprising that there have been many attempts to escape from them, and to ac- count for the distortion of lines in some other way, but * The formula for computing the change of wave-length produced by a given velocity along the line of sight is very simple. Let A. be the real wave- length of the ray ; A.', the apparent wave-length as affected by the motion ; F, the velocity of light (186,330 miles a second) ; and v, the rate at which the distance between the observer and the source of light is increasing, then \' \ r= \ , which may be written A A, = A.. If the distance is decreasing, v must be taken as minus, and A,' will be less than A.. As an example, suppose that near a sun-spot a mass of hydrogen is approaching us at the rate of 50 miles a second : how much will the wave- length of the C line (\ = 6,563 units) be diminished ? A A. = 6.56S x -nrg^ = IHf = I'W units. That is, the C line will be shifted up (toward the blue) 1'77 units on the scale of Rowland's map. 100 THE SUN. without any satisfactory success. There have been diffi- culties raised also in regard to the mathematical theory of the matter. These have been met, however; and what amounts to an experimental verification of the correctness of the received view has been reached by measurements of the displacement of lines in the spec- tra of the eastern and western limbs of the sun. The eastern limb is moving toward us, the western from us, CHANGES IN THH C LINE (September 22, 1870). in consequence of the sun's rotation, each with a ve- locity of about 1'16 miles per second. The resulting displacement of the lines is, of course, very slight only about -j-J-g- of the distance between the two D lines but, small as it is, it has been satisfactorily detected and measured by several observers Zollner, Vogel, Lang- ley, and the writer, among the earlier. The values determined have ranged generally some- what larger than 1-16. My own result was 1*42 0'07, and was obtained in 1876 with the first grating-spectro- scope used in astronomical work. A later determination made by Crew, at Baltimore, with a much more powerful instrument s gave 1'IS. The THE SPECTROSCOPE AND THE SOLAR SPECTRUM. 1Q1 most complete and satisfactory series of observations of this sort is, however, that made by Duner in 1887-'S9, which not only gave a good determination of the snn's rotation period (25'56 days), agreeing closely with that deduced from spot observations, but also brought out clearly the " equatorial acceleration " (page 140). Cornu has made a beautiful application of this prin- ciple to enable one to discriminate immediately between the lines in the solar spectrum which are really "solar" and the " telluric " lines, which are due to our own at- mosphere. A small image of the sun is formed upon the slit-plate of the spectroscope by a lens which can be made to swing back and forth three or four times a second. This makes the solar image oscillate across the slit-plate in an east and west direction, and to the ob- server all the true solar lines appear to quiver, while the telluric lines stand fast. In the motion-distortions of lines Lockyer finds strong confirmation of his ideas. It not unfrequently happens that in the neighborhood of a spot certain of the lines which we recognize as belonging to the spec- trum of iron give evidence of violent motion, while, close to them, other lines, equally characteristic of the laboratory spectrum of iron, show no disturbance at all. If we admit that what we call the spectrum of iron is really formed in our experiments by the superposition of two or more spectra belonging to its constituents, and that on the sun these constituents are for the most part restricted to different regions of widely varying pressure, temperature, and elevation, it becomes easy to see how one set of the lines may be affected without the other. But the same facts are also explicable on various other hypotheses. CHAPTER IV. SUN-SPOTS AND THE SOLAR SURFACE. Granulation of Solar Surface. Views of Langley, Xasmith, Sccchi, and others. Faculae. Nature of the Photosphere. Janssen's Photo- graphs of Solar Surface the Reseau Photospherique. Discovery of Sun-Spots. General Appearance and Structure of a Spot. Its For- mation and Disappearance. Duration of Sun-Spots. Remarkable Phenomena observed by Carrington and Hodgson. Observations of Peters. Dimensions of Spots. Proof that Spots are Cavities. Sun- Spot Spectrum." Veiled Spots." Rotation of Sun. Equatorial Ac- celeration. Explanations of the Acceleration. Position of Sun's Axis and Sccchi's Table for its Position-Angle at Different Times of the Year. Proper Motions of Spots. Distribution of Spots. WHEN an observer, provided with suitable telescopic appliances, examines the surface of the sun, he finds a most interesting field before him. At first view, in- deed, it is less impressive than the moon ; there is not so much to attract the immediate attention no moun- tain-ranges and craters, no shadows, rills, or rays. But, if the telescope is a good one and the atmos- pheric conditions favorable, the details soon begin to come out : the surface is seen to be far from uniform, composed of minute grains of intense brilliance and ir- regular form, floating in a darker medium, and arranged in streaks and groups. If the magnifying power em- ployed is rather low, the general effect of the surface is much like that of rough drawing-paper, or of curdled milk seen from a little distance ; and, generally speak- ing, a low power is all that can be used, because the SUN-SPOTS AND THE SOLAR SURFACE. 103 heat of the sun commonly keeps the air in a state of great disturbance, so that it is only occasionally that the solar surface can be scrutinized with such powers as we con- tinually employ upon the moon and planets. But now and then times come favorable minutes, and even hours when the telescopic power can be pushed to its maxi- mum, and we get such views as that which Professor Langley has presented in the beautiful drawing of which our frontispiece is a reproduction. The grains, or " nod- ules," as Herschel called them, are then seen to be ir- regularly rounded masses, measuring some hundreds of miles each way, sprinkled upon a less brilliant back- ground, and making much the same impression as snow- flakes sparsely scattered over a grayish cloth, to use the comparison of Professor Langley. If the telescope has a diameter of not less than nine inches, and if the see- ing is absolutely exquisite, then these grains themselves are sometimes resolved into " granules," little luminous dots not more than a hundred miles or so in diameter, which, by their aggregation, make up the grains, just as they in their turn make up the coarser masses of the solar surface. Professor Langley estimates these gran- ules to constitute perhaps about one fifth of the surface of the sun, while they emit at least three quarters of the light. He and Secchi seem to be so far the only ob- servers who have ever fairly seen them. The " grains " have been known for years and described by many observers, but with some very embarrassing discrep- ancies. Nasmjth, in 1861, described them as " wil- low-leaves " in shape, several thousand miles in length, but narrow, with pointed ends ; and figured the surface of the sun as a sort of basket-work formed by the inter- weaving of such filaments. Fig. 28 is copied from one of his pictures. His statement excited a good deal of 104 GROUP OF SOLAB SPOTS OBSERVED AND DBAWN BY NASMYTH (June 5, 1864). SUN-SPOTS AND THE SOLAR SURFACE. 105 pretty warm discussion. Dawes entirely denied the existence of any such forms ; while Stone and Secchi assigned them much smaller dimensions, and compared them to rice-grains. Huggins agreed completely with FIG. 29. GRANTEES AND PoREB OF THE StTN's SURFACE. (After Hllggi neither, but represents the " make-up " of the solar sur- face by a drawing from which Fig. 29 is taken. This is unquestionably a very correct delineation of what is seen with a good telescope under circumstances fair, but not the best possible. 106 THE SUN - On portions of the sun's disk, however, the element- ary structure is often composed of long, narrow, blunt- ended filaments, not so much like " willow-leaves " as like bits of straw, lying roughly parallel to each other a "thatch-straw" formation, as it has been called. This is specially common in the penumbrse of spots, or in their immediate neighborhood. If one were to speculate as to the explanation of the grains and thatch-straws, it might be that the grains are the upper ends of long filaments of luminous cloud, which, over most of the sun's surface, stand approxi- mately vertical, but in the penumbra of a spot are in- clined so as to lie nearly horizontal. This is not certain, though ; it may be that the cloud-masses over the more quiet portions of the solar surface are really, as they seem, nearly globular, while near the spots they are drawn out into filamentary forms by atmospheric cur- rents. Whatever the explanation may be, the appearance of things in the immediate neighborhood of a spot is often pretty fairly represented by Mr. Nasmyth's pict- ures, though that of Professor Langley is decidedly more accurate in details, and represents far better see- ings. Near the edges of the disk the light falls off very rapidly, and certain peculiar formations, called the fac- ulae, are there much more noticeable than near the cen- ter of the disk. These f aculse (Latin, " a little torch ") are irregular streaks of greater brightness than the gen- eral surface, looking much like the flecks of foam which mark the surface of a stream below a waterfall. Not unfrequently they are from five to twenty thousand miles in length, covering areas immensely larger than any terrestrial continent. SUN-SPOTS AND THE SOLAR SURFACE. 107 The figure, taken from a photograph by De La Rue, gives a reasonably correct idea of the general appear^ ance of these objects, and of the darkening at the limb of the sun. No woodcut, however, is quite competent to give the delicate flocculence of the details. Until lately these faculae have been considered to be simply elevated portions of the photosphere mountain- ous billows of shining cloud which rise above the gen- eral level, and protrude through the denser portions of the solar atmosphere. Occasionally, when one of them passes over the edge of the disk, it can be seen to pro- ject like a little tooth the reader should not forget, however, that the elevation, to be perceptible at all, must be at least two hundred and twenty-five miles, or some forty-five times as high as any Himalaya. If they are elevations rising from the photosphere, the reason why they are so much more conspicuous 108 THE SUN. near the limb is simply this : The luminous surface is covered, as has been intimated before, with an atmos- phere which is not very thick compared with the di- mensions of the sun, but still sufficient to absorb a good deal of the light. Light coming from the center of the disk penetrates this atmosphere at , as is apparent from the figure, under the most favorable conditions, and is but slightly reduced in amount. The edges of the disk, on the other hand, are seen through a much greater thickness of atmosphere, as at 5, and are, therefore, of course, much obscured, the amount of absorption being by some observers put as high as seventy five per cent. If, now, to take an extreme case, we suppo:e a facula high . enough to lift its summit quite through this at- mosphere, it will itself suffer no diminution of brill- iance while the sun's rotation carries it from the center of the disk to the limb, but it will have passed from a background of brightness almost equal to its own, on which it would be seen only with difficulty, to another seventy-five per cent, or so darker, and will thus become very conspicuous. What is true of faculae of such extreme dimensions is, of course, also measura- bly true of those of inferior elevation. The recent pho- to-spectrographic work of Hale and Deslandres suggests, SUN-SPOTS AND THE SOLAR SURFACE. 109 however, a different explanation of the faculae. Their spectrum (as long ago frequently observed by the writer, visually) shows the great H and K bands of calcium always reversed by a thin bright line running down the middle of each ; and while the reversal directly over a spot is usually " single," it is usually " double " * in the faculous region surrounding it i. e., the bright line is itself double, as in Fig. 73, page 231. This makes it more or less probable that the faculae, instead of being mere protrusions from the photosphere, are really luminous masses of calcium vapor floating in the solar atmosphere possibly, as Professor Hale thinks, identical with the prominences themselves. SPECTBOHELIOGBAPH PHOTOGRAPH OF Suu'a DISK, WITH F ACUL.fi. But Deslandres and Maunder dissent from this, and say that while these objects shown by the spectroscope are clearly connected with the prominences, they are as clearly not identical with them. * Such double reversal is a very common phenomenon in laboratory experiments upon metallic spectra. HO THE SUN. Fig. 31* is from one of Professor Hale's spectro- heliographic photographs, made by the apparatus de- scribed in Chap. VI, page 233. The faculse are found to some extent over the whole surface of the sun, though only sparingly in the polar re- gions, but they are especially abundant in the immediate neighborhood of spots, as Fig. 30 well shows. There are, however, numerous faculae without neighboring spots. Except near the spots, the faculae change form and place, for the most part, rather slowly, persisting some- times for several days without any very apparent alter- ation. Still, close observation and micrometric measure- ment will always detect some movement or deforma- tion, even within an interval of only an hour or two ; and near the spots the changes are often so rapid and extreme as to puzzle even a skilled draughtsman to keep up with them. This, of course, shows that the faculae are not to be identified with mountains ; they, are not permanent and stable, nor is the surface of the sun continental or oce- anic even, but either a sheet of flame or of cloud rolling and tossing, and never at rest. When we come to study the minute details of the granulation, we find move- ments at the rate of a thousand miles an hour to be the rule rather than the exception. And, although this is not the proper place to treat the subject at length, we may add that all we can learn as to the temperature and constitution of the sun makes it hardly less than certain that the visible surface, which is called the photosphere, is just a sheet of self-luminous cloud ; precisely like the clouds of our own atmosphere, with the exception that the droplets of water which constitute terrestrial clouds are replaced on the sun by drops of molten metal, and that the solar atmosphere in I PHOTOGRAPHS OF A PORTION OF THE SUN, BY JANSSEN. Meudon, June 1, 18 Interval, 50 minutes SUN-SPOTS AND THE SOLAR SURFACE. HI which they float is the flame of a burning, fiery furnace, raging with a fury and an intensity beyond all human conception. Looking at it ninety-three million miles away, we fail at first to see, in such objects as faculse and granules, the evidence of such commotion ; but, when we convert our micrometric measurements of barely perceptible changes into miles and velocities, and fig- ure to ourselves the scale of movement, we gradually comprehend their meaning, and begin to understand what we are dealing with. A great advance in our knowledge of the structure of the solar surface was gained through the photograph- ic work of Janssen, mentioned in a previous chapter.* Many of his pictures (in which the disk of the sun measures about eighteen inches in diameter) show the details of the surface nearly if not quite as well as any visual observations; and with the advantage that, while the observer with the eye could only command a small field of view, one can, on these photographic plates, command the whole at once, and catch the relations of different parts. On examining one of these magnifi- cent plates, one is at first struck with a sort of " smudg- iness " (to use the expression of Mr. Huggins in de- scribing them), which might give the impression that it was not properly cleaned before coating with the collodion. A closer examination, however, shows that the peculiarity is not in the plate but in the image ; there are patches of clear definition, half an inch or so in diameter upon a picture of the size mentioned, sep- arated by streaks and patches where everything is in- distinct and confused. One might naturally attribute this to the disturbance of the air in the telescope-tube, and to clouds of vapor * See page 52. 112 THE SUN. rising from the damp collodion surface when struck by the flash of sunlight during its exposure ; but Janssen has found that pictures taken in immediate succession show the same "smudges" on the same parts of the sun, which, of course, would not happen if they were the result of accidental currents of air or vapor in the telescope-tube. He infers, therefore, that the phenome- non is solar, and has given it the name of the Reseau Photospherique, or " Photospheric Keticulatioii," since the streaks and patches of indistinctness cover the sur- face like a net. The discovery of this feature in the structure of the solar surface is among the most interesting and im- portant results of astronomical photography. While pictures taken in immediate succession ex- hibit the same details of reticulation, those taken at intervals of an hour or two show great changes, es- pecially near spots and faculse. We present on the opposite page a pair of such photographs, borrowed from the " Annuaire " of the Bureau of Longitudes for 1879. The original pictures were taken by Janssen, at Meudon, on June 1, 1878, with an interval of fifty min- utes between them. They show clearly the peculiar characteristics of the reseau pJiotosplierique, as well as the nature and extent of the changes which occur 'in so short a time. Compare, especially, the granulation in the lower right hand corner of each picture, and imme- diately around the upper spot, remembering all the while that the scale of the picture is about forty-six thousand miles to the inch, and that the little spot at the top of the figure is nearly seven thousand miles in diameter. The idea of M. Janssen is that the regions of indis- tinctness are those where we look down upon the sur- SUX-SPOTS AND THE SOLAR SURFACE. H3 face through a portion of the sun's atmosphere which is at the moment especially agitated, while the parts where the details of the granulation are clear and well defined are those which, at the moment, are covered by an atmosphere unusually quiet and homogeneous. These regions are continually interchanging with each other, just as areas of storm and fine weather sweep over the surface of the earth, but with inconceivably greater swiftness. It is not, however, certain that the disturbed por- tions of the solar atmosphere, which produce the in- distinctness in question, lie near the sun's surface. It may be that they are high up, and it would not be an un- reasonable conjecture to suppose that the streamers and luminous masses of the corona may be concerned in the phenomenon ; it is almost certain that any great aggrega- tion of chromospheric matter would modify the appear- ance of whatever might be situated beneath it. The simple fact is, of course, that we are looking down upon the granules and other features of the sun's surface, not through an atmosphere shallow, cool, and quiet like the earth's, but through an envelope of matter, partly gase- ous and partly, perhaps, pulverulent or smoke-like, many thousand miles in depth, and always most profoundly and violently agitated. But, if there happens to be a well-formed group of spots upon the solar surface, they will be sure to claim the attention of one who, for the first time, looks at the sun through the telescope, quite to the exclusion of everything else. The umbra, with its central nuclei, and overlying bridges, veils, and clouds ; the penumbra, with its delicate structure of filaments and plumes ; the surrounding faculae and the agitated surface of the pho- tosphere in the whole neighborhood of the disturbance ; 114 THE SUN. above all, the continual change and progress of phe- nomena combine to make a fine sun-spot one of the most beautiful and intensely interesting of telescopic objects. Even before the days of telescopes there are numer- ous records of dark spots seen by the naked eye upon the disk of the sun, especially in the annals of the Chi- nese. In the year 807 A. D., a large spot was visible in Europe for some eight days, and was supposed by many to be the planet Mercury, as was the case with a spot observed by Kepler in 1609 ; indeed, in all cases where such appearances were noted, they were invariably as- cribed to bodies intervening between the earth and the sun. The idea of such imperfections upon the disk of a celestial body was utterly repugnant to the theologi- cal philosophy of the middle ages, and was admitted only slowly and grudgingly even after the demonstra- tion of the fact was complete. In 1610 and 1611 the discovery seems to have been made independently by Fabricius, Scheiner, and Gali- leo Fabricius, according to our modern rules of scien- tific priority, being entitled to the credit as the first to publish the fact in a work, " De Maculis in Sole Obser- vatis," which appeared at Wittenberg in June, 1611. The discovery was, of course, a necessary corollary to the invention of the telescope, which first came into use in Holland in 1608 or 1609. Fabricius's first observa- tion was made in December, 1610. Galileo, in a letter responding to the account of Schemer's discovery, and published early in 1612, claims to have seen the sun- spots with his newly-constructed telescope as early as October, 1610. Scheiner appears to have first seen sun- spots at Ingolstadt in March, 1611 ; but his ecclesiasti- cal superior warned him against believing his own eyes SUN-SPOTS AND THE SOLAR SURFACE. H5 in opposition to the authority of Aristotle, and it was not until November and December that he published an account of the matter in three letters to one Welser, a burgomaster of Augsburg, some months after the work of Fabricius had been printed. There is no reason whatever to doubt the word of Galileo, and his experi- ence in losing the credit of this discovery, in conse- quence of his slowness of publication, seems to have been the origin of his curious method of publishing his subsequent discoveries in the form of anagrams, the in- terpretation of which was withheld for a time. At the very outset of his observations, Fabricius, as well as Galileo, recognized that the spots are objects upon the surface of the sun, and that this body rotates on its axis, carrying them with it. Schemer at first maintained that they were planets moving very near the sun, but not in contact with it. Many shared this opinion, and Tarde, a French astronomer, went so far as to name them the Bourbonian stars, in honor of the Bourbon dynasty. Schemer's further observations soon convinced him, however, of the correctness of Galileo's opinion and arguments. Some twenty years later Scheiner published an enormous volume, the "Rosa Ursina," containing an account of his observations and apparatus. His telescope was mounted equatorially, and arranged to throw the sun's image upon a screen in pre- cisely the manner employed by some of the best mod- ern observers. He determined the time of the sun's rotation and the position of his equator with a very creditable degree of accuracy. Since then observations upon these objects have been more or less kept up all the time, but not with any regular assiduity until within the last thirty years. It was soon found that they are only transitory and 116 THE SUN. cloud-like in their nature, and interest in them there- fore flagged, until their relations to the constitution of the sun began to be recognized. A well-formed solar spot consists, generally speak- ing, of two portions a very dark, irregular, central portion called the umbra, surrounded by a shade or fringe called the penumbra, less dark, and for the most part made up of filaments directed radially inward. The appearance of things, under ordinary circumstances of seeing, is as if the umbra were a hole, and the pe- numbral filaments overhung and partly shaded it from our view, like bushes at the mouth of a cavern. I say as if, and very possibly this is the actual case, the cen- SPOT or JCLY 16, 1866. tral portion being a real cavity filled with less luminous matter, and depressed below the general level of the photosphere, while the penumbra overhangs the edge. The figure, copied from Secchi, is a fair represen- tation of such a spot, and may be compared with the SUN-SPOTS AND THE SOLAR SURFACE. H7 photographs of Janssen, which exhibit pretty ranch the same peculiarities, though with less of minute detail. The drawings of ^Nasmyth and Langley * show so much more of the detail than is ordinarily seen, that they are really less satisfactory representations of what one may expect when he observes a spot for the first time. Sev- eral points at once strike the attention. In the first place, the nearly circular form of the spot, which is the ordinary form during the middle life of one of these objects. While forming, and when on the point of dis- appearing, it is usually much more irregular. It is to be noticed also that there is nothing like a gradual shading off, either between the umbra and the penum- bra or between the penumbra and the surrounding por- tions of the photosphere ; on the contrary, the line of separation is strongly marked in each case, the penum- bra being much brighter at the inner edge, and darker at the outer, so that it contrasts distinctly both with the umbra and with the neighboring surface of the sun. This brightness of the inner penumbra seems to be due to the crowding together of the penumbral filaments where they overhang the umbra. Again, it is observ- able that there is a general antithesis between the irreg- ularities of the contour of the outer and inner edges of the penumbra. For the most part, where an angle of the penumbral matter crowds in upon the umbra, it is matched by a corresponding outward extension into the photosphere, and vice versa. It is noticeable also that many of the penumbral filaments are terminated by little detached grains of luminous matter, and there are also fainter veils of a substance less brilliant, but some- times rose-colored, which seem to float above the um- bra. Otherwise the umbra in the figure appears to be * See frontispiece, and page 104. 118 TIIE SUN - uniformly dark ; * but, if we had been actually observing the object on the 16th of July, 1866, when this pict- ure was made, we should have found even the umbra full of detail made up of cloudy masses of a brilliance really intense, and dark only by contrast with the still intenser brightness of the solar surface, as becomes ap- parent when the light from other portions is excluded. Probably we should have been able also to detect among these clouds one or more of the minute circular spots, first discovered by Dawes, much darker than the rest of the umbra, and looking like the mouths of tubu- lar orifices penetrating to unknown depths. If we were able to continue our watch for some time, we should see the details continually changing. The faint veils of overlying cirrus would probably melt away, and be replaced by others in some different po- sition ; the bright granules at the tips of the penumbral filaments would seem to sink and dissolve, and fresh portions would break off to replace them. We should find a continual indraught of the luminous matter over the whole extent of the penumbra. Almost certainly the spot would change its form and size, quite percep- tibly from day to day, and sometimes even from hour to hour. Of course, we should find it steadily moving over the solar disk from the east toward the west, and as it neared the edge it would become apparently ellip- * The umbra appears not black, but of a deep purplish tint. It is questionable, however, whether this color is real, or only due to the sec- ondary spectrum of the telescope object-glass. The principal reason for suspecting this to be the case is in the fact that, during a transit of Mercury or Venus, the planet's disk is found to present precisely the same tint, while there ia no imaginable explanation for its really being anything but black. It is certain, too, on optical grounds, that any ordinary object-glass must show a purplish fringe extending inward over any dark spot upon a white background. SUN-SPOTS AND THE SOLAR SURFACE. H9 tical in form ; the penumbra on the edge of the spot nearest the center of the sun would grow narrower and, perhaps, disappear entirely, and at last the spot, appearing like a mere line of darkness, but probably accompanied by an attendant crowd of faculae, would pass out of sight behind the limb, perhaps to reappear again after a fortnight at the eastern edge. I say per- haps, because, quite as often as not, these short-lived objects, are seen but once, not lasting through even a single revolution of the sun. 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 eigh- teen months. There are cases, however, where the dis- appearance of a spot is very soon followed by the ap- pearance of another at the same point, and sometimes 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. 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 pe- numbra only upon one side, and usually irregular in form. It is noticeable, also, that in such cases, when any considerable change of form or structure 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 happens, the parts usually seem to repel each other and fly asunder with great 120 THE SUN - 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 hun- dred miles, even with a very high magnifying power. Velocities of three or four hundred miles an hour are usual, and velocities of one thousand miles, and even more, are by no means exceptional. At times, though very rarely, a different phenome- non of the most surprising and startling character ap- pears in connection with these objects : patches of in- tense brightness suddenly break out, remaining visible for a few minutes, moving, while they last, with veloci- ties as great as one hundred miles a second. One of these events has become classical. It oc- curred on the forenoon (Greenwich time) of Septem- ber 1, 1859, and was independently witnessed by two well-known and reliable observers, Mr. Carrington and Mr. Hodgson, whose accounts of the matter may be found in the monthly notices of the Royal Astronomi- cal Society for November, 1859. Mr. Carrington at the time was making his usual daily observation upon the position, configuration, and size of the spots by means of an image of the solar disk upon a screen, being then engaged upon that eight years' series of observations which lies at the foundation of so much of our present solar science. Mr. Hodgson, at a dis- tance of many miles, was at the same time sketching details of sun-spot structure by means of a solar eye- piece and shade-glass. They simultaneously saw two luminous objects, shaped something like two new moons, each about eight thousand miles in length and two thou- sand wide, at a distance of some twelve thousand miles from each other. These burst suddenly into sight at SUN-SPOTS AND THE SOLAR SURFACE. 121 the edge of a great sun-spot, with a dazzling brightness at least five or six times that of the neighboring por- tions of the photosphere, and moved eastward over the spot in parallel lines, growing smaller and fainter, until in about five minutes they disappeared, after traversing a course of nearly thirty-six thousand miles. Their pas- sage did not seem in any way to change the configura- tion of the spot over which they passed. Mr. Carring- ton found his drawing, which was completed just before they appeared, still quite correct after they had vanished. Of course, it is possible to question the connection be- tween this phenomenon and the spot near which it ap- peared ; but, as somewhat similar appearances have been seen by other observers since then, and always in the neighborhood of spots, it is probable that there is some relation in the case. Opinions have differed widely as to the explanation. Some have maintained that the phenomenon was simply due to the fall of a couple of immense meteors into the sun's atmosphere, others that it was caused by some sudden and powerful eruption from beneath, such as the spectroscope often reveals to us nowadays ; an eruption, however, of most unusual brilliance and violence, for not one of the outbursts since then observed by the spectroscope has ever been visible without its aid. The event occurred in the midst of a remarkable magnetic storm : from August 28th to September 4th there were auroras every night all over the world, and the earth currents were often so strong as greatly to interfere with telegraphic communication. On the night of September 1st, however, as Mr. Ellis has lately shown from the original records, the magnetic disturb- ance was not specially intense, so that the occurrence observed by Carrington and Hodgson could not have 122 TIIE SUN - been the cause of the magnetic storm more likely it was a consequence, if there was any connection. 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 presence of numerous and brilliant faculse,* among which, " pores " or minute black dots are scattered. These enlarge, and between them appear grayish patches, ap- parently caused by a dark mass lying veiled below a thin layer of luminous filaments. The veil grows grad- ually thinner, and vanishes, giving us at last the com- pleted spot with its perfect penumbra. The " pores," some of them, coalesce with the principal spot, some disappear, and others constitute the attendant 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 the penumbra' filaments becomes con- fused, and, as Secchi expresses it, the luminous matter of the photosphere 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 disturbance is not unfrequently renewed at the same point after a few days, and a fresh spot appears just where the old one was overwhelmed. * This is Secchi's view. Lockyer maintains that the spots appear be- fore the faculae. SUX-SPOTS AXD THE SOLAR SURFACE. 123 We transcribe from a paper by the late Dr. Peters, of Hamilton College, a very graphic account of the ap- pearance and decay of certain sun-spots, based upon his observations at Naples in 1845-'46. It is printed in Volume IX of the " Proceedings of the American As- sociation for the Advancement of Science." He says : "The spots arise from insensible points, so that the exact moment of their origin can not be stated ; but they grow very rapidly in the beginning, and almost always in less than a day they arrive at their maximum of size. Then they are stationary, I would say in the vigorous epoch of their life, with a well-defined penumbra of regular and rather simple shape. So they sustain themselves for ten, twenty, and some even for fifty days. Then the notches in the margin, which, with a high magnifying power, always appear somewhat serrate, grow deeper, to such a degree that the penumbra in some parts becomes interrupted by straight and narrow luminous tracks already the period of decadence is approaching. This begins with the following highly interesting phenomenon : Two of the notches from opposite sides step for- ward into the area, over-roofing even a part of the nucleus ; and suddenly from their prominent points flashes go out, meeting each other on their way, hanging together for a moment, then breaking off and receding to their points of starting. Soon this electric play begins anew and continues for a few minutes, ending finally with the connection of the two notches, thus establishing a bridge, and dividing the spot in two parts. Only once I had the fortune to witness the occurrence between three advanced points. Here, from the point A a flash proceeded toward B, which sent forth a ray to meet the former when this had arrived very near. Soon this seemed saturated, and was suddenly repelled ; however, it did not retire, but bent with a sudden swing toward C ; then again, in the same manner, as by repulsion and attraction, it re- turned to B ; and, after having thus oscillated for several times, A adhered at last permanently to B. The flashes proceeded with great speed, but not so that the eye might not follow them dis- -tinctly. By an estimation of time and the known dimension of space traversed, at least an under limit of the velocity may be found ; thus, I compute this velocity to be not less than two hun- 124 THE SUN. dred millions of metres (or about one hundred and twenty thou- sand miles) in a second (sic). "The process described is accomplished in the higher photo- sphere, and seems not to affect at all the lower or dark atmos- phere. With it a second, or rather a third, period in the spot's life has begun, that of dissolution, which lasts sometimes for ten or twenty days, during which time the components are again sub- divided, while the other parts of the luminous margin, too, are pressing, diminishing, and finally overcasting the whole, thus end- ing the ephemeral existence of the spot. " Eather a good chance is required for observing the remark- able phenomenon which introduces the covering process, since it is achieved in a few minutes, and it demands, moreover, a per- fectly calm atmosphere, in order not to be confounded with a kind of scintillation which is perceived very often in the spots, especially with fatigued eyes. The observer ought to watch for it under otherwise favorable circumstances when a large and ten- or twenty-days'-old spot begins to show strong indentations on the margin." Dr. Peters, so far as we know, is the only observer who describes the remarkable phenomenon of flashes extending across an umbra with electrical velocity ; and for this reason, and because his instrument was not of the highest power a three-and-a-half -inch refractor perhaps his account must be received with a little re- serve until further confirmed. At the same time, there is nothing in the nature of the sun, or of a sun-spot, so far as at present known, to make the statement in itself impossible ; and certainly Dr. Peters holds deservedly a very high rank among astronomers for acuteness and accuracy of observation and description. It must not be understood that the life-history of a spot, just sketched, applies to all, or even with exact- ness to a majority, of them. Almost every one has its own idiosyncrasies, departing in some respect or other from the usual course of things. Spots of unusual mag- SUN-SPOTS AND TIIE SOLAR SURFACE. 125 nitude and activity often seem to have no quiet middle life ; there is no time in their history when they are not doing something or other surprising, and more or less unprecedented. We have spoken of the filaments which compose the penumbra as directed inward toward the center of the spot. This is the general rule, but the exceptions are numerous, and nothing can show better than Pro- fessor Langley's exquisite drawing how wide the di- vergence often is from this law. While at the left- hand and upper portions of the great spot (which, though " typical," is not a specimen of a quiescent spot) the filaments present the ordinary appearance, at the lower edge and upon the great overhanging branch they are arranged very differently. Yery curious, and rare, also, though we have ourselves seen a similar thing on two or three occasions, is the feathery brush which reaches in below the " branch," so closely resembling a frost-crystal upon the window-pane in a winter's morn- ing. What may be the cause of such formations it is now quite impossible to say. Probably analogies drawn from our terrestrial clouds will go further toward an explanation than any others yet proposed. Usually the penumbral filaments are brightest at the inner end where they apparently project over the um- bra, and under ordinary circumstances of vision the end appears blunt and even club-shaped. With the great twenty -three-inch telescope at Prince- ton, and on a few occasions, when the seeing has been fine enough to permit the use of powers of from six hun- dred and upward, the writer has found that, in many cases at least, the apparently club-like, almost bulbous, ends of the penumbral filaments are really fine, sharp- pointed hooks, reminding one of the curling tips of 126 THE SUN. flames, or grass-blades bending over. Ordinarily they are seen as club-like, simply because of their brightness and the irradiation and diffraction effects of moderate- sized object-glasses. Not unfrequently the penumbral filaments are curved and spirally arranged, showing a marked cyclonic action. In such cases the whole spot usually turns slowly around, sometimes completing an entire revolution in a few days. More frequently, however, the spiral motion persists but a short time, and occasionally, after continuing for a while in one direction, the motion is reversed. Very often, in spots of considerable extent, there are opposite spiral movements in different portions of the umbra ; indeed, this is rather the rule than the exception. Neigh- boring spots show no tendency to rotate in the same direction. The number of spots in which a decided cyclonic motion appears is relatively quite small, not exceeding, according to the observations of Carrington and Secchi, more than two or three per cent, of the whole. Of course, these facts are sufficient to show that this kind of motion, when it occurs, is not attribut- able to anything like that action of the terrestrial atmos- phere which determines the right- and left-handed rotation of our great storms in the southern and northern hemispheres. It is probably caused in sun-spots by merely accidental circumstances which convert the pe- numbral indraught into a whirl of no great rapidity or certain direction. It does not seem possible to find in this occasional cyclonic motion, as Faye attempts to do, the key and explanation of the whole series of sun-spot phenomena. The dimensions of sun-spots are sometimes enor- mous. Many groups have been observed covering areas of more than one hundred thousand miles square, SUN-SPOTS AND THE SOLAR SURFACE. 127 and single spots have been known to measure forty or fifty thousand miles in diameter, the central umbra alone being twenty-five or thirty thousand miles across. A spot, however, measuring thirty thousand miles over all, would be considered large rather than small. An object of this size upon the sun's surface can easily be seen without a telescope when the brightness is reduced either by clouds, or nearness to the horizon, or by the use of a shade-glass. At the transit of Venus, in 1882, every one saw the planet readily without tele- scopic aid. Her apparent diameter was about 67" at the time, which is equivalent to about thirty one thou- sand miles on the solar surface. Probably a very keen eye would detect a spot measuring not more than twenty- three or twenty-four thousand miles. Hardly a year passes, at times when spots are numer- ous, without furnishing several as large as this ; so that it is rather surprising than otherwise that we have not a greater number of sun-spot records in the pre tele- scopic centuries. The explanation probably lies in two things : the sun is too bright to be often or easily looked at, and when spots were seen they would be likely to be taken for optical illusions rather than realities. During the years 1871 and 1872 spots were visible to the naked eye for a considerable portion of the time. On several occasions pupils of the writer have noticed them of their own accord, without having had their at- tention previously directed to the matter. The largest spot yet recorded was observed in 1858. It had a breadth of more than one hundred and forty- three thousand miles, or nearly eighteen times the diam- eter of the earth, and covered about one thirty-sixth of the whole surface of the sun. Other very large ones appeared in 1892 and 1893. 128 THE SUX. Fig. 33, taken by the publisher's permission from Flummarion's Popular Astronomy, represents a very ONE OP TUB LARGEST RUN-SPOTS : SEVEN TIMES THE SIZE OF THE EARTH. OBSERVED OCTOBER 14, 18S3. VISIBLE TO THE NAKED EYB. large and interesting spot which appeared in October, 1883. It is from the drawings of Tacchini. Spet- tros. Ital., Vol. XIII. NATURE OF THE SPOTS. It has been intimated that the spots are depressions below the general level of the solar surface. For more than a century this has been the accepted doctrine, and it is probably correct ; at the same time it can hardly be SUX-SPOTS AXD THE SOLAR SURFACE. 19 regarded as absolutely settled, since it has been called in question recently by high authorities, and is still de- bated. In December, 1894, Mr. Hovvlett, who has been for more than thirty years a persistent observer of the sun, presented to the Royal Astronomical Society all his sun-spot drawings, several thousand in number, and covering the whole period from 1859 to 1893. He took the opportunity to express very strongly his opinion that the facts are against the theory that the spots are "hollows" as usually supposed, and was supported in his view by a number of good observers, who made it clear that if the spots are really depressed at all, they must be very shallow compared with their diameter. The idea was first clearly brought out by Dr. Wil- son, of Glasgow, in 1769, and his demonstration was based upon the behavior of the penumbra of a spot which he observed in November of that year. He found that, when the spot appeared at the eastern limb or edge of the sun, just moving into sight, the penum- bra was well marked on the side of the spot nearest to the edge of the disk, while on the other edge of the spot, that next the center, there was no penumbra vis- ible at all, and the umbra itself w r as almost hidden, as if behind a bank. When the spot had moved a day's journey farther inward toward the center of the disk, the whole of the umbra came into sight, and the pe- numbra on the inner edge of the spot began to be visible as a narrow line. After the spot was well advanced upon the disk, the penumbra was of the same width all around the spot ; but, when the spot approached the sun's western limb, the same phenomena were repeated as at the eastern that is, the penumbra on the inner edge of the spot narrowed much faster than that on the outer, disappeared entirely, and finally seemed to 10 130 TOE SUN. hide from sight much of the umbra, nearly a whole day before the spot passed from view around the limb. Of course, it is hardly necessary to point out what the fig- ure at once makes evident, that this is precisely the way things would go if the spot were a saucer-shaped de- DIAGRAM ILLUSTRATING THE FACT THAT SUN-SPOTS AUE HOLLOWS IN THB PHOTOSPHERE. pression in the sun's surface, the bottom of the saucer corresponding to the umbra, and the sloping sides to the penumbra. The observation of a single spot would hardly settle the question, because we frequently have spots with a one-sided penumbra. In fact, when spots are either in the process of formation or of dissolution the penumbra is seldom of uniform width all around. De La Kue, Stewart, and Loewy made, therefore, some years ago, a careful discussion of something more than six hundred cases of spots, with measurable penum- brae, and found that, in a little over seventy-five per cent, of all the cases, the penumbra was widest on the SUN-SPOTS AND THE SOLAR SURFACE. 131 edge of the spot nearest the limb, as Wilson's theory requires ; in a little more than twelve per cent, there was no noticeable difference; and in the remaining twelve per cent, it was widest on the inner edge. Father Sidgreaves, on the other hand, in following up the discussion raised by Mr. Howlett, gets an opposing verdict from the Stonyhurst sun-spot drawings. Out of one hundred and eighty-seven sketches, which he se- lected as fair tests of the Wilsonian theory, only forty- seven favored it, and one hundred and forty were opposed. But we suppose he has included as " opposed " all that did not distinctly indicate depression. Others, Secchi especially, have investigated the mat- ter by caref ully measuring, from day to day, the position on the sun's disk of some selected point in the umbra of a spot. The work is not easy, and rather unsatisfactory, on account of the rapid changes, which make it difficult to identify the point of reference in successive observa- tions ; still, the result appears decisive, showing, as an ordinary rule, that what may be called the " floor " of the umbra is depressed from two to six thousand miles, and sometimes more, below the general level of the photosphere. But the refraction of the solar atmos- phere makes the result uncertain. On a few occasions, when a spot of unusual size and depth passes over the limb of the sun, a distinct depres- sion is observed in the outline. Cassini describes such an instance in 1719. Herschel, De La Rue, Secchi, and others have given us several other observations of the same kind. Usually, however, the faculae, which sur- round the spot, mask this effect entirely, and often actually give us a number of little projecting hillocks in place of the expected depression. 132 THE SUN. SPECTRUM OF SUN-SPOTS. The spectrum of a sun-spot also furnishes an argu- ment in the same direction, tending to show that the dark portion is a cavity filled with gases and vapors, which produce the ohscuration, in part, at least, by ab- sorbing the light emitted from the floor of the depres- sion. It is not difficult to set the instrument in such a manner that the image of a sun-spot shall fall precisely upon the slit of the spectroscope.- In this case the spec- trum will be seen to be traversed by a longitudinal dark stripe, which is the spectrum of the umbra of the spot : on each side is the spectrum of the penumbra, which is usually only a trifle fainter than that of the general sur- face of the sun. The width of the stripe, of course, de- pends upon the diameter of the spot. Along the whole length of the spot-spectrum the background is darkened, showing a general absorption ; and in the upper part of the spectrum, from F to H, this seems to be pretty much all that can be noticed. The middle portion of the spectrum, however, under extremely high dispersion is different in this respect, as was discovered by the writer in 1883, and has since been abundantly confirmed by Duner and others. In many spots, especially large ones that are nearly circular and quiescent with a very dark nucleus, the spectrum of the nucleus between E and F is not continuous, but is made up of countless fine, dark lines, for the most part touching or slightly overlapping, leaving here and there, however, unoccu- pied intervals which look like (and may be) bright lines. Each dark line is spindle-shaped i. e., thicker in the middle where the spectrum is darkest, and tapers to a fine, faint, hair-like mark at each end ; most of them can be traced across the penumbra-spectrum, and even out upon the general surface of the sun. The aver- SUX-SPOTS AND THE SOLAR SURFACE. 133 age distance between the lines is about half that be- tween the two components of J s , so that within the 5 group the total number of dark lines is some 300, and there are seven or eight of the bright lines. This structure is most easily seen in the part of the spectrum between E and F ; above F the lines are crowded so closely that it is diflicult to resolve them, and below E they appear to grow wider, more diffuse, and fainter. It seems to indicate that the principal absorption which darkens the center of a sun-spot is not such as would be caused by minute solid or liquid particles by smoke or cloud which would give a continuous spectrum ; but it is a true gaseous absorption, producing a veritable dark- line spectrum, in which the lines are countless and con- tiguous. In the lower part of the spectrum, especially between C and D, the spot-spectrum is .full of interesting details and peculiarities, which deserve a far more thorough and prolonged study than they have yet received. Many of the dark lines of the ordinary spectrum are wholly unmodified in the spectrum of the spot ; in fact, this seems to be the case with the majority of them. Others, however, are much widened and darkened, and some, which are hardly visible at all in the ordinary spectrum, are so strong and black as to be very conspicu- ous : these are usually spindle-shaped, much wider in the center of the nucleus than at its edges and in the penumbra, so that they are often called " fish-bellies." Certain other lines, which are strong in the ordinary spectrum, thin out and almost disappear in the spot- spectrum, and some are even reversed at times. There are also a number of bright lines, not very brilliant, to be sure, but still unmistakable, and there are some dark shadings of peculiar appearance. 134 THE SUN. The annexed figure (Fig. 35), which represents a small portion of the spectrum of a spot observed by the writer in 1872. shows nearly all of these peculiarities. The portion represented lies between C and D, the scale attached being that of KirchhofFs map. PORTION or SUN-SPOT SPECTRUM BETWEEN C AND D. Speaking in a general way, the lines of hydrogen, iron, titanium, calcium, sodium, and vanadium are spe- cially affected. The hydrogen lines are often reversed ; those of iron, titanium, calcium, and vanadium are usu- ally thickened, and those of sodium are often enormously widened, and occasionally both widened and doubly reversed, as shown in Fig. 36, which represents their appearance in the spectrum of a spot observed on Sep- tember 22, 1870. It will be noticed that at the same time the helium-line, D 3 , which usually is invisible on the solar surface, was quite conspicuous as a dark shade. On this occasion the lines of magnesium also behaved in the same manner as those of sodium. As has already been mentioned (page 109), the H and K bands are always reversed in the sun-spot spec- trum. Usually, over the spot itself, the reversal is only " single," but double reversal is not very uncommon. SUN-SPOTS AND THE SOLAR SURFACE. 135 Mr. Lockyer announces, as a result of a long series of observations, that there is a striking difference between the spot spectra at the time of maximum and minimum sun-spot frequency ; the lines that are most conspicuous by widening and darkening being by no means the same in the two cases. The most remarkable change is in the lines of iron, which are usually conspicuous, but almost vanish from the spot-spectrum at the sun-spot maximum. At times, also, the spectrum of a spot gives evidence of violent motion in the overlying gases by distortion and displacement of the lines. When the phenomenon FIG. 36. DI Da n* REVEBSAL OF TOE D-LiNE8. occurs, it is more usually at points near the outer edge of the penumbra than over the central portion of the spot ; but, occasionally, the whole neighborhood is vio- lently agitated. In such cases it often happens that lines in the spectrum side by side are affected in en- tirely different ways one will be greatly displaced, -while its neighbor is not disturbed in the least, showing that the vapors which produce the lines are at different levels in the solar atmosphere, and do not participate to any great extent in each other's movements. It is an important fact that the same thing is often true of lines which are ascribed to a single substance : of two iron lines, for instance, one may be disturbed 136 THE SUN. and another unaffected. Mr. Lockyer lays great stress on this as supporting his dissociation hypothesis ; but other explanations are also available, see page 91. In a few instances the gaseous eruptions in the neighborhood of a spot are so powerful and brilliant that, with the spectroscope, their forms can be made out on the background of the solar surface in the same way that the prominences are seen at the edge of the sun. In fact, there is probably no difference at all in the phenomena, except that only prominences of most unusual brightness can thus be detected on the solar surface. An occurrence of this kind fell under the writer's observation on September 28, 1870. A large spot showed in the spectrum of its umbra all the lines of hydrogen, magnesium, sodium, and some others, re- versed. Suddenly the hydrogen lines grew greatly brighter, so that, on opening the slit of the spectroscope, two immense luminous clouds could be made out, one of them nearly 130,000 miles in length, by some 20,000 in width, the other about half as long. They seemed to issue at one extremity from two points near the edge of the penumbra of the spot. After remaining visible about twenty minutes, they faded gradually away, with- out apparent motion. In addition to spots, such as we have been dealing with, there are occasionally seen on the solar surface dark-gray patches, which Trouvelot, who first called attention to them in 1875, has named " veiled spots," * considering that they are essentially of the same nature as other spots, but differing in this, that the disturb- ance which generates them is not sufficiently powerful to reach the surface and break entirely through the * For Trouvclot's account, of them, see " American Journal cf Science and Art," March, 1876, Third Series, vol. xi. SUN-SPOTS AND THE SOLAR SURFACE. 137 photosphere. Over these veiled spots the bright gran- ules are less numerous and smaller than elsewhere, but much more mobile ; sometimes, and frequently indeed, they are overlaid by faculae. The changes of form and appearance in these objects are very rapid, affairs of a minute or two only, according to Trouvelot. They are found all over the solar surface, not being at all con- fined to the regions occupied by the ordinary spots, but sometimes occurring within eight or ten degrees of the sun's pole. They have been little observed, how- ever, and information respecting them is as yet very meager. ROTATION OF SUN AND PROPER MOTIONS OF SPOTS. We have already mentioned that the spots travel across the disk of the sun, from the eastern edge to the western, in such a manner as to show that they arc attached to the surface, and that the sun rotates upon its axis. The true period is about twenty-five days,* the apparent or " synodic " period being some two days longer, because the earth itself is continually moving forward in its orbit. When we come, however, to study the motions of the spots more carefully, we find that they have move- ments of their own (proper motions, as astronomers call them), both in latitude and longitude, so that no observa- tions of any single spot, however carefully conducted, * It is perhaps worth noting that, between the sun and the earth's magnetism, there is an unquestionable, though still unexplained, connec- tion, which shows itself in many ways. Among the numerous periodic variations of this magnetism Ilornstein finds one with a period of 26'32 days. Assuming this to be due to the sun's synodic rotation, he gets 21-55 days for the true rotation. Very similarly Bigelow deduces 24'86. Ycedcr's " aurora period " (27'28 days) gives 25 - 38 all of which may be taken for what it is worth. 138 THE SUN. can furnish an accurate determination of the position of the sun's axis and its period of rotation. This fact does not seem to have been comprehended by the early ob- servers (though a neglected remark of Schemer's indi- cates that he had a glimpse of the truth), and hence we have serious discordances between their diiferent results, which range from 25 '01 days, the result obtained by Delambre in 1775, to 25*58 days, as determined by Cassini about a hundred years earlier. The different values for the inclination of the sun's equator to the ecliptic lie between 6 and 7^, and those for the lon- gitude of the node between 70 and 80. The most reliable recent results are those of Carrington and Spoerer. The former makes the mean period of the sun's rotation 25'3S days, while Spoerer gives it as 25-23. THE EQUATORIAL ACCELERATION. The researches of Carrington,* between 1853 and 1861, first brought out clearly the fact that, strictly speaking, the sun, as a whole, has no single period of rotation, but different portions of its surface perform their revolutions in different times. The equatorial re- gions not only move more rapidly in miles per hour than the rest of the solar surface, but they complete the entire rotation in shorter time. If we deduce the period by means of spots near the sun's equator, we shall find it to be very nearly 25 days, a trifle less according to Carrington. Spots at a solar latitude of *A memoir by Laugier, presented to the French Academy in 1844, but never published in extenso, contains, according to Fare, data which would lead to the same result. The summary, given in the " Comtes Rendus," fails, however, to indicate any appreciation of the systematic variation of rotation rate from equator to poles, and in no way invali- dates Carrington's claim to be considered the discoverer of the law. SUN-SPOTS AND THE SOLAR SURFACE. 139 20 have, on the other hand, a period nearly 18 hours longer ; at 30 the period rises to 26 days, and at 45 to 27|, though in this latitude there are so few spots that the determination is not very reliable. Beyond this latitude we have nothing satisfactory, and it is not possible to determine, with any certainty, whether this retardation continues to the pole or not. It is a curious circumstance, probably connected with this remarkable law of surface-movement, that the spots mostly lie between ten and thirty-five degrees of latitude on each side of the sun's equator ; and it is this fact which makes it difficult to ascertain the exact laws of the solar rotation, since our observations are confined to such a limited range of latitude. As yet, no points have been found near the sun's poles perma- nent and definite enough to permit precise observations covering a sufficient interval of time. By a discussion of all his observations, more than 5,000 in number, of 954 different groups of spots, Mr. Carrington deduced the expression X = 865' 165' sinH for the daily motion of the surface of the sun in dif- ferent solar latitudes, I representing the latitude in the formula, and X the daily motion in minutes of solar longitude. This, as was said before, would make the rotation period of the sun's equator a little less than 25 days. The expression, however, is purely empirical, and no imaginable theoretical explanation can be given for the fractional exponent . Faye, assuming on theoretical grounds that this ex- ponent ought to be 2, finds from the same observations the formula X = 862' 186' sinV, an expression which agrees with all but a few of the observations nearly as well as Carrington's. Spoerer, from observations of his own, made be- 140 THE SUN. tween 1862 and 1868, and combined with those of Secchi and others, derives the still different formula, X = 1011' 203' sin(41 13'+ I). Tisserand, from observations of 325 spots in 1874- '75, deduces the expression X = 857''6 157''3 sinV. But this is probably less reliable than either of the pre- ceding, being founded on a much smaller number of ob- servations. Wilsing, of Potsdam, in 1888 published a discussion of several hundred faculce shown on their phothelio- graph plates, and deduced a rotation-period of 25'23 days ; but he found no indications of equatorial acceler- ation, and concluded that this peculiarity of the photo- sphere, where the spots have their residence, does not extend to the region of the faculse a very perplexing fact, if real. Still more recently, however, Stratonoff, of Pulkowa, from a discussion of their plates, finds from the faculse a result quite in accordance with those of Carrington and Spoerer. We have already referred to the evidence of the sun's rotation given by the spectroscope (page 100), and have specially quoted the remarkable work of Duner. His results show (we think conclusively, though objections have been raised in certain quar- ters) that the region in which the dark lines of the spectrum originate share perfectly the motion of the photosphere. His observations, moreover, have this great advantage over those made on spots and faculae, that they extend as far as 75 on each side of the sun's equator. The observations~~are very well represented by the equation X = 846' 272'"4 sin'Z. This would cor- respond to a rotation-period of 25'53 days at the sun's equator and about 37'5 at the pole but the polar period is very uncertain. SUN-SPOTS AND THE SOLAR SURFACE. 141 While either of the formulae given above agrees fairly with the facts observed, neither of them can be regarded as logically established upon a sound physical explanation. The cause of this peculiar surface-drift is not yet known. Sir John Herschel was disposed to attribute it to the impact of meteoric matter striking the sun's surface mainly in the neighborhood of the equator, and so continually accelerating its rotation, as a boy's peg- top is whipped up by the skillfully applied lash. Per- haps there is nothing absurd in the idea that a sufficient quantity of meteoric matter may reach the sun, or that the meteors move, for the most part, in the plane of the sun's equator, and direct, i. e., with and not against the motion of the planets so that their fall would be mostly confined to the equatorial regions, and would thus hasten, and not retard, the surface motion. But then the duration of the sun's rotation period should continually grow shorter, an effect which does not appear from a comparison of Scheiner's results witli those most recently obtained. Of course, it may be that such an acceleration has actually occurred, only too small to be yet detected ; still, it would seem probable that any " driving," sufficient to establish nearly two days' difference between the rotation periods at the equator and at latitude 40, must have produced a very sensible effect within three hundred years. It is more probable that the equatorial acceleration is connected in some way with the exchange of matter which, if the sun is for the most part gaseous, as now seems likely, must continually be going on between the outside and inside of the globe. If the photosphere is formed of masses falling, such an effect would be a necessary consequence. If we suppose that the out- 142 THE SUN. rushing streams of lieated gas and vapor, as they rise, continue in the gaseous condition until they reach the summit of their ascent, and remain at this height long enough to acquire sensibly the rotation velocity corre- sponding to their altitude, and that then the products of condensation, resulting from their cooling, fall down- ward, and thus falling constitute the photosphere, we should have precisely the actual phenomenon. The ro- tation velocity of each visible element of the photosphere would be that corresponding to a greater altitude, and therefore greater than that naturally belonging to its observed position, and this difference would vary from the equator, where it would be a maximum, to the poles, where it would vanish. Of course, it is not necessary to such an effect that the conditions supposed should be rigidly complied with ; it will suffice to admit that in the photosphere the falling masses are more conspicuous than those which are ascending or stationary, and it would seem hardly possible that it should be otherwise. Whether, however, the effect thus produced would account in measure as well as kind for the observed phenomena, is a question requiring for its answer a more thorough mathematical investigation than the writer has yet been able to undertake. If we consider only the spots, it would seem entirely possible that they may be produced by matter which has fallen from a height of even fifteen or twenty thou- sand miles, and that fall would be quite sufficient to ac- count for their whole acceleration. The fact that rapid changes in the configuration of a spot are generally accompanied by an eastward rush of the whole, also favors the idea that a downfall of something from above is concerned in the matter. SUN-SPOTS AND THE SOLAR SURFACE. 143 If we rightly understand the matter, this theory of the equatorial acceleration is in substantial accordance, so far as it goes, with that formulated some years later by Mr. Lockyer, and given in the last chapter of his " Chemistry of the Sun." But his " dissociation theory " apparently has an important role to play in providing the " hundreds of millions of tons " of falling matter that produce the phenomena by their " down rush." Schaeberle also attributes the equatorial acceleration to the falling back of material that has been projected to a great elevation above the photosphere. The idea of Faye appears to have been nearly the reverse of that here suggested. He attributes the for- mation of the photosphere to gaseous matter not falling from above, but ascending from, below, and starting from a stratum at a certain depth below the surface ; by supposing the depth of this stratum to vary with the latitudes, being greatest at the poles of the sun and least at the equator, it is easy to explain on this hy- pothesis the accelerated motion of the surface at the equator, and to justify his formula, which makes the retardation at higher latitudes proportional to the square of the sine of the latitude ; but no reason is evident why the depth of this stratum should vary. Certain later investigations in 188G upon the rota- tion of fluid masses, by Jukowsky, of Moscow, as ap- plied to solar conditions by his colleague Belopolsky, perhaps warrant a hope that the phenomena of surface- drift in longitude, and even the periodicity of the spots, ultimately find a rational explanation as necessary re- sults of the slow contraction of a non-homogeneous and mainly gaseous globe* The subject is difficult and obscure ; but if it can be proved, as seems not impossi- ble, that, on mechanical principles, the time of rotation 144 THE SUX. of the central portions of such a whirling mass must be shorter than that of the exterior, then there will be, of necessity, an interchange of matter between the inside and outside of the sphere, a slow surface-- drift from equator towarrd the poles, a more rapid inter- nal current along and near the axis, from the poles toward the equator, a continual " boiling up " of internal matter on each side of the equator, and, finally, just such an eastward drift near the equator as is actually observed. Moreover, the form of the mass, and the intensity of the drift and consequent " boiling up " from underneath might, and probably would, be subject to great periodical variations. As to Zollner's idea that the equatorial acceleration is due to the friction between a liquid sheet, constitut- ing the photosphere, and a solid nucleus below, it is hardly necessary to say that this view is in complete opposition to those held by almost all astronomers, and seems to be untenable in its fundamental assumptions. On the whole, however, the writer sympathizes with Duner in his conclusion : " I must confess that this dif- ference between the rotation periods in the different (solar) latitudes appears to me incomprehensible, and constitutes one of the most difficult problems of astro- physics." No theory yet presented is really satisfactory. THE POSITION OF THE SUN'S AXIS. The plane of the sun's rotation is slightly inclined to that of the earth's orbit. According to Carrington, the angle is 7 15', while Spoerer makes it 6 57'. This plane cuts the ecliptic at two opposite points called the nodes, one of which is in longitude 73 40', according to Carrington, and 74 36' according to Spoerer. The axis of the sun is therefore directed to a point in the SUX-SPOTS AND THE SOLAR SURFACE. 145 constellation of Draco, not marked by any conspicuous star. Astronomers define its position by saying that its right ascension is 18 h 44 m , and its declination is 64. It is almost exactly half- way between the bright star a Lyrse and the polar star. The earth passes through the two nodes on or about the 3d of June and the 5th of December. At these times the spots move apparently in straight lines across the sun's disk, and his poles are situated on its circum- ference. During the summer and autumn, from June to December, the sun's northern pole is inclined toward the earth ; during the winter months, the southern. The angle which the sun's axis appears to make with a north and south line in the sky (technically, the position.-angle of the sun's axis) changes considerably during the year, varying 26 each side of zero. As it is often very desirable for an amateur to know this angle approxi- mately, we insert the following little table, giving the position angle of the sun's north pole referred to the center of the disk. The table is derived from the much more extensive one in Secchi's " Le Soleil " : POSITION ANGLE OF SUN'S AXIS. JANUARY 4, JULY 6 0-00. Jan. 15, June 25.. 5 west. Dec. 24, July 17. . I 5 east. Jan. 26, June 14.. . 10 west. Dec. 15, July 29. 10 east. Feb. 7, June 2.... 15 west. ; Dec. 3, Aug. 11. . 15 east. Feb. 22, May 18. . 20 west. 1 Nov. 19, Aug. 27. 20 east. March 18, April 25 26 west. Oct. 29, Sept. 20. 25 east. April 5 26* 20' west. Oct. 10 26 20' east. It is understood, of course, that the table is only approximate, because the numbers change slightly ac- cording to the place of the current year in the leap- year cycle ; but the results obtained from it are always 11 146 THE SUN. correct within about |, wliich is near enough for most purposes. Fig. 37 illustrates these points, giving the position- angle of the sun's axis, and the aspect of his equator at different times of the year as seen from the earth. For FIG. 37. SEPT. 4. DEC. 5. MARCH 7. POSITION-AXULK OF ScVS AXIS, AND ASPECT OF HIS EQUATOR. the sake of clearness, however, the inclination of the sun's equator to the ecliptic is considerably exaggerated in the lower row of figures : the equator never appears so strongly curved as there represented. PROPER MOTION OF SPOTS. After making due allowance for the equatorial acceleration, it is found that almost every spot has more or less motion of its own. Between latitudes 20 north and 20 south, Mr. Carrington finds, on the whole, a slight tendency to motion toward the equator, the move- ment amounting to a minute or two of arc per diem ; from 20 to 30 on both sides of the equator, there is SUN-SPOTS AND THE SOLAR SURFACE. . 147 a somewhat more decided motion toward the poles. Faye has also shown that many spots move in small ellipses upon the surface of the sun, completing their circuits in a day or two, and repeating them with great regularity for weeks, and even months. Whenever a spot is passing through sudden changes, it generally moves forward upon the solar surface, as has already been mentioned, with something like a leap ; and, when a spot divides into two or more, the parts generally sep- arate with a very considerable velocity, as if (we do not say because) there was a repulsion between them. DISTRIBUTION OF SUN-SPOTS. The sun-spots, as has already been said, are not dis- tributed over the sun's surface with anything like uni- formity. They occur mainly in two zones on each side of the equator and between the latitudes of 10 and 30. On the equator itself they are comparatively rare ; there are still fewer beyond 35 of latitude, and only a single spot has ever been recorded more than 45 from the solar equator one observed in 1846 by the late Dr. Peters, then in Naples. The figure shows the distribution of 1,386 spots ob- served by Carrington. The figure is constructed in this way : The circumference of the sun, on the left- hand side of the figure, is divided into five-degree spaces from the equator each way, and at each of them is erected a radial line whose length in four hundredths of an inch is proportional to the number of spots ob- served within 2 of latitude on each side. Thus, the line drawn at 20 north latitude, and marked " 151," is | of an inch long, and means that 151 spots were recorded between 17^ and 22 north latitude. It is at once evident from mere inspection that the THE SUN. distribution follows no simple law of latitude. On the northern hemisphere, the distribution, during the eight years over which the observations extend, was not very irregular, though there is a distinct minimum at 15, DlSTBIBTJTION OF SUN-SPOTS AND PEOTtTBKEANCES. and two maxima at about 11 and 22 of latitude. On the southern hemisphere the minimum at 15 is very marked, and the numbers at 10 and 20 are far in excess of those in the northern hemisphere. Of the whole number, 711 were in the southern hemisphere, as against 675 in the northern. The minimum at 15 of latitude was special to the date of observation, and had its origin in a law discov- ered by Spoerer a few years ago to be discussed later (page 156). His own observations from 1861 to 1867 show nothing of the kind. They give the following SUX-SPOTS AXD TIJE SOLAR SURFACE. 149 distribution of 1,053 spots in latitude, viz. : -{- 35, 4 ; + 30, 4 ; + 25, 16 ; + 20, 50 ; + 15, 133 ; + 10, 198; +5, 114 in all, 519 spots north of the solar equator. 40 spots were on the equator, or within 2 of it. South of the equator we have, in latitude : 5, 113 ; - 10, 206 ; - 15, 109 ; - 20, 38 ; - 25, 19 ; - 30, 7 ; - 35, 1 ; - 40, 1 in all, 494 southern spots. In 1866, a year of spot minimum, there were only 94 spots in all, and of these 94 all but two were situated within 17 of the equator. It is to be noticed that at times when spots are abundant their mean latitude is greater than when they are few, or, in other words, an increase in the number of spots generally carries with it a widening of the zones in which the spots appear. All the observations concur in showing this. The cause of this distribution of the spots in zones is not known. It is probably connected with the origin of the spots themselves, and very possibly has something to do with the law of surface-motion just discussed. At least it is certain, as Faye pointed out some years ago, that, while at the solar poles and equator adjoining por- tions of the photosphere have no relative motion with reference to each other, yet in the middle latitudes this is not true ; here each element of the surface has a different velocity from those immediately north and south of it, so that they drift by each other like the filaments of a liquid current which is suffering retarda- tion, producing, as Faye supposes, whirlpools and eddies which, according to his view, generate the spots. As regards the sun's northern and southern hemi- spheres, there is often a great inequality. Thus, from 1672 to 1704 absolutely no spots were recorded in the northern hemisphere, and when a few appeared in 1705 ]50 THE SUN. and 1714, the French Academy was formally notified of the fact as something very remarkable. We do not know that anything quite like this has ever happened since ; but the inequality between the two hemispheres is often very marked for mouths together, though in the long run there seems to be no difference. It is a question of much theoretical importance whether spots do or do not appear repeatedly at the same points ; for if this is really the case, it would make it almost certain that below the photosphere there must be a coherent nucleus, carrying with it in its rota- tion such volcanic or otherwise peculiar regions as to cause the breaking out of spots above them. There would be no difficulty in accounting for two or three dissolutions and reappearances in the same region with- out any such hypothesis, since a great disturbance in the solar atmosphere would not subside entirely for a long time. The observations of Spoerer show that this actually happens, and that, for a period of several months, spots and faculae often recur several times at the same point. But his observations do not give any real support to the idea of a solid nucleus, nor has he himself ever favored such a view, although some (and among others the writer), misunderstanding certain ex- pressions of his, have supposed that he did. CHAPTER Y. PERIODICITY OF SUN-SPOTS; THEIR EFFECTS UPON THE EARTH, AND THEORIES AS TO THEIR CAUSE AND NATURE. Observations of Schwabe. Wolf's Numbers. Proposed Explanations of Periodicity. Connection between Sun-Spots and Terrestrial Magnet- ism. Remarkable Solar Disturbances and Magnetic Storms. Effect of Sun-Spots on Temperature. Sun-Spots, Cyclones, and Rainfall. Researches of Symons and Meldrum. Sun-Spots and Commercial Crises. Galileo's Theory of Spots. Herschel's Theory. Secchi's First Theory. Zollner's. Faye's. Secchi's Later Opinions. Theo- ries of Lockyer, Schaeberle, and others. IT was early noticed that the number of sun-spots is very variable, but the discovery of a regular periodicity in their number dates from 1851, when Schwabe, of Dessau, first published the result of twenty -five years of observation. During this time he had examined the sun on every clear day, and had secured an almost per- fect record of every spot that appeared upon the solar surface. He began his work without any idea of ob- taining the result he arrived at, and says of himself, that, " like Saul, he went to seek his father's asses, and found a kingdom." His observations showed unmis- takably that there is a pretty regular increase and de- crease in the number of sun-spots, the interval from one maximum to the next being not far from ten years. Subsequent observations and a thorough examination of all known former records fully confirm this conclu- 152 THE SUN. sion, except that the mean period appears to be some- what greater, eleven and one ninth years being the value at present generally received. Professor R. Wolf, of Zurich, has, been especially indefatigable in his investi- gations upon this subject, and has succeeded in disinter- ring from all sorts of hiding-places a nearly complete history of the solar surface for the past hundred and fifty years. Among other things he finds among the unpublished manuscripts of Horrebow (a Danish as- tronomer who flourished a century ago) a distinct in- timation (in 17T6) that zealous and continued observa- tion of the sun-spots might lead to " the discovery of a period, as in the motions of the other heavenly bodies," with the added remark that " then, and not till then, it will be time to inquire in what manner the bodies which are ruled and illuminated by the sun are influ- enced by the sun-spots " alluding, perhaps, to certain ideas then, as now, more or less current, and illustrated by the attempt of Sir W. Herschel, a few years later, to establish a relation between the price of wheat and the number of sun-spots. Wolf has brought together an enormous number of observations, and with immense labor has combined them into a consistent whole, deducing a series of " rel- ative numbers," as he calls them, which represent the state of the sun as to spottedness for every year since 1745. His "relative number" is formed in rather an arbitrary manner from the observation of the spots : representing this number by r, the formula is, r^= #(/+ 10g\ in which g is the number of groups and isolated spots observed, and f the total number of spots which can be counted in these groups and singly, while k is a coefficient which depends upon the ob- server and his telescope. Wolf takes it as unity for PERIODICITY OF SUN-SPOTS. 153 154 THE SUN". himself, observing with a three-inch telescope and power of 64. For an observer with a larger instrument, k would be a smaller quantity, while a less powerful instrument and less assiduous observer would receive a "&" greater than unity, as probably seeing fewer spots than Wolf himself would reach with his instrument. These rela- tive numbers, as tested by the most recent photographic results of De La Hue and Stewart, are found to be quite approximately proportional to the area covered by the spots. We give on the opposite page a figure deduced from the numbers, published by Wolf in 1877, in the " Me- moirs of the Royal Astronomical Society," and showing their course year by year since 1772. The continua- tion* of the curve to 1880 is from numbers subse- quently published by tu m in the astronomical periodi- cals. The horizontal divisions denote years, and the height of the curve at each point gives the " relative number" for the date in question. For example, in 1870, about the middle of the year, the relative num- ber was 140, while early in 1879 it ran as low as 3. The dotted lines are curves of magnetic disturbance, with which at present we have no concern. Our dia- gram, on account of the smallness of the page, only goes back to 1772, but Wolf's investigations reach to 1610, and he gives, in the paper from which were derived the numbers used in constructing our diagram, the follow- ing important table of the maxima and minima of sun- spots since that date, dividing the results into two series, the first of which, from the paucity of observations, is to be considered of much inferior weight to the second. * It did not seem worth while to re-engrave the plate in order to bring the curve down to date, but the main results since 1880 are stated nu- merically a page or two later. PERIODICITY OF SCX-SPOTS. 155 FIRST SERIES. SECOND SERIES. Minima. Maxima. Minima. Maxima. 1610-8 1615-5 1745-0 1750-3 8-2 10.5 10-2 11-2 1619-0 1626-0 1755-2 1761-5 15-0 13-5 111 8-2 1634-0 1639-5 1766-5 17697 iro 9-5 9-0 8-7 1645-0 1649-0 1775-5 1778-4 10-0 11-0 92 9-7 1655-0 1660-0 . 1784-7 1788-1 11-0 15-0 13-6 16-1 1666-0 1675-0 1798-3 1804-2 13-5 10-0 12-3 12-2 16795 16850 1810-6 1816-4 10-0 8-0 12-7 13-6 1689-5 1693-0 1823-3 1829-9 8-5 12-5 10-6 7-3 1698-0 1705-5 1833-9 1837-2 140 12-7 96 10-9 1712-0 1718-2 18435 1848-1 11-5 9-3 12-5 120 1723-5 1727-5 1856-0 1860-1 10-6 11-2 11-2 10-5 1734-0 1738-7 1867-2 1870-6 Mean period. Mean period. Mean period. Mean period. 11'20 2-11* 11-20 2-06 11-16 1-54 10-94 2-52 0-64 0-63 0-47 0-76 From these data, 'Wolf derives a mean period of ll'lll years, with an average variability of 2'03 years, and an uncertainty of 0'307, due chiefly to the difficulty of fixing the precise date of maximum or minimum. After the great maximum of 1871*6, when the rel- ative number reached 140, there was an unusually pro- tracted down-slide until 1879, when, as the figure shows, *The upper number, 2'11, indicates that the individual periods have an average variation of 2'11 years on one side or the other from the mean period. The lower number, 0*64, is the so-called " probable error " of the period. Similarly in the three other columns. 156 THE SUN. a very low minimum occurred. After that a feeble maximum (only 64) arrived pretty quickly near the end of 1883, followed by an average minimum in the middle of 1889. The next and last maximum was passed in 1893 ; it was not a very high one, perhaps about 70 but Wolf died in 1893, and we have no authentic figures later than 1891. A moment's inspection of the curve shows that the maxima differ greatly in intensity, and that the period is not at all fixed and certain like that of an orbital mo- tion, but is subject to great variations. Thus, between the maxima of 1829*9 and 1837*2 we have an interval of only 7'3 years, while between 1788 and 1804 it was 16'1 years.* A portion of this great variableness of period may, perhaps, be due to the incompleteness of our observations, but only a portion. It is quite likely that a fluctuation of much longer period, not far from sixty years, is, to some extent, responsible for the effect by its superposition upon the principal (eleven-year) oscillation. Another important fact is that the interval from a minimum to the next following maximum is only about 4% years on the average, while from the maximum to the next following minimum the interval is 6'6 years. The disturbance which produces the sun-spots springs up suddenly, but dies away gradually. Still another fact, as yet unexplained, and probably of great theoretical importance, has recently been brought out by Spoerer. Speaking broadly, the dis- turbance which produces the spots of a given sun-spot period first manifests itself in two belts about 30 north * Some astronomers contend that there ought to be another maximum inserted about 1795. Observations about this time arc few in number and not very satisfactory. PERIODICITY OF SUN-SPOTS. 157 and south of 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, however, 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 to the newly beginning outbreak ; and it appears that the true sun- spot cycle is from twelve to fourteen years long, each be- ginning in high latitudes before the preceding one has expired near the equator. Fig. 40 illustrates this, embodying Spoerer's results from 1855 until 1880. The dotted curves show Wolf's 18 55 18 60 18 65 18 70 18 75 18 30 L. 30" 26" 22 18 14 10 W. 100 50 \ \ 'S \ \ \ \ / \ \ \ ^ *- ~~- - r ' \, ^ -- ^. /' SPOERER'S CURVES OF SUN-SPOT LATITUDE. 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 first series began in 1856 and ended in 1868 ; the second broke out in 1866 and lasted until 1880. During these 158 THE SUN. years it happened that there was very little difference between the northern and southern hemispheres of the sun. EXPLANATIONS OF SUN-SPOT PERIODICITY. There is no question of solar physics more interest- ing or important than that which concerns the cause of this periodicity, but a satisfactory solution remains to be found. It has been supposed by astronomers of very great authority that the influence of the planets in some way produces it. Jupiter, Venus, and Mer- cury have been especially suspected of complicity in the matter, the first on account of his enormous mass, the others on account of their proximity. De La Hue and Stewart deduced from their photographic observa- tions of sun-spots, between 1862 and 1866, a series of numbers, strongly tending to prove that, when two of the powerful planets are nearly in line as seen from the sun, then the spotted area is much increased. They have investigated especially the combined effect of Mer- cury and Venus, Jupiter and Venus, and Jupiter and Mercury, as also the effect of Mercury's approach to, or recession from, the sun. In all four cases there seems to be a somewhat regular progression of num- bers, though much less decided in the third and fourth than in the first and second. The irregular variations of the numbers are, however, so large, and the duration of the observations so short, that it is hardly safe to build heavily upon the observed coincidences, since they may be merely accidental. In fact, so far as we can learn, the observations since 1866 furnish no confirma- tion of this theory. An attempt to connect the eleven-year period with that of the planet Jupiter also breaks down. "While, PERIODICITY OF SUN-SPOTS. 159 for a certain portion of time, there is a pretty good agreement between the sun-spot curve and that which represents the varying distance of Jupiter from the sun, there is complete discordance elsewhere. About 1870 the maximum spottedness occurred when the planet was nearest the sun, but at the beginning of the century the reverse was the case. Loomis suggested that the con- junctions and oppositions of Jupiter and Saturn may be at the bottom of the matter. These occur at inter- vals of 9'93 years, from a conjunction to an opposition, or vice versa. But, when we come to test the matter, we find that, in some cases, sun-spot minima have coin- cided with this alignment of the two planets ; in other cases, maxima. It is, indeed, very difficult to conceive in what man- ner the planets, so small and so remote, can possibly produce such profound and extensive disturbances on the sun. It is hardly possible that their gravitation can be the agent, since the tide-raising power of Venus upon the solar surface would be only about TJ-^ of that which the sun exerts upon the earth ; and in the case of Mercury and Jupiter the effect would be still less, or about -j-gVj of the sun's influence on the earth. Making all allowances for the rarity of the materials which com- pose the photosphere, it is quite evident that no planet- lifted tides can directly account for the phenomena. If the sun-spots are due in any way to planetary action, this action must be an occasion rather than a cause. A minute disturbance may, so to speak, " pull the trigger " and bring on an explosion. The touch of a child's finger fired the Flood Eock mine. Several astronomers, among others Professor B. Peirce, seem to have adopted an idea before alluded to first suggested, we believe, by Sir John Herschel 160 THE SUN. that the spots are caused by meteors falling upon the sun. According to this view, the periodicity of the spots could be simply accounted for by supposing the meteors to move in a very elongated orbit, with a pe- riod of 11*1 years, adding the additional hypothesis that at one part of the orbit they form a flock of great density, while elsewhere they are sparsely distributed. This meteoric orbit would have to lie nearly in the plane of the sun's equator, and have its aphelion near the orbit of Saturn. Of course there is no necessity to limit our hypothesis to a single meteor stream. What we know of meteor-showers encountered by the earth, makes it likely that there may be several, of different periods ; and thus we may account for some of the ob- served irregularities of the sun-spot period. The hy- pothesis has many excellent points, and we shall have occasion to recur to it again. At the same time, it may be said here that it seems very difficult to make it ex- plain the enormous dimensions and persistence of many sun-spot groups, and the distribution of the spots on the sun's surface in two parallel zones, with a minimum at the equator. The irregularity in the epochs of max- ima and minima is also much greater than would have been expected. On the whole, it seems rather more probable that the periodicity is in the sun itself, depending upon no external causes, but upon the constitution of the photo- sphere and the rate at which the sun is losing heat. Perhaps we may compare small things with great by referring to the periodic explosions of the Icelandic geysers, or the " bumping " of ether and many other liquids in a chemist's test-tube. Looking at it in this light, we should imagine the course of events to consist of a gathering of deep-lying forces during a season of PERIODICITY OF SUN-SPOTS. 161 external quiescence, followed by an outburst, which relieves the internal fury ; the rest and the paroxysms recurring, at somewhat regular intervals, simply because the forces, materials, and conditions involved, change only slowly with the lapse of time. If sucli be really the case, it is clear, of course, that this periodicity is never likely to be very regular, and will not long keep step with any planetary march. Time of itself, therefore, will by-and-by solve the prob- lem for us, or at least will refute any false hypothesis resting upon the recurrence of planetary positions. TERRESTRIAL INFLUENCE OF SUN-SPOTS. Even more important than the problem of the cause of sun-spot periodicity, is the question whether this pe- riodicity produces any notable effects upon the earth, and, if so, what ? In regard to this question the astro- nomical world is divided into two almost hostile camps, so decided is the difference of opinion, and so sharp the discussion. One party holds that the state of the sun's surface is a determining factor in our terrestrial meteor- ology, making itself felt in our temperature, barometric pressure, rainfall, cyclones, crops, and even our financial condition, and that, therefore, tho most careful watch should be kept upon the sun for economic as well as scientific reasons. The other party contends that there is, and can be, no sensible influence upon the earth produced by such slight variations in the solar light and heat, though, of course, they all admit the connection between sun-spots and the condition of the earth's magnetic elements. It seems pretty clear that we are not in a position yet to decide the question either way ; it will take a much longer period of observation, and observations con- 12 162 TIIE SUN - ducted with special reference to the subject of inquiry, to settle it. At any rate, from the data now in our pos- session, men of great ability and laborious industry draw opposite conclusions. It certainly is not so plain that the sun-spots have not the influence which their worshipers, I had almost called them, claim for them, as to absolve us from the duty of investigating the matter in the most thor- ough manner. On the other hand, it is also by no means certain that we shall find the labor of investiga- tion fruitful in precisely the manner and degree desired. Those who search for truth with honest endeavor may, nevertheless, be sure of their reward in some way. I have said that there is no doubt as to the con- nection between the sun-spots and terrestrial magnetism. In 1850, Lament, of Munich, called attention to the fact that the average daily excursions of the magnetic needle have a period which, from the few decades of observation at his command, he fixed at ten and one third years. Perhaps a word of explanation is needed here. Every one knows that the compass-needle does not point exactly north, and its divergence from the true meridian is different in different places. On the At- lantic coast of the United States, for instance, the north pole of the magnet points west of north, and on the Pacific coast east of north. What is more : at any par- ticular place the direction of the needle is continually changing, these changes being like the changes in the temperature of the air, in part regular and predictable, and partly lawless, so far as we can see. One of the most noticeable of the regular magnetic changes is the so-called diurnal oscillation ; during the early part of the day, between sunrise and one or two MAGNETISM AND SUN-SPOTS. 163 o'clock p. M., the north pole of the needle moves toward the west in these latitudes, returning to its mean position about 10 P. M., and remaining nearly stationary during the night. The extent of this oscillation in the United States is about 15' of arc in summer, and not quite half as much in winter ; but it differs very much in different localities and at different times, and also and this is Lamont's discovery the average extent of this diurnal oscillation at any given observatory increases and de- creases pretty regularly during a period of 10 years, according to his calculations. As- soon as Schwabe an- nounced his discovery of the periodicity of the solar spots, Sabine in England, Gautier in France, and Wolf in Switzerland, at once and independently perceived the coincidence between the spot-maxima and those of the magnetic oscillation. Faye at one time attempted to impugn this conclusion. In order to make his point, he insisted that the magnetic maximum is shown by Cas- sini's observations to have occurred early in 1787, and, dividing the interval between this and the last magnetic maximum, near the close of 1870, by 8, the number of intervening periods, he gets 10'45 years for the mean magnetic period, instead of ll'll. The reply is, that the observations both of the sun-spots and of the mag- netic elements near the close of the eighteenth century are so meager and unsatisfactory that the evidence as to the precise time of maxima and minima is very in- complete. In 1885, however, Faye yielded to the con- stantly accumulating weight of evidence, and gave in his adhesion to the received conclusion, which is now practically undisputed. The convincing evidence as to the reality of the as- serted connection lies in the closeness with which, ever since we have been in possession of continuous and sat- 1(54 THE SUN. isfactory observations, the magnetic curve copies that of the sun-spots. In Fig. 39 the dotted curves represent the mean amount of magnetic oscillation as deduced by Wolf from various series of observations. From 1820 to 1895 the record is almost continuous, and the coinci- dence of the curves is such as to make it impossible to doubt the connection.* The argument is much strengthened by an examina- tion of records of the aurora borealis. Occasionally so- called " magnetic storms " occur, during which the com- pass-needle is sometimes almost wild with excitement, oscillating 5 or even 10 within an hour or two. These " storms " are generally accompanied by an aurora, and an aurora is always accompanied by magnetic disturb- ance. Now, when we come to collate aurora observations with those of sun-spots, as Loomis has done with great care and thoroughness, we find an almost perfect paral- lelism between the curves of auroral and sun-spot fre- quency. We find also, as Shearman, of Toronto, and Dr. Veeder, of Lyons, 1ST. Y., have pointed out, that auroras often run in series, so to speak, following each other for several months at nearly regular intervals of 27*275 days, which is very closely the period of the sun's ap- parent equatorial (synodic) rotation ; this of course makes it more or less probable that their appearance is connected somehow with the way in which certain por- tions of the sun's surface present themselves to the earth. Dr. Veeder's idea is that disturbed regions upon * A discussion, by Balfour Stewart, of the observations at Kew, be- tween 1856 and 1867, brings out the correspondence very beautifully, and seems to show that the magnetic changes lag behind the sun-spots about five months. MAGNETISM AND SUN-SPOTS. 165 the sun are specially influential upon the earth's mag- netism at the moment when they are near the eastern edge of the sun, and just coming in sight to us on the earth. There is no obvious reason, however, why a disturbance on the sun should thus propagate itself more vigorously in a direction tangential to the sun's surface and in the plane of the sun's equator than in any other direction ; and while Dr. Yeeder is certainly able to marshal a great number of coincidences in support of his opinion, there are also numerous cases where the region of solar disturbance was near the middle of the sun's disk, as, for instance, the great magnetic storms and auroras of February 13, 1892, and November 17, 18S2. In this connection it seems worth while to quote from an article by Mr. Maunder, of the Greenwich Observa- tory, with respect to these spots, and two other groups of almost equal magnitude which appeared together in April, 1882. He writes : u In a period of nearly nineteen years, therefore " (from 18V3 to 1892), ' we have three magnetic storms which stand out pre- eminently above all others during that interval. In that same period we have three great sun-spot displays counting the two groups of April, 1882, together which stand out with equal dis- tinctness far above all other similar displays. And we find that the three magnetic storms were simultaneous with the greatest de- velopment of the spots. Is there any escape from the conclusion that the two have a real and binding connection? It may be di- rect; it may be indirect and secondary only; but it must be real and effective." "Knowledge," May, 1892. It is not easy to frame any satisfactory theory to ac- count for this connection between solar disturbances and terrestrial magnetism. It can hardly be in the way of temperature, for the influence of sun-spots in this re- spect is so slight that it is still an open question whether we do or do not get from the sun more than the average 166 THE SUN. amount of heat during a sun-spot maximum. Probably it is more immediate and direct ; perhaps in some way kindred with the action which drives off the material of a comet's tail, and proves that other forces besides grav- itation are operative in interplanetary space. Or, not at all impossibly, as Mr. Maunder suggests, it may be indirect ; an action of some cosmic cause upon sun and earth together. There are a number of observed instances which, though not sufficient to demonstrate the fact, still ren- der it very probable that "every intense disturbance of the solar surface is propagated to our terrestrial mag- netism with the speed of light. An instance fell under the writer's notice in the course of a series of spectro- scopic observations at Sherman. On August 3, 1872, the chromosphere in the neighborhood of a sun-spot, which was just coming into view around the edge of the sun, was greatly disturbed on several occasions during the forenoon. Jets of luminous matter of intense bril- liance were projected, and the dark lines of the spectrum were reversed by hundreds for a few minutes at a time. There were three especially notable paroxysms at 8.45, 10.30, and 11.50 A. M. local time. At dinner the pho- tographer of the party, who was determining the mag- netic constants of our station, told me, without knowing any thing about my observations, that he had been obliged to give up work, his magnet having swung clear off the scale. Two days later the spot had come around the edge of the limb. On the morning of August 5th I began observations at 6.40, and for about an hour wit- nessed some of the most remarkable phenomena I have ever seen. The hydrogen lines, with many others, were brilliantly reversed in the spectrum of the nucleus, and at one point in the penumbra the C line sent out what MAGNETISM AND SUX-SPOTS. 167 looked like a blowpipe-jet, projecting toward the up- per end of the spectrum, and indicating a motion along the line of sight of about one hundred and twenty miles per second. This motion would die out and be renewed again at intervals of a minute or two. The figure gives an idea of the appearance of the spectrum. The dis- turbance ceased before eight o'clock and was not re- newed that forenoon. On writing to England, I re- FIG. 41. ceived from Greenwich and Stonyhurst, through the kindness of Sir G. B. Airy and Rev. S. J. Perry, copies of the photographic magnetic records for those two days. Fig. 42 is reduced from the Greenwich curve. That obtained at Stonyhurst is essentially the same. It will be seen that on August 3d, which was a day of gen- eral magnetic disturbance, the three paroxysms I noticed at Sherman were accompanied by peculiar twitches of the magnets in England. Again, August 5th was a quiet day, magnetically speaking, but just during that hour when the sun spot was active, the magnet shivered and 163 THE SUN. trembled. So far as appears, too, the magnetic action of the sun was instantaneous. After making allowance for longitude, the magnetic disturbance in England ap- pears strictly simultaneous, so far as can be judged, FIG. 42. ENWICH (August 8 and 5, 1872). with the spectroscopic disturbance seen on the Rocky Mountains, and the difference can not have been more than about ten minutes. But the time at Sherman was. not noted with any great precision. Of course, as has been said, no two or three coinci- dences such as have been adduced are sufficient to es- tablish the doctrine of the sun's immediate magnetic action upon the earth, but they make it so far probable as to warrant a careful investigation of the matter an investigation, however, which is not easy, since it im- plies a practically continuous watch of the solar surface. MAGNETISM AND SUN-SPOTS. 1G9 It may be added, too, that many striking disturb- ances which have been observed upon the sun, in the ascent of lofty prominences, received no magnetic re- sponse from the earth ; and there have also been great auroras with no obvious solar correlative. Indeed, there is every reason to suppose that a large proportion of all the magnetic disturbances at any given observa- tory are purely local, having nothing whatever to do with the sun. Some also which are not local have been traced to the action of the moon, and it is not at all improbable that others yet are due to causes oper- ating in interplanetary space. Solar disturbances are not the cause of our magnetic storms, but only one cause of some of them ; and very likely a cause only in the sense that the pulling of a trigger " causes " the flight of a rifle-ball : there need be no proportionality between such a cause and its effect. It would be unfair to our readers to pass without notice the remarks of Lord Kelvin in a recent presiden- tial address to the Royal Society (November, 1892), ex- pressing his dissent from the accepted view of the relation we have been discussing. Taking the magnetic storm of June 25, 1885, as an example, he computes that " In tin* eight hours of a not very severe magnetic storm as much work must have been done by the sun in sending magnetic waves out in all directions through space as he actually does in four months of his regular heat and lipht. This result," he adds, "it seems to me, is absolutely conclusive against the supposition that terrestrial magnetic storms are due to magnetic action of the sun, or to any kind of action taking place witliin the sun, or in connection with hurricanes in his atmosphere, or anywhere near the sun outside. It seems as if we may also be forced to con- clude that the supposed connection between magnetic storms and sun-spots is unreal, and that the seeming agreement between the periods has been a mere coincidence." 170 THE SUN. And yet, with all deference to so high an authority, and without questioning the accuracy of his calculations, they seem really to be no more conclusive than a com- putation to show that the work done by an explosion vastly exceeded the power of the person who pressed the firing button. The nature of the mechanism by which the connection is established may, and still does, remain uncertain, but the statistics leave no doubt as to the reality of the connection itself. It is not, per- haps, outside the limits of possibility, as before hinted, that both the solar and terrestrial disturbance have a common origin in some invasion of power or matter from outer space that the solar tumult is the brother and not the father of our own aurora. As to the effect of sun-spots upon terrestrial temper- ature, no conclusion seems possible at present. The spots themselves, as Henry, Secchi, Langley, and others have shown, certainly radiate to us less heat than the general surface of the sun. According to the elaborate determinations of Langley, the umbra of a spot emits about fifty-four* per cent, and the penumbra about eighty per cent, as much heat as a corresponding area of the photosphere. The direct effect of sun-spots is, therefore, to make the earth cooler. As the total area covered by spots, even at the time of maximum, never exceeds -g-J-^ of the whole surface of the sun, it follows * The most recent observations, those made at Daramona, in Ireland, by W. E. Wilson, in 1893, with a " radio-micrometer " and other apparatus of the highest order, give about forty-six per cent, for this ratio. All ob- servers find that it increases near the limb of the sun, and both Langley and Frost have encountered cases where the umbra of a spot was appar- ently warmer than the surrounding photosphere : a fact which, if not the result of some error of observation, is difficult to explain on the theory that spots arc cavities, though a necessary consequence if they, like the faculze, arc masses floating at some elevation above the photosphere. MAGNETISM AND SUN-SPOTS. 1Y1 that directly they may diminish our heat-supply by about y^ of the whole. Whether this effect would be sensible or not, is a question not easily answered. But, while the direct effect would be of this nature, it is quite probable that it is at least fully compensated by another of the opposite character. We get our light and heat from the photosphere which is covered by an atmosphere of gases, and in this atmosphere a consider- able absorption occurs. Now, if the level of the photo- spheric surface be disturbed, so that it is covered with waves and elevations of any considerable height, as compared with the thickness of the overlying atmos- phere, then, as Langley has shown, the radiation will at once be increased ; since, while the absorption is in- creased by a certain percentage for those portions of the photosphere which are depressed below their ordi- nary level, it is much more decreased for those that are raised. The reason of this is that, when a luminous object is immersed in an absorbing medium it loses much more light for the first foot of submergence than for the second, and more for the second than for the third ; so that when it has reached a considerable depth it re- quires an additional submergence of many feet to di- minish its radiation as much as the first foot did. If, therefore, sun-spots are accompanied by considerable vertical disturbance of the photosphere, as is almost certain, we must have as a result an increased radia- tion on account of the disturbance, offsetting, more or less entirely, the opposite effect which is at first view most obvious. Then, again, it is altogether probable that spots are either due to, or accompanied by, an eruptive action the internal, and hotter, gases bursting through the pho- 172 THE SUN. tospliere with unusual abundance during seasons of spot-maximum. This must necessarily tend to increase the emission of heat from the sun, and possibly by a considerable amount. But, on the other hand, any considerable increase in the thickness of the chromo- sphere, such as might result from abundant and long- continued eruption, would work in the opposite direc- tion. It is impossible, therefore, to predict, a priori, which effect will predominate, or to say whether the mean temperature of the earth ought to be raised or lowered during a sun-spot maximum ; and thus far no compari- son of observations has settled the matter to general satisfaction. At least, no longer ago than 1878, Balfour Stewart, who ought to know if any one, writes, " It is nearly, if not absolutely, impossible, from the observa- tions already made, to tell whether the sun be hotter or colder, as a whole, when there are most spots on his surface." On the one hand, Jelinek, from all temperature observations available in Germany up to 1870, found the influence of sun-spots entirely inappreciable, though from the same observations he did deduce minute effects produced by the changes in the distance and phase of the moon. On the other hand, Mr. Stone, while astrono- mer royal at the Cape of Good Hope, and Dr. Gould, in South America, consider that the observations taken at their stations show a distinct though slight diminution of temperature at the time of a sun-spot maximum : according to Dr. Gould the difference at Buenos Ayres between maximum and minimum amounts to about lf Fahr. He also considers that the meteorological records of the Argentine Republic between 1875 and 1885 show a distinct connection between the sun-spots METEOROLOGY AND SUN-SPOTS. If3 and the force and direction of the winds at the various stations. At the Cape of Good Hope, Mr. Stone finds the difference to be about three fourths of a degree from thirty years' observations at least, if we rightly interpret his curve of temperatures, for it is not quite clear what unit of temperature is used in constructing his diagram. At Edinburgh, Piazzi Smyth finds in the records of the rock thermometers a marked eleven-year perio- dicity, of which the range amounts to about a degree (Fahr.), and the maxima, instead of coinciding with the sun-spot minima, come about two years behind them. On the whole, perhaps, as things now stand, it would be fair to say that there is a small balance of probability in favor of the statement that years of sun-spot maxi- mum are a degree or so cooler than those of spot-mini- mum ; but the balance is very slight indeed, and the next investigation of somebody else may carry it to the other side. As regards the influence of sun-spots upon storms and rainfall, the evidence, if not entirely conclusive, as it is considered by Mr. Lockyer and some other high authorities, is at least considerably stronger. In 1872 Mr. Meldrnm, director of the observatory at the Mau- ritius, published a comparison between the number of cyclones observed in the Indian Ocean and the state of the sun, and pointed out that the number of cyclones was greatest at the time of a sun-spot maximum. We quote his words (" Nature," vol. vi, p. 358) : " Taking the maxima and minima epochs of the sun-spot period, and one year on each side of them, and comparing the number of cyclones in these three-year periods, we get the following results : 174: THE SUN. TEARS. No. of cyclones in each year. Total No. of cyclones. Maxima Minima.... !1847 .... i I 1 ! 15 8 21 9 14" 1848 1849 !1855 1856 1857 1859 1860 1861 1866 1867 Maxima . 1868 1870. . 1871 1872 Subsequently Mr. Meldrum made more extensive comparisons, including not only cyclones proper, but other great storms, and brings out essentially the same results. At the same time it is to be noted that the yearly numbers vary enormously, and, on referring to his second paper (" Nature," vol. viii, p. 495), it will be found that the number for the sun-spot maximum, 1847-'49, is only twenty-three, while that for the mini- mum, 1866-'68, is twenty-one. (Mr. Meldrum coaxes the first sun-spot maximum a little by using the years 1848-'50 in his comparison ; rather unwarrantably, it would seem, since the epoch of spot maximum was 1848'! : by using those years, he gets twenty-six instead of twenty-three.) The variations from year to year are so extreme that it is sufficient to say that the observations can hardly be considered as demonstrative without much further con- firmation from other sources. Mr. Meldrum has attempted to supply this confirma- tion by tabulating the rainfall at a number of stations METEOROLOGY AND SUN-SPOTS. 175 in and near the Indian Ocean, and obtains a result con- firmatory on the whole, though there are several discrep- ancies. Mr. Lockyer, from observations of the rainfall at the Cape of Good Hope and Madras, gets corrobora- tive figures. Mr. Meld rum, in a still later paper published in the " Monthly Notices of the Mauritius Meteorological So- ciety," for December, 1878, discusses at length the rainfall of more than fifty different stations in all parts of the earth, and also the levels of many of the princi- pal European rivers. The discussion covers nearly all the available data from 1824 to 1867. It is only just to Mr. Meldrum to say that the treatment seems to be sufficiently thorough, perfectly fair, and the result on the whole is in favor of his opinion that there is a real connection between the annual rainfall and the state of the solar surface. He finds the average rainfall for the earth to be about 38'5 inches annually ; the range be- tween the maximum and minimum is about four inches ; and the rainfall maximum occurs about a year after the sun-spot maximum, though with a good deal of varia- tion at different stations. In some countries, indeed, and at some times (in the United States, for instance, between 1834 and 1843), the results conflict with the theory, but the general accordance is striking, and seems to warrant his concluding statement that " the mean rainfalls of Great Britain, the Continent of Europe, America, and India, as represented by all the returns that have been received, have, notwithstanding anom- alies, varied directly as Wolfs sun-spot numbers have varied, and the epochs of maximum and minimum rain have nearly coincided with those of the sun-spots. The rainfalls at five stations in the southern hemisphere, for shorter periods, give similar results." 176 THE SUN. Mr. Symons, from the British rainfall of the past one hundred and forty years, gets an equivocal result. American stations, so far as they have been tested, are on the whole rather in opposition to those of the Indian Ocean, indicating somewhat less rain than usual during a sun-spot maximum. But, as any one can see by con- sulting Mr. Symons's paper in "Nature," vol. vii, pp. 143-145, in which he has tabulated an immense number of rainfall statistics, the evidence is extremely conflict- ing altogether different in force and character from that which demonstrates the magnetic influence of solar disturbances. Still other attempts have been made to establish a connection between sun-spots and various terrestrial phenomena. Thus, Dr. T. Moft'at, in 1874, published results tending to show that in sun-spot years the aver- age quantity of atmospheric ozone is somewhat greater than during a spot-minimum. Another eminent physician, whose name escapes us, endeavored, some years ago, to show that the visitations of Asiatic cholera are periodical, and that their period depends upon that of the sun-spots, being just once and a half as long about fifteen years. This periodicity may be real, perhaps / but, if so, the fact that the chol- era maxima are alternately synchronous with the max- ima and minima of the spots, would be sufficient to disprove the idea of any casual connection between the phenomena. One of the most interesting of the essays in this direction, is that of Professor Jevons, who sought to show a relation between sun-spots and commercial crises. The idea is by no means absurd, as some have declared it is a mere question of fact. If sun-spots have really any sensible effect upon terrestrial meteor-. INFLUENCE OF SUX-SPOTS. 177 ology, upon temperature, storms, and rainfall, they must thus indirectly affect the crops, and so disturb financial relations ; in such a delicate organization as that of the world's commerce, it needs but a feather-weight, rightly applied, to alter the course of trade and credit, and pro- duce a "boom" (if we may be forgiven the use of so convenient a word) or a crash. We have not time or space to discuss ' Mr. Jevons's paper, but must content ourselves with saying that, to us at least, the facts do not seem fairly to warrant his conclusion. It can do no harm to reiterate arid emphasize what was said a few pages back, that the question of sun-spot influence can not be considered settled ; and that the only method of deciding it is by a continuous series of careful observations, conducted specially for the pur- pose, or at least conducted with reference to the con- ditions of the problem, since the same observations would also be useful as data for various other investi- gations. While it is not at all unlikely that investigation will result in establishing some real influence of sun- spots upon our terrestrial meteorology and determin- ing its laws, it is practically certain that this influence is extremely slight, and so masked and veiled by other influences more powerful . that it is extremely diffi- cult to bring to light. SUN-SPOT THEORIES. Naturally, the remarkable phenomena of the sun- spots have invited speculation as to their cause. As has been mentioned already, some of the early observers believed the spots to be planetary bodies cir- culating around the sun, very near its surface. This 13 178 THE SUN. opinion Galileo unanswerably refuted by pointing out that in that case the spot, in its movement around the sun, ought to be visible less than half the time. He, on the other hand, proposed the theory that they are clouds, floating in the solar atmosphere. This view, in one form or another, has since been held by many astronomers of great authority. Derham believed these clouds to be eruptions from solar volca- noes, and in our own times Capocci has adopted and maintained the same theory. Peters seems to have con- sidered it favorably in 1846, at least so far as the vol- canic part of the hypothesis is concerned, while Kirch- hofl: seems to have assented to Galileo's original opinion unmodified. If the statement be interpreted to mean that sun-spots are masses of cloudy matter, less luminous than the photosphere, and floating in, not above, the photosphere, probably a very large proportion of the students of solar physics would to-day agree to it. Gal- ileo, however, believed the spot-clouds to be high above the shining surface, which we now know not to be the fact; for the observations of Wilson, in 1769, men- tioned a few pages back, and the whole body of obser- vations since then, have made it almost certain that the umbra of a sun-spot lies several hundred miles below the level of the photosphere.* Lalande, however, was not disposed to accept Wil- son's doctrine, and maintained that the sun-spots are the tops of solar mountains projecting above the lumi- nous surface islands in the ocean of fire. In this hy- pothesis the penumbra is accounted for by the shelving sides of the mountains seen through the serni-trans- parent flame. It will be noticed that the theories * But we must not overlook Mr. Hewlett's conclusions (p. 129), nor the observations of E. Wilson and Frost (note to p. 170). SUN-SPOT THEORIES. 179 already mentioned, as well as that of Sir William Herschel, which we must now present, all proceed upon the assumption that the central core of the sun is solid. About the beginning of the present century, Sir William Herschel, after a careful study of the facts, but much influenced by the belief that the sun must (for theological reasons) be a habitable body, proposed an hypothesis which stood unchallenged for nearly half a century. FIG. 48. PHOTO-SPHERE:. PENUMBRAL CLOUD. BODY OF SUN. W. HEKSCHEL'S SUN-SPOT THEOET. He supposed the central portion of the sun to be solid ; its surface cool, non-luminous, and habitable. Around this he placed two envelopes of cloud the outer one, the photosphere, incandescent, blazing with unimaginable fury ; the inner one non-luminous, dark itself, but capable of reflecting light from its upper sur- face, and acting as a screen to protect the underlying country from the heat of the photosphere. The spots he supposed to be caused by temporary openings in the clouds, through which we could look down upon the dark surface of the central globe ; the penumbra being caused by the intermediate cloud-layer, opening less widely than the photosphere. The figure illustrates this theory. As to the cause of the openings he uttered no decided opinion, though suggesting that they might be due to volcanic eruptions, forcing their way Up through the higher atmosphere. His son, Sir John Herschel, many years later, pro- 180 THE SUN. posed an explanation which would make the spots to be great whirling storms boring down through the photo- sphere and clouds, instead of eruptions pushing their way outward. According to him, the rotation of the sun causes an accumulation of the solar atmosphere at the sun's equator a thickening of the layer which ob- structs the radiation of heat. This being so, there should be on the sun, as on the earth, though for an entirely different reason, a temperature higher in the equatorial regions than elsewhere ; and then would follow a long train of consequences, among them these : the solar at- mosphere would be disturbed by currents like the trade- winds on the earth ; there would be stormy zones on each side the equator, and these storms would furnish an explanation of the spots. To a certain extent, the cause adduced must actually exist. The sun's rotation must necessarily thicken the atmospheric layer which overlies the photosphere (i. e., it must, if the surfaces of the photosphere and chromo- sphere can be regarded as level surfaces), and this cause must tend to raise the actual temperature of the sun's equator, while at the same time it must diminish its radiation to the earth, and so render the solar equator apparently cooler, as tested by our observations from the earth. But, so far as can be judged, this effect is quite insensible, as it should be, since the sun's rotation is so slow ; and the motions of the spots show no such systematic drift north or south as solar trade-winds would necessarily produce. The elder Herschel's theory satisfies all the tele- scopic appearances of sun-spots quite as well, perhaps, as any yet proposed. It breaks down in its assumption that the principal portion of the sun is a solid mass, an assumption which is now almost universally regarded SUX-SPOT THEORIES. 181 as incompatible with what we know of the solar tem- perature, radiation, and constitution. It seems to modern physicists an unavoidable con- clusion that the sun's central mass must be gaseous, or at least not solid. Setting out with this idea, Faye and Secchi independently, about 1868, proposed the theory that the spots are openings in the photosphere, through SECCHI'S FIKST SPOT-THBOET. which the internal gases are bursting outward. "We present one of Secchi's figures illustrating this view. But it was abandoned by its proposers as soon as it was clearly pointed out that in that case the spectrum of the umbra of a sun-spot should be composed of bright lines ; and Secchi himself and others had shown that it is not so at all, but a spectrum due to increased absorption, 182 THE SUN. and probably indicating, not an up-rush of heated gases through the photosphere, but a descent of cooler and less luminous matter from above. In this connec- tion the observations of the writer and of Duner may be referred to (page 132). But the theory has great vitality. Mr. Proctor, in his " Old and New Astron- omy," maintained it, and it continually turns up in the speculations of popular writers. About 1870 Zoll- ner proposed a peculiar theory which has many good points about it, but seems obnoxious to fatal objec- tions, and has found very few defenders. He conceives the surface of the sun to be liquid & molten mass over- laid by an atmosphere of vapor. This liquid surface he imagines to be here and there covered at times by slag- like masses of much lower radiating power, the result of local cooling. Around their edges the solar flames burst out with redoubled fury, but at the center the cooler mass of scoria determines a downward current, so as to establish a powerful circulation in the solar at- mosphere downward at the center of the spot, outward in all directions at the surface of the slag, upward all around its margin, and inward, toward the center, in the upper air. This theory admirably agrees with the spectroscopic phenomena ; but the hypothesis of a con- tinuous liquid shell, cool enough to permit the forma- tion of scoriae, seems inconsistent with other phenom- ena, which make it impossible to admit so low a tem- perature at so great a depth. At present, opinion, for the most part, seems to be divided between two rival theories proposed by Faye and Secchi. Faye conceives the sun-spots to be the effect of solar storms ; Secchi believes them to be dense clouds of eruption -products settling down into the photo- SUN-SPOT THEORIES. 183 sphere near, but not at, the points where they were ejected. Faye, it will be remembered, supposes the sun's pe- culiar law of rotation to be due to the hypothetical fact that the ascending masses of vapors (which form the photosphere by their condensation) start from a stratum whose depth below the visible surface regularly dimin- ishes from the equator toward the poles. Hence re- sult currents parallel to the equator, and the conse- quence is that, generally speaking, neighboring portions of the photosphere have a relative drift. At the equa- tor and at the poles this drift vanishes, but is most con- siderable in the middle latitudes. Now, it is Faye's theory that, in consequence of this relative drift, eddies are formed, as explained on a preceding page ; these eddies become cyclones or whirls precisely analogous to those seen in water where a rapid current is obstruct- ed by an obstacle. In such a case, as every one knows r tunnel-shaped vortices are formed, down which floating materials and air are carried to considerable depths. Our terrestrial whirlwinds and tornadoes are produced, according to Faye (but in opposition to the generally received theories), in a similar manner, beginning from above, and penetrating downward until the point of the whirling vortex reaches and sweeps the earth. Now, such a vortex, on the solar scale, is the essence of a sun- spot, according to Faye. It is evident at once that this theory gives a reason- able explanation of the distribution of the spots in two parallel zones on each side of the sun's equator, and that the drifting action, in which the cause of the spots is supposed to lie, is a vera causa. The theory accords very well, also, with the phe- nomena which accompany the subdivision of spots, 184 THE SUN. since whirls in water and cyclones m the terrestrial atmosphere behave in precisely the same sort of way. It fairly meets, too, the spectroscopic indications. The cavity filled with descending vapors would naturally give just such a kind of spectrum as that which is ordi- narily observed. Moreover, the gases carried down in the vortex below the photosphere, especially the hydro- gen, would boil up again all around the whirlpool, and thus we could account for the ring of faculse and prom- inences which, as a general rule, environs every spot of considerable magnitude. Some of the more obvious objections can also be easily disposed of. Thus, it has been said that, if the sun-spots are such vortices, they ought to be circular in outline. Faye replies that we see, not the vortex itself, but a great cloud of cooler gases, sucked down from above and gathered into the storm from all sides, and the form of this cloud would depend upon a multitude of circumstances. But there are other objections which are not so easily met. It the theory be true, all spots are whirls and ought to show a vortical motion, and, what is more, all spots north of the equator ought to whirl in the same direction, and against the hands of a watch (as seen from the earth), while those in the sun's southern hemi- sphere should revolve in the contrary direction, pre- cisely as cyclones do in the atmosphere of the earth. Now, this is not the case at all. As we have seen, only a very small percentage of the spots show any trace of vorticose motion ; and, so far from observing any uniformity in the direction of rotation on each side of the equator, we frequently find different members of the same group of spots, or even different portions of the self-same spot, revolving oppositely. In fact, when we come to look into the matter nu- SUN-SPOT THEORIES. , 185 merieally, we find that the drift, which Faye makes the determining factor of sun-spot genesis, is far too slight to produce such effects. It is very easy to compute this drift if we assume the correctness of Faye's own formula for the motion of a point on the sun's surface in any given solar lati- tude, viz., V = 862' - 186' sin 2 X; V in this formula being the number of minutes of solar longitude passed over by any given point in twenty-four hours. If we apply this formula to two points on the solar surface, one in latitude 20 and the other in latitude 20 1', i, e., about 123 miles north of the first, we shall find that the first has a daily motion of 840-242' and the second 840-207', a difference of only -035', or (in this latitude) 4- IT miles. That is to say, if we take two points on the solar surface, on the same meridian, in latitude 20, at a distance of 123 miles, the one nearer the equator will, at the end of twenty-four hours, have drifted about 4 miles to the eastward of the other. If we make the same calculation for latitude 45, we get a result a trifle greater about 4r| miles per day. With these figures it is easy to see why the sun-spots do not behave more like the disturbances of our terres- trial atmosphere, in exhibiting cyclonic motion as a regular and invariable characteristic, instead of an occa- sional and rather a rare phenomenon. Secchi's latest theory is based essentially upon the idea, certainly borne out by observation, that eruptions are continually breaking through the photosphere, and carrying up metallic vapors from the regions beneath. He imagines that these vapors, after becoming consid- erably cooled, descend upon the photosphere and form depressions in it, which are filled with these less lumi- nous and absorbent materials. It is difficult to see why 186 THE SUN - the effect should remain so persistent, or why, even if the eruption be long maintained, the cloud should con- tinue to descend in the same place. In fact, as was said only a few moments ago, a spot is generally sur- rounded by a ring of eruptions, and things take place as if they were all pouring their ejections into the same receptacle as if there were, in fact, some such down- ward suction through the center of the spot as the the- ory of Faye supposes, an aspiration capable of drawing in toward the spot all erupted materials in the vicinity. The sun-spot theories of Lockyer and Schaeberle, already referred to on page 143 in connection with the explanation of the equatorial acceleration of the sun's rotation, agree with this theory of Secchi's in attributing the spots to the downfall of matter from a great ele- vation. Schaeberle supposes it to be matter simply blown out by eruptions, some of it with force enough to carry it out even to the orbits of Jupiter and Saturn ; on its return it penetrates and chills the photosphere. Mr. Lockyer, if we understand him rightly, in his sug- gestions which form the closing chapter of the " Chem- istry of the Sun," is rather disposed to think that the " iron " and such other substances as by their fall pro- duce the spots are formed by the union and combina- tion of their elementary constituents, which have as- cended in a " dissociated " condition to the upper regions of the solar atmosphere. There, where the temperature is no longer above the " dissociation " point, the atoms recombine into molecules of iron-vapor, etc. ; the vapors condense into clouds and liquid masses, and these de- scend upon the photosphere. They absorb heat all the way down, by revaporization and new dissociation chill- ing the photosphere where they pierce it and causing a " splash " or up-rush of the photospheric matter and its SUN-SPOT THEORIES. 1S7 underlying gases all around the spot, wliicli we recog- nize as faculae, prominences and metallic eruptions. According to these theories the faculae and eruptions are consequences of the formation of the spot : accord- ing to Secchi they precede and cause it. It would seem easy to decide the question by observation, but it does not appear to be so ; on the whole, however, the weight of evidence is pretty strongly in favor of the opinion that faculae and pores and a general disturbance of the region are usually obvious before the spot manifests it- self ; and it must be admitted that in some cases the appearances puzzlingly resemble the emergence of a dark tnas&from beneath. Probably both Lockyer and Schaeberle would cheer- fully accept Sir John Herschel's theory to a certain ex- tent that some of the spots may be due to the fall upon the sun of great meteors from outer space. "While it is hardlv possible that, directly, a meteor, such as we know meteors upon the earth, could by its fall produce even a small sun-spot, it is not easy to say what might be the indirect effects consequent upon its passage through the photosphere, and its disturbance of the dynamical equilibrium. The writer some time ago suggested a modification of Secchi's theory, which seems to remove some of the objections, and appears on the whole more probable than any of the others. It may be that the spots are depressions in the photospheric level, caused not directly by the pressure of the erupted materials from above, but by the diminution of upward pressure from below, in consequence of eruptions in the neighborhood ; the spots thus being, so to speak, sinks in the photosphere. Undoubtedly the photosphere is not a strictly continu- ous shell or crust, but it is heavy as compared with the 188 THE SUN. uncondensed vapors in which it lies, just as a rain-cloud in our terrestrial atmosphere is heavier than the air, and it is probably continuous enough to have its upper level affected by any diminution of pressure below. The gaseous mass below the photosphere supports its weight and the weight of the products of condensation, which must always be descending in an inconceivable rain and snow of molten and crystallized material. To all intents and purposes, though nothing but a layer of clouds, the photosphere thus forms a constricting shell, and the gases beneath are imprisoned and compressed. Moreover, at a high temperature the viscosity of gases is vastly increased, so that quite probably the matter of the solar nucleus resembles pitch or tar in its consist- ency more than what we usually think of as a gas. Consequently, any sudden diminution of pressure would propagate itself gradually from the point where it oc- curred. Putting these things together, it would seem that, whenever a free outlet is obtained through the photosphere at any point, thus decreasing the inward pressure, the result would be the sinking of a portion of the photosphere somewhere in the immediate neigh- borhood, to restore the equilibrium ; and, if the erup- tion were kept up for any length of time, the depression in the photosphere would continue till the eruption ceased. This depression, filled with the overlying gases, would constitute a spot. Moreover, the line of frac- ture, if we may call it so, at the edges of the sink would be a region of weakness in the photosphere, so that we should expect a series of eruptions all around the spot. Tor a time the disturbance, therefore, would grow, and the spot would enlarge and deepen, until, in spite of the viscosity of the internal gases, the equilibrium of pressure was gradually restored beneath. So far as we SUN-SPOT THEORIES. 189 know the spectroscopic and visual phenomena, none of them contradict this hypothesis. As regards the limitation of the spots to certain latitudes, this, as has been said already, almost certainly will find its explanation in that of the equatorial accel- eration. Faye, Belopolsky, Lockyer, and Schaeberle, all present such explanations. Schaeberle's discussion of the subject may be found in " Astronomy and Astro- Physics," for April, 1894. We shall have occasion to refer to it again in connection with the corona. Whatever may be the cause of spots, it is probable that the annexed figure gives a fair idea of the arrange- ment and relations of the photosplieric clouds in the CONSTITUTION OF A SUN-SPOT. neighborhood of one. Over the sun's surface gener- ally, these clouds probably have the form of vertical columns, as at a a. Just outside the spot, the level of the photosphere is usually raised into faculae, as at b b. These faculae are for the most part overtopped by eruptions of hydrogen and metallic vapors, as indicated by the shaded clouds. Of these metallic eruptions we shall have more to say in the chapter upon the chromo- sphere and prominences, only remarking here that, 190 THE SUN. while the great clouds of hydrogen are found every- where upon the sun, these spiky, vivid outbursts of metallic vapors seldom occur, except just in the neigh- borhood of a spot, and then only during its season of rapid change. In the penumbra of the spot the photo- spheric filaments become more or less nearly horizontal, as atpp ; in the umbra, at u, it is quite uncertain what the true state of affairs may be. We have conjecturally represented the filaments there as vertical also, but de- pressed and carried down by a descending current. Of course, the cavity o o is filled by the gases which overlie the photosphere ; and it is easy to see that, looked at from above, such a cavity and arrangement of the luminous filaments would present the appearances actu- ally observed. Oppolzer, of Vienna, in 1893 proposed a new theory based largely upon Hann's researches on the tempera^ ture effects of vertical atmospheric currents. Such cur^ rents are supposed to rise periodically from the polar regions of the sun, to drift slowly toward its equator, and to descend in the spot zones, becoming heated and " dried " in their descent, thus forming in the photo- sphere cavities which are filled with metallic vapors in purely gaseous condition. In many ways the theory admirably corresponds with facts, explaining better than any other the peculiar character of the sun-spot spectrum, Spoerer's law of sun-spot latitudes, and the otherwise puzzling observa- tions of Langley and Frost upon sun-spot temperatures (page 1YO, note). But the polar streams themselves are unaccounted for, and it remains to be seen how this "meteorological" theory will withstand other adverse criticisms. CHAPTER VI. THE CHROMOSPHERE AND THE PROMINENCES. Early Observations of Chromosphere and Prominences. The Eclipses of 1842, 1851, and I860. The Eclipse of 1868. Discovery of Janssen and Lockyer. Arrangement of Spectroscope for Observations upon Chromosphere. Spectrum of Chromosphere. Lines always present. Lines often reversed. Ultra-Violet Studies of Hale and Des- landres. Motion Forms. Double Reversal of Lines. Distribution of Prominences. Magnitude. Classification of Prominences as qui- escent, and eruptive or metallic. Isolated Clouds. Violence of Mo- tion. Observations of August 5, 1872. Theories as to the Forma- tion and Causes of the Prominences. WHAT we see of the sun under ordinary circum- stances is but a fraction of his total bulk. While by far the greater portion of the solar mass is included within the photosphere the blazing cloud-layer, which seems to form the sun's true surface, and is the princi- pal source of his light and heat yet the larger portion of his volume lies without, and constitutes an atmos- phere whose diameter is at least double, and its bulk therefore sevenfold that of the central globe. Atmosphere, however, is hardly the proper term ; for this outer envelope, though gaseous in the main, is not spherical, but has an outline exceedingly irregular and variable. It seems to be made up not of overlying strata of different density, but rather of flames, beams, and streamers, as transient and unstable as those of our own aurora borealis. It is divided into two portions, separated by a boundary as definite, though not so 192 THE SUN. regular, as that which parts them both from the photo- sphere. The outer and far more extensive portion, which in texture and rarity seems to resemble the tails of comets, and may almost, without exaggeration, be likened to " the stuff that dreams are made of," is known as the " coronal atmosphere," since to it is chiefly due the " corona " or glory which surrounds the darkened sun during an eclipse, and constitutes the most impressive feature of the occasion. At its base, and in contact with the photosphere, is what resembles a sheet of scarlet fire. The appearance, which probably indicates a fact, is as if countless jets of heated gas were issuing through vents and spiracles over the whole surface, thus clothing it with flame which heaves and tosses like the blaze of a conflagra- tion "like a prairie on fire," to quote the vividly de- scriptive phrase of Professor Langley. This has received the name of chromosphere, a designation first proposed by Frankland and Lockyer in 1869, and signifying " color-sphere," in allusion to the vivid redness of the stratum, caused by the pre- dominance of hydrogen in these flames and clouds. It was called the " sierra " by Airy in 1842, and Proctor and some other writers prefer that name to the later and more common appellation. Here and there masses of this hydrogen mixed with other substances rise to a great height, ascending far above the general level into the coronal regions, where they float like clouds, or are torn to pieces by contend- ing currents. These cloud-masses are known as solar " prominences," or " protuberances," a non-committal sort of appellation applied in 1842, w r hen they first attracted any considerable attention, and while it was a warmly-disputed question whether they were solar, THE CHROMOSPHERE AND THE PROMINENCES. 193 lunar, phenomena of our own atmosphere, or even mere optical illusions. Jt is unfortunate that no more appro- priate and graphic name has yet been found for objects of such wonderful beauty and interest. Until recently, the solar atmosphere could be seen only at an eclipse, when the sun itself is hidden by the moon. Now, however, the spectroscope has brought the chromosphere and the prominences within the range of daily observation, so that they can be studied with nearly the same facility as the spots and faculse, and a fresh field of great interest and importance is thus opened to science. It seems hardly possible that the ancients should have failed to notice, even with the naked eye, in some one of the many eclipses on record, the presence of blazing, star-like objects around the edge of the moon, but we find no mention of any thing of the kind, al- though the corona is described as we see it now. On this ground some have surmised that the sun has really undergone a change in modern times, and that the chromosphere and prominences are a new development in the solar history. But such mere negative evidence is altogether insufficient as a foundation for so impor- tant a conclusion. The earliest recorded observation of the prominences is probably that of Yassenius, a Swedish astronomer, who, during the total eclipse of 1733, noticed three or four small pinkish clouds, entirely . detached from the limb of the moon, and, as he supposed, floating in the lunar atmosphere. At that time this was the most natural interpretation of the appearance, since the fact that the moon has no atmosphere was not yet ascer- tained. The Spanish admiral, Don Ulloa, in his account of 14 194 THE SUN. the eclipse of 1 778, describes a point of red light which made its appearance on the western limb of the moon about a minute and a quarter before the emergence of the sun. At first small and faint, it grew brighter and brighter until extinguished by the returning sunlight. He supposed that the phenomenon was caused by a hole or fissure in the body of the moon ; but, with our present knowledge, there can be little doubt that it was simply a prominence gradually uncovered by her motion. The chromosphere seems to have been seen even earlier than the prominences : thus Captain Stannyan, in a report on the eclipse of 1706, observed by him at Berne, noticed that the emersion of the sun was pre- ceded by a blood-red streak of light, visible for six or seven seconds upon the western limb. Halley and Louville saw the same thing in 1715. Halley says that two or three seconds before the emersion a long and very narrow streak of a dusky but strong red light seemed to color the dark edge of the moon on the western edge where the sun was about to reappear. Louville's account agrees substantially with this, and he further describes the precautions he used to satisfy himself that the phenomenon was no mere optical illu- sion, nor due to any imperfection of his telescope. In eclipses that followed that of 1733, the chromo- sphere and prominences seem to have attracted but lit- tle attention, even if they were observed at all. Some- thing of the sort appears to have been noticed by Ferrers in 1806, but the main interest of his observation lay in a different direction. In July, 1842, a great eclipse occurred, and the shadow of the moon described a wide belt running across southern France, northern Italy, and a portion of Austria. The eclipse was carefully observed by THE CHROMOSPHERE AND THE PROMINENCES. 195 many of the most noted astronomers of the world, and so completely had previous observations of the kind been forgotten, that the prominences, which appeared then with great brilliance, were regarded with extreme surprise, and became objects of warm discussion, not only as to their cause and location, but even as to their very existence. Some thought them mountains upon the sun, some that they were solar flames, and others, clouds floating in the sun's atmosphere. Others re- ferred them to the moon, and yet others claimed that they were mere optical illusions. At the eclipse of 1851 (in Sweden and Norway), similar observations were repeated, and, as a result of the discussions and comparison of observations which followed, astronomers generally became satisfied that the prominences are real phenomena of the solar atmosphere, in many respects analogous to our terrestrial clouds ; and several came more or less confidently to the conclusion, now known to be true (see Grant's " History of Physical Astrono- my "), that the sun is entirely surrounded with a con- tinuous stratum of the same substance. Many, how- ever, remained unconvinced : Faye, for instance, still asserted them to be mere optical illusions, or mirages. In the eclipse of 1860, photography was for the first time employed on such an occasion with anything like success. The results of Secchi and De La Rue removed all remaining doubts as to the real existence and solar character of the objects in question, by exhibiting them upon their plates gradually covered on one side and un- covered on the other side of the sun by the progress of the moon. Secchi thus sums up his conclusions, which have been justified in almost all their details by later obser- vations ; they require few and slight corrections : 196 THE SUN, " 1. The prominences are not mere optical illusions ; they are real phenomena pertaining to the sun. . . . 2. The prominences are collections of luminous mat- ter of great brilliance, and possessing remarkable pho- tographic activity. This activity is so great that many of them, which are visible in our photographs, could not be seen directly even with good instruments. 3. Some protuberances float entirely free in the so- lar atmosphere like clouds. If they are variable in form, their changes are so gradual as to be insensible in the space of ten minutes. (Generally, but by no means al- ways, true.) 4. Besides the isolated and conspicuous protuber- ances there is also a layer of the same luminous sub- stance which surrounds the whole sun, and out of which the protuberances rise above the general level of the so- lar surface. . . . 5. The number of the protuberances is indefinitely great. In direct observation through the telescope the sun appeared surrounded with flames too numerous to count. . . . 6. The height of the protuberances is very great, especially when we take account of the portion hidden by the moon. One of them had a height of at least three minutes, which indicates a real altitude of more than ten times the earth's diameter. . . ." But. their nature still remained a mystery; and no one could well be blamed for thinking it must always remain so to some degree. At that time it could hard- ly be hoped that we should ever be able to ascertain their chemical constitution, and measure the velocities of their motions. And yet this has been done. Before the great Indian eclipse of August 18, 1868, the spec- troscope had been invented (it was, indeed, already in THE CHROMOSPHERE AND THE PROMINENCES. 197 its infancy in 1860), and applied to astronomical research with the most astonishing and important results. 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. They found the spectrum of the prominences observed 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, completely demonstrated the fact that the prominences are enor- mous 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 prevented 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 conclu- sion, that hydrogen is the most conspicuous component of the prominences, but found that the yellow line must 198 THE SUN. be referred to some other element than sodium, being somewhat more refrangible than the D lines. He found also that, by slightly moving his telescope 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 sta- tion 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 accounts 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 inde- pendently, 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 occurred 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 sufficient power, even in broad daylight. The principle is simply this : Under ordinary circumstances the protuberances are invisible, for the same reason as the stars in the day- time : they are hidden by the intense light reflected from the particles of our own atmosphere near the sun's place in the sky, and, if we could only sufficiently weaken this aerial illumination, without at the same time weakening their light, the end would be gained. THE CHROMOSPHERE AND THE PROMINENCES. 199 And the spectroscope accomplishes precisely this very thing. Since the air-light is reflected sunshine, it of course presents the same spectrum as sunlight, a con- tinuous band of color crossed by dark lines. K"ow, this sort of spectrum is greatly weakened by every increase of dispersive power, because 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 dis- persive power of the spectroscope. The bright lines are only more widely separated not in the least dif- fused 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 dispersion, 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 telescope, be examined with a spectroscope, one might 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 experi- ment as early as 1867, but without success ; partly be- cause 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 existence. At any rate, as soon as the discovery was announced, 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 per- ceive a thing known to exist does not require one half 200 THE SUN. the instrumental power or acuteness of sense as to dis- cover 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 under- taken its construction, and partly due to other causes, he received the new spectroscope just as the report of Herschel's and Tennant's observations reached England. Hastily adjusting the instrument, not yet entirely com- pleted, he at once applied it to his telescope, and with- out difficulty found the lines, and verified their position. He immediately also discovered them to be visible around the whole circumference of the sun, and conse- quently that the protuberances are mere extensions of a continuous 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, Jans- sen, Lockyer, Zollner, and others, that by giving a rapid motion of vibration or rotation to the slit of the spec- troscope it would be possible to perceive the whole con- tour and detail of a protuberance at once, but it seems THE CHROMOSPHERE AND THE PROMINENCES. 201 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 suf- ficient 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 protuber- ance becomes visible, and, if not too large, the whole can be seen at once : with the widening of the slit, how- ever, the brightness of the background increases, so that the finer details of the object are less clearly seen, and a limit is soon reached beyond which further widen- ing is disadvantageous. The higher the dispersive pow- er of the spectroscope the wider the slit that can be used, and the larger the protuberance that can be exam- ined as a whole within certain limits, however. It is not difficult with our latest spectroscopes, diffraction instruments especially, to reach a dispersion so great that even the C line becomes broad and hazy, like the OP A PROMINENCE IN FULL SUNSHINE. b lines in an ordinary instrument. In that case each luminous point in the prominence itself is represented in the image of the prominence, net by a point, as it should be to give clear definition, but by a streak at right angles to the spectrum lines. Mr. Huggins's first successful observation of the 202 THE SUN. form of a solar protuberance was made on February 13, 1869. Fig. 46, copied from the " Proceedings of the Royal Society," presents his delineation of what he saw. As his instrument had only the dispersive power of two prisms, and included in its field of view a large portion of the spectrum at once, he found it necessary to sup- plement its powers by using a red glass to cut off stray FIG. 47. SPECTROSCOPE, WITH TRAIN OP PRISMS. light of other colors, and by inserting a diaphragm at the focus of the small telescope of the spectroscope to limit the field of view to the portion of the spectrum immediately adjoining the C line. With the instru- ments now in use, these precautions are seldom neces- sary. It may be noticed, in passing, that Mr. Huggins had previously (and has subsequently) made many experi- ments with different absorbing media, in hopes of find- ing some substance which, by cutting off all light of other color than that emitted by the prominences, should render them visible in the telescope ; thus far, however, without success. The spectroscopes used by different astronomers for THE CHROMOSPHERE AND THE PROMINENCES. 203 observations of this sort differ greatly in form and power. Fig. 47 represents the one long used at the Shattuck Observatory of Dartmouth College, and sev- eral of our American observatories are supplied with instruments similarly arranged. The light passes from the collimator c, through the train of prisms p, near their bases, and, by two reflections in a rectangular prism, r, is transferred to the upper story, so to speak, of the prism-train, and made to return to the telescope t, finally reaching the eye at e. It thus twice traverses a train of six prisms, and the dispersive power of the instrument is twelve times as great as it would be with only one prism. The diameter of the collimator is a little less than an inch, and its length ten inches. The whole instrument only weighs about fourteen pounds, and occupies a space of about 15 in. X 6 in. X 5 in. It is also automatic, i. e., the tangent screw m keeps the train of prisms adjusted to their position of min- imum deviation by the same movement which brings the different portions of the spectrum to the center of the field of view, and the milled head f focuses both the collimator and telescope simultaneously. The spectroscope is attached to the equatorial tele- scope, to which it belongs, by means of the clamping rings a, a. These slide upon a stout metal rod, firmly fastened to the telescope in such a way that the slit 0 73 as the rise of tern- 288 THE SUN. perature produced by a sunbeam three inches in diam- eter, absorbed by a mass of matter equivalent to 4,638 grains of water (we do not indicate the minutiae of the process by which the weight of the tin vessel, ther- mometer, stirrer, etc., are allowed for). Nothing more is now necessary to enable us to compute just how much heat is received by the earth in a day or a year, except, indeed, the determination of the very troublesome and somewhat uncertain correction for the absorption of heat by the earth's atmosphere a correction deduced by means of observations made at varying heights of the sun above the horizon. Herschel preferred to express his results in terms of melting ice, and put it in this way : the amount of heat received on the earth's surface, with the sun in the zenith, would melt an inch thickness of ice in two hours and thirteen minutes nearly. Since there is every reason to believe that the sun's radiation is equal in all directions, it follows that, if the sun were surrounded by a great shell of ice, one inch thick and a hundred and eighty-six million miles in diameter, its rays would just melt the whole in the same time. If, now, we suppose this shell to shrink in diameter, retaining, however, the same quantity of ice by increasing its thickness, it would still be melted in the same time. Let the shrinkage continue until the inner surface touches the photosphere, and, allowing for atmospheric absorption, the ice -envelope would become more than a mile thick, through which the solar fire would still thaw out its way in the same time at the rate, according to HerschePs determinations, of more than forty feet a minute. Herschel continues that, if this ice were formed into a rod 45-3 miles in diameter, and darted toward the sun with the velocity THE SUN'S LIGHT AND HEAT. 289 of light, its advancing point would be melted off as fast as it approached, if by any means the whole of the solar rays could be concentrated on the head. Or, to put it differently, if we could build up a solid column of ice from the earth to the sun, nearly two miles and a half in diameter, spanning the inconceivable abyss of ninety- three million miles, and if then the sun should concen- trate his power upon it, it would dissolve and meltj not in an hour, nor a minute, but in a single second : one swing of the pendulum, and it would be water ; seven more, and it would be dissipated in vapor. In formulating this last statement we have, however, employed, not Herschel's figures, but those resulting from later observations, which increase the solar radia- tion almost fifty per cent., making the thickness of the ice-crust which the sun would melt off of his own sur- face in a minute to be much nearer sixty feet than forty. To put it a little more technically, expressing it in terms of the modern scientific units, the sun's radiation amounts to more than 1,200,000 calories per minute for each square metre of his surface, the calory* or heat- unit, being the quantity of heat which will raise the temperature of a kilogramme of water one degree cen- tigrade. An easy calculation shows that, to produce this amount of heat by combustion would require the hourly burning of a layer of anthracite coal more than nineteen feet (five metres) thick over the entire surface of the sun nine tenths of a ton per hour on each square foot of surface at least nine times as much as the consump- * This is the engineers' " calory." For many scientific purposes the " small calory," a thousand times less, is more conveniently used viz., the amount of heat which will raise the temperature of one gramme of water one centigrade degree. 20 290 THE SUN. tion of the most powerful blast-furnace known to art. It is equivalent to a continuous evolution of about twelve thousand horse-power on every square foot of the sun's whole area. As Sir William Thomson (now Lord Kel- vin) has shown, the sun, if it were composed of solid coal, and produced its heat by combustion, would burn out in less than five thousand years. Of this enormous outflow of heat the earth of course intercepts only a small portion, about ^,2Tnr.in5,innF- But even this minute fraction is enough to melt yearly, at the earth's equator, a layer of ice more than one hun- dred and thirty-two feet thick. If we choose to express it in terms of " power," we find that this is equivalent, for each square foot of surface, to more than seventy- two tons raised to the height of a mile ; and, taking the w r hole surface of the earth, the average energy re- ceived from the sun is over sixty mile-tons yearly, or one horse-power continuously acting, to every twenty- five square feet of the earth's surface. Most of this, of course, is expended merely in maintaining the earth's temperature ; but a small portion, perhaps -j-^ of the whole, as estimated by Helmholtz, is stored away by animals and vegetables, and constitutes an abundant revenue of power for the whole human race.* * Several experimenters have contrived machines for the purpose of utilizing the solar heat as a source of mechanical energy, among whom Ericsson and Mouchot have been most successful. M. Pifre describes some results from a machine of Mouchot's construction, claiming to have utilized more than seventy per cent, of the heat which falls on the mirrors of the instrument something over twelve calories to a square metre. We do not mean, of course, that this percentage of the total solar energy ap- peared as mechanical power in the engine, but only in its boiler. The machine had a mirror-surface of nearly a hundred square feet, and gave not quite a horse-power. Ericsson's engine, exhibited for several years in the American Institute Fairs in New York, about 1886, was still more efficient and powerful, driving a two and a half horse-power engine vigor- THE SUN'S LIGHT AND HEAT. 291 If we inquire what becomes of that principal portion of the solar heat which misses the planets and passes off into space, no certain answer can be given. Remem- bering, however, that space is full of isolated particles of matter (which we encounter from time to time as shooting- stars), we can see that nearer or more remotely in its course each solar ray is sure to reach a rest- ing-place. It has been suggested that the sun sends heat only to- ward its planets ; that the action of radiant heat, like that of gravi- tation, is only between masses. But scientific investigation so far fails to prove it. The energy radiated from a heated globe is found to be alike in all direc- tions, and wholly independent of the bodies which receive it, nor is there the slightest reason to suppose the sun any way differ- ent in this respect from every other incandescent mass. Pouillet's experiments were made about the same time as Herschel's, but with a dif- ferent apparatus, though based on the same principles. He named his instrument the pyrheliometer, or " meas- urer of solar fire." Fig. 95 represents it. The little snuffbox-like vessel, a, b, of silver-plated copper, black- ened on the upper surface, contains a weighed quantity ously. It is quite likely that such machines will prove practically useful in countries where sunshine can be depended on at certain seasons, as in Egypt and California. 292 THE SUN. of water, and a thermometer is immersed in it, the mer- cury in its stem being visible at d. The disk, e, e, makes it easy to point the instrument squarely to the sun, by directing it so that the shadow of a falls concentrically upon this disk. The button at the lower end is for the purpose of agitating the water in the vessel a, a, by sim- ply turning the whole thing on its axis, in the collar