UC-NRLF B M 5ED fihfl ASTRONOMY TEN YEARS' WORK OF A MOUNTAIN O BS E R V ATO R Y Mount San Antonio from Mount Wilson. TEN YEARS' WORK OF A MOUNTAIN OBSERVATORY A brief account of the Mount Wilson Solar Observatory of the Carnegie Institution of Washington BY GEORGE ELLERY HALE WASHINGTON, D. C. PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON 1915 CARNEGIE INSTITUTION OF WASHINGTON PUBLICATION No. 235 NOTE. The eleven departments of research of the Carnegie Insti- tution of Washington, of which the Mount Wilson Solar Observatory is one, are located in various places selected because of their suitability for the several purposes in view. Information regarding the work of the Institution may be obtained on application to the Office of Administration, Washington D. C. ASTRONOMY LIBRARY GIBSON BROTHERS, PRINTERS WASHINGTON, D. C. ASTRO NO to LIBRARY FIG. i. Mount Wilson from Pasadena TEN YEARS' WORK OF A MOUNTAIN OBSERVATORY Ten years have elapsed since the inception of the Mount Wilson Observatory, and the time is opportune for a general survey of the work of the past and the possibilities of the future. An exten- sive equipment of buildings and instruments, involving heavy 'initial 'cost and large annual ex- penditure for operation and maintenance, has been provided on a 6,ooo-foot mountain summit and in the valley near its base. Activities of many kinds, including the preparation of new plans of research, the invention, design, and construction of instruments, the building of a mountain road, (3) M776894 4 the transportation of many hundreds of tons of materials, the erection of brick, steel, and concrete structures, the execution of an observational pro- gram, the measurement and reduction of thou- sands of photographs, and the imitation of celes- tial phenomena by laboratory experiments, have taxed the capacity of a large staff of workers. The equipment is now so nearly complete and the plan of investigation so definitely outlined that a brief description of some typical methods and results may be of service to the visitor and to the general reader interested in the progress of astrophysical research. OLD METHODS AND NEW. An observatory, like any other laboratory of research, may concentrate its attention upon either one of two widely different objects: the accumu- lation of great stores of data in existing depart- ments of knowledge, or the opening up and explo- ration of new fields of investigation. In both cases extensive series of routine observations are re- quired, but the point of view and the mode of attack are essentially different. In known fields, long since efficiently occupied, standard methods of observation and instruments obtainable from skilled makers are available for use. With the aid of such instruments and methods, perhaps modi- fied and perfected in various details, observations of great precision and importance can be obtained. Moreover, by the preparation of a suitable scheme, well exemplified in Kapteyn's work on the struc- 5 ture of the universe, these observations can be made to serve, not merely for the tabulation of accurate data, but for the solution of the greatest problems of astronomy. Results of the highest importance are therefore within the reach of the investigator equipped with standard instruments. His studies may develop new points of view or new data which will lead into new fields of research, but his position and needs are very different from those of the man whose re- searches force him to leave the familiar path. Discarding, perhaps, the instruments which have proved their strength and weakness by many years of use, he replaces them with others possessing new advantages and defects. In departing from ac- cepted standards and in preparing to overcome difficulties, the initiator of new methods almost necessarily becomes an instrument-maker, and hence a machine-shop may be his first requirement. He can not afford to intrust construction to instru- ment-makers thousands of miles away, with whom he is unable to discuss details of the design, neces- sarily subject to frequent modification in the light of newly acquired ideas. To be most efficient, he must be his own designer and builder, ready to take immediate advantage of those new points of view and new possibilities of attack which his investigations are certain to disclose. A laboratory or observatory like that of Mount Wilson, planned for the exploration of unfamiliar fields, can thus possess no fixed and standard equipment. Its mode of attack and its means of progress must grow with its work and develop with the disclosure of new and unexpected possi- bilities. ADVANTAGES OF A MOUNTAIN SITE. It was Newton who first pointed out the impor- tance of making astronomical observations from a mountain top: "For the Air through which we look upon the Stars, is in a perpetual Tremor; as may be seen by the tremulous Motion of Shadows cast from high Towers, and by the twinkling of the fix'd stars. . . . The only remedy is a most serene and quiet Air, such as may perhaps be found on the tops of the highest Mountains above the grosser Clouds." (Opticks, third edition, p. 98.) But height is not the only essential; indeed, very great altitudes are to be avoided. The summit of the Rocky Mountains is a notoriously bad place for astronomical work, because of the unstable atmospheric conditions and the frequent storms. Long periods of unbroken weather, free from rain and with little cloud, are associated with that tranquillity and steadiness of the atmosphere which Newton so much desired. These are to be found on the mountains of the Sierra Madre range, in the semi-tropical climate of Southern California. Mount Wilson, selected by Hussey after many tests of other elevated points in the northern and southern hemispheres, was accordingly chosen as the observation station. Rising abruptly from the San Gabriel Valley to a height of nearly 6,000 feet, and lying some 30 miles from the Pacific Ocean, it offered more advantages than any other site known. One of these was the possibility of establishing the shops, laboratories, and offices in the city of Pasadena, within easy reach of large foundries, supply houses, sources of electric light and power, and other facilities demanded by the nature of the work. Throughout the dry season, FIG. 2. Mount Wilson from Mount Harvard. day after day is clear and tranquil and the wind velocity is remarkably low. The broad and diver- sified mountain summit, protected from the sun's heat by spruces, pines, and undergrowth, affords numerous locations perfectly adapted for the vari- ous instruments; the water-supply is abundant, and the completion of the mountain road makes the distance to the Pasadena office only 16 miles, covered in i\ hours (ascending) by auto-stage. LIFE HISTORY OF THE STARS. But the scheme of research is the prime con- sideration. Let us suppose, as in the case of the Mount Wilson Observatory, that the chief object is to contribute, in the highest degree possible, to the solution of the problem of stellar evolution. What was the origin of this earth on which we live? We know that it is a member of a solar system, one of several planets moving harmoniously about a great central sun, on which they depend for light FIG. 3. Saturn. and heat. But how was the earth formed? Through what successive stages did it pass in its early life? How were its constituent parts sepa- rated from that great vaporous mass which, as most astronomers believe, once united the planets and the sun? By what process, extending over millions of years, have the intensely hot solar gases condensed toward a center, leaving behind those rotating and revolving spheres, the planets and their satellites ? Or must this Laplacian view give way to the radically different conceptions of the planetesimal hypothesis? What is the nature of the central sun, on which our lives depend ? What is its relationship to other stars, and what part does it play in the universe ? How is this universe organized, what bodies does it comprise, what is its structure, and what are its bounds? If the processes of creation and evolution are still at work within it, may we not expect to find existing exam- FIG. 4. Nebula N. G. C. 7217. pies of the various stages through which our own solar system has passed? And may we not hope, by learning the relationship of stars in moving groups, and by tracing these groups back to their former positions in space, to reconstruct a picture of the universe as it was long before the solar sys- tem had taken form? A large project, it may be said, too ambitious for the ideal of any institution. But while no single 10 observatory may hope to cover the ground com- pletely, it may reasonably expect to contribute toward the solution of the problem and thus to aid in attaining a clearer comprehension of that great process of evolution which has produced the earth and its inhabitants. It is evident, from the nature of the case, that the plan of attack must be broad and elastic, utilizing a variety of powerful methods toward a common end. But in the presence of innumerable interesting objects for study, there is danger of a scattering of effort and a mere multi- plication of observations. Every possible astro- nomical observation might have a bearing on our problem and thus seem to justify its making; but unless it were made as an element in some general plan, an indefinite amount of energy might be spent without avail. A common scheme should tie together many diverse investigations, multi- plying the intrinsic importance of each because of its bearing on all the others. THE SUN. Let us begin with the sun. Its prime interest and importance, as the source of the light and heat on which we all depend, would be sufficient reason for its special consideration. But equally signifi- cant is the fact that the sun is the only one of all the stars which lies near enough to the earth to be studied in detail; each of the others is reduced by distance to an infinitesimal point of light, which the most powerful of telescopes can not magnify into an II appreciable disk. We may safely infer, from many observations of recent years, that thousands of the stars are almost identical in character with the sun, though some are much larger or smaller and some are in earlier or later stages of development. But if we wish to know what a star really is we FIG. 5. Direct photograph of the Sun. must approach it closely, and this is possible only in the case of the sun. Indeed, because the sun was regarded as so important, offering so many opportunities to increase our knowledge of its nature, the observatory was conceived primarily 12 for solar research. But the necessity for seeking, among the stars and nebulae, for evidence as to the past and future stages of solar and stellar life, rendered a broadening of scope advisable from the outset. Much attention is therefore devoted to the sun as the chief among the stars, but the essential means of attacking the more distant objects of the universe have also been provided. AUXILIARIES OF THE TELESCOPE. Ten years ago the possibilities of the spectro- heliograph, as a means of increasing our knowledge of the invisible atmosphere of the sun, had become apparent. This instrument, which was clearly FIG. 6. The Kenwood Spectroheliograph. susceptible of further improvement and develop- ment, was accordingly chosen as one of our chief auxiliaries for the study of the sun. Of still wider range of application was the solar spectroscope, previously used almost exclusively as a visual instrument of small dimensions attached to a moving telescope tube. Rowland had invented 13 the concave grating, and used it for his epoch- making photographic studies of laboratory spectra and the spectrum of sunlight. But his solar image was small and unsteady and there remained a most promising opportunity to apply a powerful photo- graphic spectroscope to the investigation of sun- spots, the chromosphere, and other details of a large solar image. Solar spectroscopy was far behind laboratory spectroscopy and new types of FIG. 7. The Snow Telescope. instruments were clearly demanded. It was evi- dent that for efficient use in photography the spectroscope of the day must be greatly lengthened, thus making it too long to serve as an attachment to a moving telescope. Accordingly it became necessary to develop a suitable form of fixed tele- scope, capable of forming a large and sharply defined image of the sun on the slit of a long fixed spectrograph. THE SNOW TELESCOPE. A step in this direction was the horizontal tele- scope, which the late Miss Snow had presented to the Yerkes Observatory through the kind interest and assistance of Dr. George S. Isham. It*con- sists of a coelostat, with a plane mirror 30 inches in diameter, rotated^by clockwork at such a rate FIG. 8. Ccelostat and second mirror of Snow Telescope. as to keep the beam of sunlight, reflected from its silvered (front) surface, in a fixed position on a second plane mirror standing above and south of it. From this mirror the beam is reflected nearly horizontally to a point 100 feet north, where it falls on a 24-inch concave mirror of 60 feet focal length, which forms a solar image about 6^2 inches 15 in diameter on the slit of the spectrograph or spectroheliograph. Loaned by the University of Chicago, and set up on Mount Wilson at a time when the Solar Obser- vatory was at work as an expedition from the Yerkes Observatory, the Snow telescope was found to have advantages and defects characteristic of a new instrument. Currents of warm air, rising from the hot soil of the mountain summit and carried across the entering beam of light, decreased the sharpness of the image during the hotter hours of the day. The sun's direct rays warped the telescope mirrors, changing the focus and blurring the details of the image after a few minutes of exposure. But the worst of these difficulties were soon overcome by observing in the early morning or late afternoon, shielding the mirrors from the sun between exposures and cooling them with blasts from electric fans. Thus controlled, the Snow telescope yielded excellent photographs of the solar atmosphere and justified the hopes we had entertained of its performance. WHAT IS A SUN-SPOT? Sun-spots, though known and studied for 300 years, offered most promising opportunities for research. Evidently there was much to be learned from an investigation of their spectra, which had never been attempted with adequate instrumental means. To produce these spectra, the light of a sun-spot was passed through a narrow slit, and i6 thence to a lens of 18 feet focal length, which rendered the rays parallel; they then met a grating of polished metal, ruled with some 15,000 lines to the inch; this analyzed the composite light into its constituent parts and returned the rays through the lens, which formed an image of the long spec- trum band on a photographic plate below the slit. With this spectrograph, constructed in our shop in Pasadena and mounted for use with the Snow FIG. 9. Direct photograph of Sun-spot. telescope, it soon became an easy matter to photo- graph sun-spot spectra. The curious widened lines, the much debated bands, and the strength- ened and weakened lines were thus accurately recorded for study. What is the cause of these peculiarities? The first step was to test by labor- atory experiments the hypothesis that some of them are due to reduced temperature of the spot vapors. In studying such phenomena the spec- troscopist is in a position much like that of an archeologist endeavoring to translate an unknown language. A bilingual inscription, containing an expression of the same fact in both celestial and terrestrial characters, is what he requires, and this a suitably equipped physical laboratory is often capable of supplying. In the solar spectrum we can photograph about 20,000 lines, distributed irregularly from the red to the violet, and throughout the invisible regions beyond. Perhaps some of these are due to iron. FIG. 10. Sun-spot Spectrum, a Solar and b Spot Spectrum widened; c from original negative, Spot Spectrum in middle. To settle this it is only necessary to vaporize some iron between the poles of an electric arc and photo- graph its spectrum beside that of the sun (Fig. n). Some 2,000 solar lines are found to coincide in position with lines of iron. As these lines are given only by iron, we may conclude at once that this element exists in the solar atmosphere. So much for the chemical identification of lines. We may next interpret their peculiarities. In the case of sun-spots we suspected that certain changes i8 in the relative intensities of the lines were due to a reduced temperature of the spot vapors. To test this, the spectrum of iron vapor in an electric arc was photographed at different temperatures. Some of the lines were found to strengthen, others to weaken relatively, as the temperature was FIG. ii. Iron (above) and Solar Spectrum (below). reduced. When compared with the iron lines in sun-spots the changes were seen to be of the same kind. The same test, applied to the vapors of chromium, nickel, manganese, titanium, and other metallic elements, previously identified in spots, FIG. 12. Effect of temperature on Spectrum of Vanadium: a In Carbon Arc; b, c, and d in Electric Furnace at temperatures of 2600, 2350, and 2150 C., respectively. 19 gave the same result. It thus became clear that sun-spots actually are regions of reduced tempera- ture in the solar atmosphere. The next step bore out this conclusion. If the solar vapors are cooler in spots than in the general atmosphere of the sun, then it may be possible for some of them to unite chemically. Thousands of faint lines in the spot spectra were measured and identified as band lines dueto chemical compounds. Fowler, who had also worked with success on the strengthened and weakened lines, found magne- sium hydride. Titanium oxide and calcium hy- dride were identified in our laboratory.- Thus we began to form a new picture of these regions of the solar atmosphere and to recognize the chemical changes at work in the spot vapors. SOLAR METEOROLOGY. Meanwhile systematic work was in progress with the spectroheliograph, which gives images of the sun in monochromatic light, showing the dis- tribution of some one vapor in its atmosphere. In the favorable California climate it is possible to photograph the sun on about 300 days of the year (in one season on 113 successive days). Every clear morning, and frequently in the after- noon, the instrument was at work, making pictures of the great gaseous clouds in the solar atmosphere. These were first observed many years ago as solar prominences, rising high above the sun's limb at total eclipses, when the bright light of the disk was 20 cut off by the moon. The spectroheliograph not only permits the prominences to be photographed on any clear day, but discloses extensive clouds of calcium, hydrogen, iron, and other vapors, which do not rise high enough to be observed in elevation at the limb, but are recorded (as flocculi) in pro- jection against the bright disk. To the eye at the telescope, or in direct photographs of the ordinary FIG. 13. The Chromosphere photographed without an IJclipse. kind, these flocculi are wholly invisible. The spec- troheliograph brings them to view by excluding from the photographic plate all light except that due to calcium or hydrogen, as the case may be. The measurement of these plates with the helio- micrometer (Fig. 55), an instrument devised and constructed in our instrument-shop, gave directly the latitudes and longitudes of the flocculi, without 21 the extensive computations required when the ordi- nary type of measuring machine is used. Their change of position. from day to day yielded a new determination of the law of the solar rotation, which was found to differ at the calcium and hydrogen levels. At the lower level of the calcium flocculi the period of rotation at the sun's equator is 24.8 days, increasing gradually to 26.8 days at 45 latitude. In other words, the gaseous sun KIG. 14. Solar Prominence 80,000 miles high. does not rotate like the solid earth, on which points in all latitudes complete a rotation in 24 hours. It turns more and more slowly as the poles are approached, points in high latitudes lagging behind those nearer the equator. If this could happen on the earth, Jacksonville, which is almost due south of Cleveland, would be far to the east of it 24 hours hence. In the higher levels of the solar 22 atmosphere, where the hydrogen flocculi float, the period of rotation for any latitude is less than for the levels below, but the difference in rotation time between pole and equator is less marked than in the lower atmosphere. FIG. 15. a, Direct photograph of Sun, August 31, 1906; b. Calcium (Hz) Flocculi at same hour. SOLAR AND STELLAR SPECTROSCOPY. Since the flocculi are constantly changing in form, they are not very satisfactory objects for rotation measurements. Much more accurate results can be obtained by measuring, with a powerful spectrograph, the velocity of approach and recession of the east and west edges of the sun. The east edge is moving toward the earth on account of the sun's rotation; this causes a dis- placement of the spectrum lines toward the violet (Fig. 1 6) . At the west edge, which is moving away, the lines are equally displaced toward the red. The double displacement, measured at different lati- 23 tudes, gives the velocity of approach and recession in kilometers per second. An investigation of this kind threw much new light on the peculiar law of the solar rotation, giving with high precision the rotation period at different levels and the change in its value from equator to pole. The Snow telescope thus proved its usefulness for a wide variety of observations, most of which we could not have made with moving telescopes of the standard type. In addition to the work already mentioned, the 1 8-foot spectrograph FIG. 16. Spectra of east (a) and west (ft) edges of Sun, showing dis- placement caused by solar rotation. yielded excellent photographs of spectra of various parts of the solar disk, revealing numerous pecu- liarities in the spectrum near the edge of the sun. Although not designed for stellar work, the Snow telescope also permitted photographs of the spec- trum of Arcturus to be taken with a powerful grating spectrograph. When compared with the spectra of sun-spots, the relative intensities of the lines were found to be similar, indicating that Arcturus is cooler than the sun, a fact of impor- tance in its bearing on the question of stellar evolution. 2 4 THE 6O-FOOT TOWER TELESCOPE. But the Snow telescope was not free from limi- tations. During long exposures its mirrors were seriously distorted by the sun's heat and the effect of heated air from the earth was plainly shown by a blurring of the solar image. To obviate or reduce FIG. 17. 6o-foot Tower Telescope. these difficulties the vertical or tower telescope was devised, and constructed in an inexpensive form. After reflection from two plane mirrors at the summit, the sun's rays pass through a 1 2-inch objective of 60 feet focal length, which forms an image of the sun on the slit of the spectrograph in the observing-room at the foot of the tower. The FIG. 1 8. Hydrogen Flocculi surrounding Sun-spots, showing right- and left-handed Vortices, September 9, 1908. mirrors, much thicker than those of the Snow telescope, are but little affected by the sun's heat. Elevated 60 feet in the air, they also escape some of the warm currents rising from the hot soil. The results are a decided improvement in the sharpness of the image and a prolongation of the period during which good observations are possible. Another advantage, quite as important, follows from the use of an underground chamber to con- 26 tain the spectrograph, now increased in length from 1 8 to 30 feet. The gain resulting from its greater stability and from the constancy of tem- perature of the grating at the bottom of the well was plainly apparent in the new photographs of spot spectra, which brought out details previously unrecognized. SOLAR VORTICES AND MAGNETIC FIELDS. The development of the spectroheliograph had also advanced another step. The dark hydrogen June 2 June 3 4 h 58 m P.M. June 3 5t>i 4 "> P.M. June 3 5 h 22 m FIG. 19. Hydrogen Flocculus sucked into Sun-spot, June 3, 1908. flocculi, first photographed at the Yerkes Observa- tory in 1903, had hitherto been recorded only with the blue hydrogen lines. In 1908 the new red- sensitive plates of Wallace, applied to the pho- tography of the sun's disk with the red (Ha) line of hydrogen, gave results of great interest. In the 27 higher part of the hydrogen atmosphere, thus revealed in projection against the disk, immense vortices were found surrounding sun-spots (Fig. 1 8) . This led to the hypothesis that a sun-spot is a solar storm, resembling a terrestrial tornado, in which the hot vapors, whirling at high velocity, are cooled by expansion, thus accounting for the FIG. 20. jo-foot Spectrograph, with Polarizing Apparatus above Slit. observed intensity changes of the spectrum lines and the presence of chemical compounds. But the observed widening of many spot lines and the doubling or trebling of some of them re- mained inexplicable until the vortex hypothesis suggested an explanation. Thomson and others had shown that electrons are emitted by hot bodies; hence they must be present in great numbers in 28 the sun. If positive or negative electrons were caught and whirled in a vortex they would produce a magnetic field, such as we obtain by passing an electric current through a coil of wire. Zeeman had discovered in 1896 that the lines in the spec- trum of a luminous vapor in a magnetic field are widened or (if the field is strong enough) split into several components (Fig. 21). Moreover, the light FIG. 21. Effect of Magnetic Field upon Lines of Iron Spectrum. In a the middle component and in b the outer components are cut out by a Nicol Prism; c, Spectrum without Magnetic Field. of these components is polarized in so characteristic a way that there can be no uncertainty in identi- fying the effect. Could this be the condition of things in sun-spots? The 3o-foot spectrograph of the tower telescope permitted the test to be made at once. The characteristic polarization phenomena appeared and one by one all of the distinctive peculiarities of the Zeeman effect were made out. Thus direct evidence, open to only one interpretation, proved the existence of magnetic fields in sun-spots, and strengthened the view that these are caused by electric vortices. This conclusion, in common with many others regarding the nature of sun- spots, could not have been obtained without the aid of the physical laboratory. A B FIG. 22. Effect of Nicol and Compound Quarter-wave Plate upon (.4) Lines of Spark in Magnetic Field and (fl) Solar Line in Spectrum of Sun-spot, a Red Components; b Violet Components; middle line of Triplet shows in B but not in A. Let us see how the effect of magnetism on light is studied. We place our iron arc or spark between the poles of a powerful magnet (Fig. 29) and pho- tograph its spectrum. The lines behave in the most diverse way, some splitting into triplets, others into quadruplets, quintuplets, sextuplets, etc. One chromium line is resolved by the magnet into 21 30 components. If a magnetic field is really at work in sun-spots we should anticipate a close correspond- ence between the behavior of each solar line and its laboratory equivalent. And this is exactly what we find (Fig. 22). Furthermore, the distance between the components of a line is directly pro- portional to the strength of the magnetic field. Thus, by determining the separation correspond- ing to a magnetic field whose strength can be measured in the laboratory, we may easily derive the strength of the field in sun-spots. FIG. 23. Waterspout off the Coast of Sicily. SUN-SPOTS AND FLOCCULI. We can not enter here into the various appli- cations of this conclusion to the explanation of solar phenomena. If we could see a single sun- spot from a point beneath the solar surface, it would probably resemble a terrestrial water-spout or tornado, though its cross-section, instead of being a few hundred feet, would be hundreds of miles. The strength of the magnetic field pro- duced, which is measured by the degree of separa- tion of the triple lines, increases with the diameter of the spot. The field is strongest near the center of the spot, where the lines of the triplet are most widely separated, and decreases to very low in- tensity at points just outside the edge of the penumbra. The spectrograph, when equipped with suitable polarizing apparatus, serves as an extraordinarily delicate means of measuring these fields, which can be observed in regions where they are not much more intense than the magnetic field of the earth. In this way it became possible, FIG. 24. Lines of Force about + and Poles of a Magnet. as described below (p. 43), to detect the compara- tively weak magnetic field of the entire sun. It has long been known that sun-spots usually occur in pairs, and our study of the Zeeman effect indicates that the two principal spots in such a group are almost invariably of opposite polarity. The natural inference is that we are here dealing with a semi-circular vortex (like half a smoke ring) the two ends of which, coming to the sun's surface 32 from below, appear to whirl in opposite direc- tions. But this hypothesis is still under investi- gation. Stormer of Christiania has developed for us the mathematical theory of the motions in the solar atmosphere of vapors within the influence of such magnetic fields, with results of great interest. FIG. 25. Two Bipolar Sun-spot Groups. The illustration (Fig. 25) shows that the hydrogen flocculi about a bipolar spot-group resemble the lines of force between two magnets of opposite polarity (Fig. 24). But similar structure might be produced by the direct hydrodynamical influence of the spot vortices upon the solar atmosphere .33 above them, and it is still a question whether this or the electromagnetic influence is predominant. The existence of electric vortices in sun-spots is indirectly shown by the presence of the magnetic fields, which presumably could not be produced in the sun by any other means. But direct-proof of the existence of these vortices was subsequently FIG. 26. Inflow at High Levels and Outflow at Low Levels above Spots. found by Evershed, of Kodaikanal, who measured the outflow of gases close to the solar surface and their inflow at higher levels. This work has been repeated and extended on Mount Wilson with highly significant results, affording not only an accurate picture of the conditions existing above a sun-spot, but also the means of assigning to each line of the solar spectrum a definite level in the sun's atmosphere. 34 WORK OF THE PASADENA LABORATORY. The value of laboratory work in the interpreta- tion of solar phenomena has already been illus- trated. Magnetic fields are detected by the split- ting and polarization of spectrum lines, differences of temperature by changes in their relative inten- sities, differences of pressure by shifts in their positions, etc. By producing such effects artifi- FIG. 27. Pasadena Laboratory. cially, with the aid of powerful electric furnaces, pressure pumps, and other physical instruments, we can imitate a great variety of celestial phe- nomena and interpret complex and obscure pecu- liarities. It is thus plainly apparent that a phys- ical laboratory is a necessary adjunct of an astro- physical observatory. Our experience has shown 35 that the literature of spectroscopy almost never contains the information we require. The only way to obtain sufficient data is to produce them ourselves, under conditions within our own control and adapted to meet the manifold requirements of the observed astronomical phenomena. More- over, the performace of a few experiments never suffices. It becomes necessary, not only to imi- tate an effect for a given line, or for a limited region of the spectrum, but to extend the observations over the whole range of spectrum available. It is easy to see that heavy tasks devolve upon our laboratory staff, both in observing and in the extensive work of measurement and reduction. It is a comparatively simple matter to show that a change of furnace temperature will modify the relative intensities of certain lines; but to measure the changes for the thousands of lines of iron, chromium, nickel, vanadium, and many other elements recognized in celestial objects is a task requiring years of continuous work. So with all of the other effects of pressure, magnetic field, change of potential of the electric discharge, etc. It is evident why the simple beginnings of this laboratory on Mount Wilson have led to larger developments in Pasadena, where heavy electric currents and other facilities are available. The intrinsic value of the laboratory results, as distinguished from their usefulness for the inter- pretation of astronomical observations, should not be overlooked. Each investigation is equally applicable to the interpretation of fundamental problems of physics, particularly those concerned with the nature of radiation. The changes of intensity of spectrum lines produced by raising and lowering the temperature of the electric fur- nace help to indicate how the shock of molecular collisions may influence the motions of the elec- trons within the atom. The phenomena of the "tube arc" throw new light on radiation processes hitherto associated mainly with high electric po- FIG. 28. Electric Furnace. tentials. The complex phenomena of the Zeeman effect (as revealed in a comparative study, with powerful spectrographs and an intense magnetic field, of the lines of a long list of elements) furnish material available for wide generalizations, impor- tant in their bearing on theories of radiation and atomic structure. Thus the maintenance of this laboratory would be highly advantageous from the standpoint of the physicist, even if it had no 37 connection with the Observatory. The doubled efficiency resulting from the combination of the two is of the same character as that which follows from the joint prosecution of the solar and stellar work. PRESSURES AND MOTIONS IN THE SOLAR ATMOSPHERE. I have mentioned the displacement of lines by pressure as a laboratory problem. The applica- tion of the results to the interpretation of line FIG. 29. Magnet for Zeeman Kffect. displacements observed in various parts of the solar atmosphere has yielded much new informa- tion. If the vapors in question are moving toward or away from the observer, displacements due to motion must be distinguished from those caused by pressure; this can be done with certainty by 38 using the laboratory data, some lines being shifted by pressure to the red and others to the violet. It is found that some of the solar gases are rising, while others are falling. The pressure at different heights is then determined, and found to range from about an atmosphere close to the solar surface to exceedingly low values, such as we know in vacuum tubes, at elevations of several thousand miles. The delicacy of this method is illustrated by the fact that the spectrographs on Mount Wilson and in Pasadena show a distinct difference in the position of certain lines in the electric-arc spectrum, caused by the difference in atmospheric pressure between mountain and valley. Thus we are enabled to sound the solar atmosphere through all its depths and to learn of its phenom- ena at different levels. The same method, when applied to stars, has given us a preliminary deter- mination of the pressure in stellar atmospheres. THE "FLASH" SPECTRUM WITHOUT AN ECLIPSE. In designing the first tower telescope, one of the objects in view was to provide a means of photo- graphing the "flash" spectrum without the aid of a total eclipse. When the moon passes between the earth and the sun it cuts off the brilliant light of the solar disk and permits the spectrum of its gaseous atmosphere to be photographed. The narrow arc of light, coming from this luminous atmosphere at the moment when the sun's disk is completely covered by the dark body of the 39 moon, is passed through a prism, and the resulting series of bright lines is recorded upon a sensitive plate. But the study of this "flash" spectrum has been seriously hampered by its momentary visibility, occurring only at intervals of years. With the 6o-foot tower telescope the numerous bright lines of the "flash" can be photographed on any day of good definition, with a spectrograph more powerful than those used in eclipse observations. In this way we not only get a marked increase in the accuracy of measuring these lines and their dis- placements, but we also find it possible to study the phenomena of levels lower than are attainable at eclipses. Some remarkable modifications of the dark-line solar spectrum at the sun's limb have also been found on these photographs. THE I5O-FOOT TOWER TELESCOPE. The success of the first tower telescope indicated that the construction of a more powerful instru- ment, giving a larger image of the sun (16 inches in diameter), would be fully warranted. To secure the necessary steadiness of mirrors and lenses mounted 160 feet above the ground, the plan was adopted of incasing each steel member (leg or cross-bracing) of a skeleton tower within the corresponding hollow member of another skele- ton tower, with sufficient clearance to prevent contact. The inner tower thus carries the instru- ments; the outer tower carries the dome to cover them, while its members serve as an efficient FIG. 30. iso-foot Tower Telescope. wind-shield. Thus the requisite steadiness has been secured, in spite of the great height of the structure. In other re- spects the new tower is also a decided improve- ment, adding still fur- ther to the sharpness of the solar image during the warmer hours and thus increasing the du- ration of the period of observation, which with this telescope lasts throughout the day. The opening up of the various new fields of solar research taxes the capacity of all three telescopes, each of which is devoted to the work for which it is best adapted. A feature of the new tower telescope, which is quite as important as the enlarged solar image, is the spectrograph,now extended to a focal length of 75 feet and mounted in a deep well FIG. 31. Section of is ' Tower Telescope. 42 beneath the tower. With this powerful instrument the D lines of sodium, which are barely separated with the standard one-prism spectroscope of the chemist, are about 1.2 inches apart in the third- order spectrum. A photograph of the solar spec- trum on this scale, including the ultra-violet but ex- FIG. 32. Summit of iso-foot Tower Telescope. eluding the infra-red, would be 70 feet long. With the aid of Koch's recording microphotometer, the gain in precision of measurement is directly pro- portional to the length of the spectrograph. Thus, as compared with the 3.5 foot Kenwood spectro- 43 graph, of which it is the direct successor, the 75- foot spectrograph gives results fully 20 times as accurate. FIG. 33. Observing Room and ys-foot Spectrograph. THE SUN AS A MAGNET. What this means in practice is illustrated by a recent investigation. As already explained, many of the spectrum lines are split into two or more components by a magnetic field. A Nicol prism and quarter-wave mica plate, placed over the slit of the spectrograph, permit us in the laboratory to cut off either component at will. The use of a compound quarter-wave plate, made of narrow strips, gives the serrated appearance of the lines shown in Fig. 22. If the lines in the solar 44 spectrum are similarly affected, and if the degree of their displacement varies from pole to equator as calculation shows it should do on a magnetized sphere, we may conclude that the whole sun is a magnet. An extended investigation (rendered difficult by the very minute displacements of the solar lines, far too small to appear to the eye in the photographs) has led us to the conclusion that the sun is a magnet, with its poles lying at or near the poles of rotation. FIG. 34. D I,ines of Solar Spectrum, with Iodine Absorption Spectrum superposed. The sun in this respect resembles the earth, which has long been known to be a magnet. The general magnetic field of the sun, although about 80 times as intense as that of the earth, is so weak compared with the magnetic fields in sun-spots that the full power of the 75-foot spectrograph was required to reveal it. As the sun rotates on its axis, it permits the magnetic phenomena of all parts of its surface to be studied. Photographs of the spectra over a wide range of latitude are therefore made daily, in order to provide material for charts which will show the exact position of the 45 magnetic poles and the intensity of the field at different levels in the solar atmosphere. Our interest in the sun's magnetism is not con- fined to the field of solar physics; its study should aid in explaining the source and fluctuations of the earth's magnetism and in the interpretation of certain stellar phenomena. It is not improbable, as Schuster has suggested, that every star, and f-^ l\