SMITHSONIAN CONTRIBUTIONS to

VOLUME I

SMITHSONIAN INSTITUTION

WASHINGTON D. C.

Contents of Volume 1

No. 1. in . Pages i-, 1-181. Pub- lished December 19, 1956. No. 2. Papers on Reduction Methods jor Photographic Meteors. Pages 183-243. Published May 13, 1957.

Smithsonian

Contributions to Astrophysics

VOLUME 1, NUMBER 1 NEW HORIZONS IN ASTRONOMY

edited by FRED L. WHIPPLE

DIRECTOR, ASTROPHYSICAL OBSERVATORY SMITHSONIAN INSTITUTION

flfe

SMITHSONIAN INSTITUTION

Washington, D. C.

1956 UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1956

For sale by the Superintendent of Documents, U. S. Government Printing Office Washington 25, D. C. - Price $1.50 Smithsonian Contributions to Astrophysics The Smithsonian Institution's Astrophysical Observatory is passing through a period of reorganization. The scientific headquarters have been moved to Cambridge, ., where a close working association with the Harvard College Observatory is being developed. These changes, as well as the many advances in science and technology, us to reevaluate the basic scientific policies and goals of the Astrophysical Observatory. With the present rapid progress in basic science and its application to technology, research in the fundamental astrophysical processes of the , , , and is going forward at a rapid pace. Concurrently, our mushrooming technology has become more and more sensitive to these phenomena of the solar , heretofore thought to be of only academic interest. In keeping with these developments, the Astrophysical Observatory must broaden the scope of its activities while continuing its long-estab- lished tradition of research—a tradition that began with the pioneering studies on by the observatory's founder, Samuel P. Langley, and continued during of active investigations under Dr. C. G. in solar and terrestrial phenomena and their interrelationships. Thus, our future program will include not only , its variations, and its effects on our , but also other phenomena that affect the earth and its . We find, for example, that corpuscular radia- tion from the sun is associated with solar activity as well as with such geophysical phenomena as the aurorae, air glow, terrestrial magne- tism, the , radio transmission, and other observed phenomena. Meteoric bodies, dissipated from or chipped from asteroidal sources, not only produce meteors in our atmosphere and provide samples of cosmic material for our museums but also ionize the high atmosphere and contribute to its chemical composition, scatter in the zodiacal , and provide us with a cosmic ballistics laboratory of ultravelocity . The Astrophysical Observatory, therefore, faces widening horizons. It will broaden the scope of its studies of and atmos- pheric effects to include related phenomena within the such as , hypervelocity ballistics, and the . It will cooperate in exploiting new methods of research and technology such as nuclear processes and artificial . To strengthen its effort the Astrophysical Observatory will not carry out its research in isolation; it will coordinate its work closely with that of other astrophysicists and geophysicists who are studying similar problems. Although research in astrophysics continues to expand, no correspond- ing growth has occurred in the avenues of publication. Observational results are often so compressed in print that they cannot be analyzed in detail by independent investigators; and advances in technical procedures or instrumentations are often presented with such brevity that their value to the potential user is small. In an effort to affect these trends and to provide a proper communi- cation for the results of research conducted at the Astrophysical Observatory, we are inaugurating this new publication, Smithsonian Con- tributions to Astrophysics. Its purpose is the "increase and diffusion of knowledge" in the of astrophysics, with particular emphasis on problems of the sun, the earth, and the solar system. Its pages will be open to a limited number of papers b\r other investigators with whom we have common interests. We earnestly hope that Smithsonian Contributions to Astrophysics will serve to promote a greater understanding, appreciation, and enjoyment of that part of the in which we are privileged to live. FRED L. WHIPPLE, Director Astrophysical Observatory Smithsonian Institution Cambridge, Massachusetts November 14, 1956 A grant from the National Science Foundation toward the publication of New Horizons in Astronomy is gratefully acknowledged.

New Horizons in Astronomy

Preface The project of collecting a series of papers oriented toward new horizons in astronomy began in the 1954-55 Panel of Astronomy to the National Science Foundation. The initial impetus came largely from the enthusiasm of Dr. , who has presented some of his thoughts in a paper entitled "The General Needs of Astronomy" (Publ. Astron. Soc. Pacific, vol. 67, p. 214, 1955). The resulting deliberations led to the establishment of an ad hoc committee on the "Needs of Astronomy," of which I was made chairman. I felt that some of the goals of this com- mittee could be attained by publishing a collection of papers by leaders in the various fields of astronomy who would present their concepts of research projects most likely to advance the science of astronomy during the next decade. The purpose is at least fourfold: To provide material that might assist the National Science Foundation, in its allocation of funds, to support the astronomical research most likely to be productive. To help younger choose the most fertile fields for active research. To encourage established astronomers to pause for a moment and reflect on the most advantageous planning in their own fieldso f research, and to supply them with a broader insight into the activities in other fields of astronomy, possibly related to their own. To encourage research in new or neglected fields of astronomy, par- ticularly those that border on other areas of physical science such as , electronics, chemistry, , and the like. The publication of New Horizons in Astronomy has been supported jointly by the Smithsonian Institution and the National Science Founda- tion. The Astrophysical Observatory shares with the National Science Foundation the hope that these papers may promote intensified efforts and increasingly significant results in astronomical research. The hearty cooperation of the many busy scientists who took time from their other duties to write these papers should establish a debt of gratitude in the minds of all those who will profit by these contributions. I am deeply indebted to Mr. Paul H. Oehser, Mr. Ernest Biebighauser, and Mr. John S. Lea, of the Editorial and Publications Division of the Smithsonian Institution, and to Mrs. Lyle Boyd, of the Harvard College Observatory, for their great interest and effort in this project; otherwise these papers would not now rest in your hands. FRED H. WHIPPLE

Contents Page Techniques and Instrumentation : I. S. Bovven 1 Solar Instrumentation: J. W. and R. B. Dunn 5 and Scintillation: Geoffrey Keller 9 Spectroscopy: Charlotte E. Moore 13 Radial Velocities: R. M. Petrie 15 : K. Aa. Strand 21 Photoelectric Astronomy: A. E. Whitford 25 Interferometers in : B. F. Burke and M. A. Tuve . 31

Related Sciences Rocketry: R. Tousey 39 Computing Machines in Astronomy: Paul Herget 45 Turbulence: S. Chandrasekhar 47 Astroballistics: Richard N. Thomas 49 Electrodynamics of Fluids and Plasmas: Max Krook 53 Hydromagnetic and Problems: Eugene N. Parker 59

Solar System Earth's Magnetism: Walter M. Elsasser 67 Time: William Markowitz 73 Geographical Positions: John A. O'Keefe 75 Celestial : G. M. Clemence 79 Meteorites: John S. Rinehart 81 Meteors: Fred L. Whipple 83 Minor Planets: Paul Herget 87 Planets, Satellites, and Comets: Gerard P. 89 and : A. B. Meinel 95 Solar-Terrestrial Relationships: and Communications: Walter Or Roberts 99 : Goldberg and Donald H. Menzel 103

Stars and the Spectral Classification: W. W. Morgan . 115 Stellar : Marshal H. Wrubel 119 Aerodynamic Problems in Stellar Atmospheres: Richard N. Thomas. 123 Emission *: Lawrence H. Aller 127

381014—56 2 CONTENTS

Page and the Galaxy—Continued Double and Multiple Stars: Otto Struve 131 : U. Van Wijk 135 Variable Stars: C. Payne-Gaposchkin 137 Interstellar Matter: Lyman Spitzer, Jr 139 21-Centimeter Research: Bart J. 141 Some Observational Problems in Galactic Structure: Harold Weaver. 149 Galactic Clusters and Associations: A. Blaauw 159 Globular Clusters Observed through a Crystal : W. A. Baum. . 165 and Evolution: M. Schwarzschild 177 Radio Sources: John D. Kraus 179 Techniques and Instrumentation

Optics

By I. S. Bowenl

Geometrical optics is one of the oldest of critical definition for of large sciences. Most of the basic laws have been aperture. telescopes have been suc- known for centuries, yet in the past two or cessfully made with apertures up to 48 inches three decades advances in optical designs have with a focal length of 120 inches. However, in revolutionized astronomical techniques. An order to avoid chromatic aberrations the focal outstanding example of this is the Schmidt length must increase as the 3/2 power of the . Until the discovery of the principle aperture. Thus the minimum of a 200- of this instrument in 1931, camera optics inch Schmidt is 1,000 inches. The length of providing critical definition over a field greater the tube would have to be twice this, or 2,000 than two or three degrees in diameter were inches; i. e., three times that of the present limited to focal ratios of d/f =1/5 or less. . This would make the cost of With the critical definition the telescope tube and the dome to house it can be achieved over wide fields at focal ratios prohibitive. as fast as dff=l. This 10- to 100-fold gain in Unfortunately the field of good definition of speed has opened up many new fields in a high-speed paraboloid is very small; less than astronomy. }{ inch in diameter for the Hale telescope in Furthermore it has been possible to design, which d/f= 1/3.3. Corrector lenses designed for special purposes, many modifications of the by Dr. F. E. and placed in front of the Schmidt , some of which achieve even plate have successfully increased this area of greater gains in speed and field. These include good definition to a diameter of 3 or 4 inches. the Baker Super-Schmidt for meteor photog- This still falls far short of the 14 X 14-inch area of raphy and the solid-block and thick-mirror critical definition attained with the 48-inch Schmidts, the Schmidt-aplanatic com- Schmidt, which operates at the even higher bination, and the twice-through corrector plates speed of d/f= 1/2.5. Because of this small for spectroscopic applications. There are un- field, the use of large telescopes such as the doubtedly many additional modifications that 200-inch is impractical for most survey pur- should be explored. For example, the focal poses. surfaces of most of these Schmidt are Another important need at the present time curved with a radius approximately equal to is an objective prism or similar device that may the focal length. Thin glass plates can be be used for the wholesale classification of spectra bent to radii greater than 18 inches and simple of stars of fainter than can now be field flatteners yield satisfactory results in reached. Thus, because of the deterioration of cameras with focal ratios up to about d/f=l. the definition of the spectra by "seeing," it is However, to exploit fully the still higher not feasible to use an objective prism on tele- speeds in the shorter focal lengths it will be scopes of appreciably more than 10-foot focal necessary to develop either field flatteners that length. Furthermore, the limit on the exposure are effective at the higher focal ratios or to find time set by the is essentially a flexible emulsion base that is stable against the same with an objective prism as for direct changes in and humidity. photography. If the usual dispersion of 200- Another urgently needed development is an 300 A/mm is used and the spectra are widened optical system that will provide a wide field of to 0.2 or 0.3 mm, the is spread over an 1 Director of Mount Wilson and Paloraar Observatories. area 1,500 to 4,000 times that of a image, SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS with the result that the for light which is incident on them. Undoubtedly the spectra is 8 or 9 magnitudes below the limit the local air turbulence due to this could be sub- for direct photography. For a 10-foot focal stantially reduced by proper light shields and length the limiting magnitude for direct pho- by artificial cooling of parts that cannot be tography isMph=21 and therefore for spectros- shielded. Useful studies could also be made to copy Mph=12 or 13. This falls far short of the determine the most effective method for thermo- magnitudes that may be reached with slit stating an optical instrument such as a spectro- spectra. Thus with the 200-inch telescope a graph without setting up internal air currents 1-night exposure enables one to reach the 15th and turbulence that often cause serious deteri- magnitude at 38 A/mm and the 18th magnitude oration of the image. at 170 A/mm. At the present time there is no Every large telescope is essentially a collector equipment available for picking up unusual of light which is focused on some type of faint objects for detailed study at these higher receiver. The properties of this receiver play dispersions with a slit spectrograph. The antici- a very large role in the proper design of the pated development of the image tube, discussed telescope. Thus up until the present century below, will render this problem even more serious practically all astronomical observations were by pushing the limit of the slit spectrograph made visually; i. e., the eye was the from 2 to 5 magnitudes still fainter. receiver. The range of wavelengths to which One suggested solution of this problem is the the eye has appreciable sensitivity is very small use, as a multislit, of a high-contrast positive (1,000-1,500 A) and the field of view of critical print of a given star field taken with the same definition is also very limited. Furthermore, telescope. With a wide-angle telescope and a the lens of the eye is a part of the optical spectrograph having wide-angle collimator and system and sets an upper limit to the size of camera optics, a large number of spectra could the cone of light that can be concentrated on be obtained simultaneously without interference the sensitive surface. For this type of receiver from the sky background. Unfortunately, if the simple two-lens refractor operating at focal Schmidt optics are used throughout, the field ratios of d/j'=1/15 to 1/20 was ideal. As a flatteners necessary to bring the curved fields of result nearly all large telescopes constructed the telescope and collimator into coincidence in the last half of the 19th century were of this destroy the off-axis collimation. Some other type. wide-angle system must be devised for the By 1900 photographic procedures were rapidly purpose. replacing visual observations. The silver-halide The efficiency of slit spectrographs would be grains, however, are sensitive in the substantially increased if larger gratings could to beyond the limit of transmission of the be made or if gratings could be ruled which pro- atmosphere while the more recent development vide higher angular dispersion and still retain a of dyes has produced plates that are sensitive high concentration of light in one order. In well out into the range. Also, in order order to avoid interference between collimator that an image be recorded on a photographic and camera optics, this grating should prefer- plate it is necessary that a certain minimum ably be of the transmission type. The so-called number of quanta fall on a unit area of the "venetian-blind type" described in "Astronom- plate. This limit is such that in order to ical Spectrographs: Past, Present, and Future," photograph faint and nebulosities in in ' 'Vistas in Astronomy,'' is one possible solution a feasible exposure time it is necessary to use a if the necessary ruling techniques can be light collector with a focal ratio of at least developed. djf=l/5. This fact resulted in a sudden shift Another problem that needs investigation is early in the present century to the use of the that of seeing. Presumably very little can be reflector, which provides the complete achro- done to reduce turbulence in the upper atmos- matism and greater speed necessary for the phere. However, in solar instruments much of efficient use of the new type of receiver. the disturbance is caused by the heating of the Not only were the properties of the photo- optical parts and their surroundings by the sun- graphic plate major factors in causing the shift NEW HORIZONS IN ASTRONOMY

from refractors to reflectors, but they have raphy of the sky. Indeed it now appears that played an important role in fixing many of the in the foreseeable future the limit of detectabil- design criteria for both the reflector itself and ity of a faint star image will approach the - the auxiliary equipment, such as spectrographs, retical limit. This limit is set by the necessity to be used with it. One of these properties is of collecting a larger number of quanta from the linear resolving power of the plate. This the star than the probable random fluctuation resolving power sets the minimum focal length in the number of quanta received from the sky of the telescope or of the spectrograph background in the area of the sky covered by camera if a given angular resolving power or the object. wavelength resolution, AX, is to be achieved. These developments are still in such an early Furthermore, in all of these instruments this stage that it is too soon to predict their possible plate resolving power is the factor effect on telescope design. Thus the linear considered in setting up optical design and resolving power will depend on the final type construction tolerances. Another property of of receiver adopted. It is also quite probable the plate, namely, that it can record a very that the new receiver will not require, like the wide field at one time, led to the development of photographic plate, a minimum concentration various wide-field optics culminating in the of quanta per unit area for a satisfactory record. Schmidt camera. Similarly the necessity of One criterion for this will depend on the pos- concentrating a certain minimum of light per sibility of eliminating spurious thermal unit area in order to obtain an image has been by refrigeration or other means. In any case the chief cause of the design and construction the much greater efficiency of the photo- of optical systems of extreme speed («?//= 1 or electric process will permit the record of a greater) for spectrograph cameras. given object to be obtained in very much less It is now becoming apparent that another time than that required by the photographic change in the receiver for astronomical ob- process. The shift in these two properties will servations is in the offing. Thus photoelectric probably eliminate the necessity for optics of surfaces have been developed, and are in use extreme focal ratio such as the types recently for television and other purposes, that have a developed for both direct and spectroscopic quantum efficiency from 10 to 100 times as observations. Regardless of the properties high as that of the best photographic plates. of the photoelectric receiver as finally developed, The electrical circuits used in some of these substantial changes in telescope and spectro- receivers permit the reduction or elimination graph design may be expected, if for no other of a uniform background of the type caused by reason than the elimination of many of the scattered starlight and the permanent aurora, present restrictions set by the properties of the which sets the present limit in direct photog- photographic plate.

Solar Instrumentation By J. W. Evansx and R. B. Dunn

The technical problems of above the regions of the earth's atmosphere are decidedly different from those in any other that absorb the solar in these wave- field of astronomy because of the abundance length regions and prevent their observation of light and the large apparent diameter of the from the ground. sun. The latitude for the exercise of ingenuity This brief description can give no concept of in securing more and more refined information the difficulties encountered in the work, nor is practically unlimited, and affords a kind of of the years of patience and imagination that satisfaction rarely found in any other field. were required to overcome them; but the re- The next 25 years promise to be the most sults so far have more than justified the effort. exciting and the most productive since the for- The expected spectroscopic features duly ap- mulation of the laws of spectrum analysis by peared, but some very unexpected emission Kirchhoff. Since 1945, the conception and lines were also recorded. The expected lines construction of new instruments and techniques strengthen our confidence in the broad features have accelerated markedly, as have the theoret- of existing solar theory, and the new lines ical advances which designate the critical present new conditions which any valid theory observations we need to determine the physical must satisfy. The observations have character of the sun. We have barely begun unveiled a hitherto inaccessible part of the to get acquainted with the new tools and to solar spectrum, but the quick glance we have exploit their possibilities, but the results are been allowed sharpens the appetite for more already impressive. details, more quantitative measurements, and Most of the new developments are projections a further extension to still shorter waves. of older devices and methods, with refinements can still do a great deal, but there that greatly increase their power. Two of can now be little doubt that a man-made satel- them, however, introduce entirely new methods lite eventually will be our platform for solar that are sure to be exceedingly fruitful: the research above the absorption of the terrestrial observation of the far ultraviolet spectrum of atmosphere. New difficulties will appear and the sun from rockets, and the recording of solar be overcome, and new data on the . spectrum will further delineate the structure The rocket observations are a fine example of the sun. Such advances, inevitably, will of cooperative effort in several fields to solve a raise new questions in place of the old ones very difficult technical problem. The record- solved, but they will concern questions of de- ing instrument is either a spectrograph designed tails which we cannot ask at present because to work in the 900- to 3,000-angstrom range or we do not know enough. an X-ray ionization for the 1- to It was discovered during War II 5-angstrom region. The recorder must be that the sun was the source of occasional mounted in the rocket with a device which keeps radio noise strong enough to interfere with it pointed at the sun with an accuracy of a few operations. This discovery has led to minutes of arc, in spite of lively gyrations in a new branch of solar astronomy which we have its supporting platform. Finally, the rocket just begun to explore in its general outlines. itself must be capable of carrying its load to Instead of the 2-to-l range in wavelength of an altitude of 100 km or more in order to rise the , the observable radio

1 Sacramento Peak Observatory, , X. Mex. emission from the sun covers a range of at SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 least 1,000 to^l. This provides a tool for contribute significantly to the daily analysis in depth. Shortwaves in the centi- coverage obtained elsewhere? Can the solar meter range originate near the visible surface activity be adequately described from present of the sun. As we go to longer waves, the solar surveys? In other words, the observer point of origin rises through the corona to should have a specific problem in mind when heights of some hundreds of thousands of building the instrument and when taking the kilometers for the longest waves of 10 meters observations. or so. Like the light spectrum, the In instrumental development there is room of the radio spectrum depends on physical for every type of genius. To get an ingenious conditions at its point of origin and in the idea and show that it works is not enough. higher layers traversed by the radiation in The inventor, or someone else, must still leaving the sun. The interpretation of much develop the instrument to the point where any that has been observed is not yet clear, but capable observer can use it routinely to obtain there can be no doubt that solar radio astronomy reliable data. A typical example of instru- will occupy an extremely important place in ment development is the coronagraph. future research. showed that the coronagraph could be built, The design and construction of new solar and he proceeded to build one. He used it instruments at observatories all over the world with the utmost skill to solve several important continues at an ever-increasing pace. Some coronal problems, and, incidentally, demon- of them, like the Babcocks' magnetograph, strated its capabilities. Since then a great are the response to a particular need. Others deal of work has resulted in developing the are new versions of tried-and-true devices, instrument, as a coronagraph, and in making like spectrographs, for which there are sure to it more convenient to operate. It is hoped be many uses not necessarily specified in ad- that due to additional scale, resolution, photom- vance. In a field where we have still so much etry, and systematic use, Lyot's ideas will to learn, we should plan our new equipment tell us even more about the sun than did his with care in order to avoid unprofitable duplica- first observations. In some cases the con- tion. veniences themselves make the difference be- In considering solar instrumentation, both tween taking or not taking the critical observa- present and future, we should always keep ia tion. mind our real object: to find out something The development of the completed instru- about the sun. Specifically, we want the statis- ment is very often the result of a team effort. tics, , and physics of the solar We need new gadgets and new ideas, but in atmosphere, and a good theoretical model the solar field we also need people who take that explains all the observations. We must existing ideas and instruments and add such approach the problem from every possible vital factors as photometric standards, direct direction. Routine daily survey programs, photoelectric photometry, differential photom- detailed studies of active regions, critical etry, seeing cancellation devices, and conveni- observations designated by theoretical models, ences that speed the operation. new types of observations—all should be used As elsewhere, solar instrumentation is getting to solve the problem. so complex that we need specialists in the opti- Each new instrument acquired should fill a cal, electronic, and mechanical aspects and in specific need. A careful investigation is often data reduction, who will cooperate with the solar required to determine whether the planned physicist to produce the final result. It is no observation is already being carried out else- longer feasible for a person to be skilled in all where. Especially in planning a survey pro- fields necessary to the solution of a problem. gram, the observer should ask himself: Should He must work with a group, and the group I make an effort to take a few prominence members must have mutual confidence in each movies, when other observatories have collected other. A great deal of intercommunication is such gross quantities of film that they defy essential. Finally, the group must be compe- any sort of reduction ? Will my daily survey of tently directed and coordinated. NEW HORIZONS IN ASTRONOMY

The value of some very simple solar observa- with light from neighboring areas. The idea of tional device in an educational institution a scarcity of sunlight is perhaps a novel one, but should never be underestimated. An example may not seem so surprising when we consider is a prominence birefringent filter. The direct that we would like right now to take the light contribution to research on the sun may be small from an area of 0.01 square millimeter in a or none, but the stimuli provided by such fa- 300-mm solar image, spread it out into a visible cilities may be decisive in attracting capable stu- spectrum nearly a hundred yards long, and dents to solar research, which greatly needs then determine the intensity in an area of 0.01 skilled workers. square mm of this spectrum. The one thing we Until a few years ago solar research was main- can do is to increase efficiency in the use of the ly concerned with broad average characteristics. quanta we receive. The photographic plate One studied the spectrum without regard for utilizes only one in a hundred of the incident small variations from point to point on the sun. quanta. Modern photoelectric surfaces are The was treated as a homogene- about 10 times as efficient, and we have already ous atmosphere of uniform temperature. More begun to take advantage of this improvement. recently the importance of the local differences The problem of the future is to find still more has been recognized and the tendency has been efficient photoelectric surfaces, and to develop toward a finer grained analysis, in which many accurate and convenient methods for recording small areas of the solar surface must be studied continuously the photoelectrons emitted from separately. The result is an enormous increase each point of the surface. in the volume of information required, since we Once the observational data are secured, the now have to consider many small areas instead problem of reducing them to numerical quan- of a single large one. Gathering data, reducing tities must be faced. This has always been a them to useful numerical quantities, and analyz- laborious process, and the vastly greater volume ing the greatly increased volume of such nu- of data now required demands a radically new merical quantities are already acute problems, method of automatic reduction. The mag- and will become more serious as time goes on. nificent spectra coming from the new In gathering data, the unaccustomed pinch spectrograph at the McMath-Hulbert Observa- of the basic limitation of information contained tory emphasize the problem decisively, and I in the available light from the sun is already take them as a clear-cut example. The being felt. A telescope receives light from a McMath-Hulbert observers find that both the given area on the sun in the form of discrete Fyaunhofer lines and the solar continuum quanta at an average rate of n per second, in a show marked variations in the direction purely random fashion. This means that the perpendicular to the dispersion. These varia- number of quanta collected by the telescope in tions are due to spectroscopic differences in a given second differs from n in a random fash- the small regions imaged on the slit of the , with a mean percentage error of -j=- spectrograph. They consist in wiggles and changes of width in the lines, and changes in Hence a 1-second observation can at best de- the of the continuum. Thus the termine the brightness of the area with this spectrograms have variations which must be percentage accuracy, and the larger n is, the measured in two directions instead of one, along more accurate the observation. When we try the spectrum and across the spectrum. Now to observe smaller and smaller areas of the sun suppose we want to exhaust the information we are, in effect, reducing n in proportion, and contained on such a plate over a wavelength the fundamental errors of observation increase. range of only 5 angstroms. Let us say that the Unfortunately we cannot make up for this by slit was 1 cm long, that the resolving power of simply observing for a longer time. Seeing is the telescope was 0.2 mm along the slit, and never good enough to yield an undistorted and the resolving power of the spectrograph was unsmeared image of the sun for more than a 0.01 A. Complete information, therefore, calls fraction of a second, and the longer our observa- for the determination of the light intensity at tion the more our small area becomes confused 500 points along the spectrum for each of 50 s SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS points along the slit. In other words, we have have stimulated the development of information to make 25,000 intensity determinations for a theory, the science of extracting the maximum complete analysis, instead of the 500 that would possible information from a minimum of raw suffice if there were no appreciable variations data in the least time. Devices for automati- along the slit. In this instance the new cally performing various standard operations in dimension has multiplied the labor of analysis data reduction, and computers which handle by a factor of 50. The information is there in the numerical results with unimaginable speed, the plate, but the labor of extracting it has have been devised. The theoretical and phys- become prohibitive. ical machinery for the solution of the astronom- The case of the wiggly lines is only one ex- ical problem are now at hand, and the comput- ample of the trend toward the study of finer ers at least are already vigorously at work for details on the solar disk, in the chromosphere, astronomers in several fields. But as yet the and in the corona. The result is an unprece- automatic reduction of astronomical data to dented mass of information in the coded form the point where they are ready to be fed into of photographic , magnetic variations in a tape, or some other analog record. To be a computer has hardly been touched. The of any use, these records must be decoded and problem for the immediate future, therefore, is presented in numerical form. Once this has to develop the devices for bridging the gap been accomplished, we are confronted with between the output of a telescope and the input endless pages of numbers, all ready for interpre- to a train of automatic reducing equipment tation by the unhappy theoretician. Classical already available. The output of such a train methods of data handling will be quite in- is numerical information, ready either for the adequate to utilize more than a tiny fraction of computer or for direct analysis by the astro- the information contained in the observations, physicist if the quantity is manageable. The and future progress depends very heavily on importance of this problem should be em- finding the means for breaking this bottleneck. phasized. Unless it can be solved the effec- These problems of data reduction and analysis tiveness of many of the instrumental develop- on a large scale are by no means limited to ments for solar research will be limited to a solar astronomy. Industrial and military needs fraction of their real potential. Astronomical Seeing and Scintillation1

By Geoffrey Keller

Anyone who has made observations with large waves seem to be caused by nonuniform telescopes is only too painfully aware that the in the ionosphere. Ryle and Hewish most serious limitations to their performance are (1950) have observed fluctuations in the inten- usually connected in one way or another with sity of radio sources at meter wave lengths of problems of astronomical seeing. Images of the order of 10 percent and fluctuations in the stars are much larger than one would expect apparent positions of the radio objects of the them to be were their size determined only by order of a minute or so of arc. Such variations the aberrations of the telescope itself. Details give rise to serious observational problems, and, of extended objects, such as the and the though some observational uncertainties due to sun, are badly smeared when the seeing is poor. scintillation and bad seeing may be overcome by As a case in point, Baum (1955) gives the aver- making repeated measurements, the loss in effi- age image diameter seen with the 200-inch tele- ciency is very great. scope at about 2.5 seconds of arc, whereas in the The most extensive efforts that have been absence of bad seeing conditions the perfectly made to date to minimize the effects of the adjusted telescope should give images of di- earth's atmosphere on astronomical observa- ameter of the order of 0.04 seconds of arc. The tions have consisted of more or less empirical total amount of light received from a bright star comparisons of various possible sites for estab- by a telescope of moderate size also fluctuates lishing new observatories. When a very large in an irregular fashion, and in a manner which investment of this sort is to be made, it becomes can easily be shown to be due not to intrinsic economically feasible to spend a considerable fluctuations in the star but to the fact that the sum on extended tests. Such tests were made starlight must through regions of the prior to the construction of the Hale telescope earth's atmosphere wherein the density varies at Mount Palomar, and currently a new series irregularly. Protheroe (1955a) has shown that of tests is being made in connection with a with a telescope of 12% inches aperture the search for a site for a proposed American Ob- fluctuations in brightness of a bright star will servatory. Such surveys generally use tele- average around 10 percent of the mean, and will scopes and record the fluctuations in size and have frequencies ranging up to a thousand cycles brightness of a star such as (chosen per second. In pioneering investigations in this frequently because of its relatively fixed position field Mikesell, Hoag, and Hall (1951) obtained in the sky). similar results. Another potential means of reducing seeing Similar effects are encountered in radio as- effects is through the use of instruments mount- tronomy, and here also the responsible agency ed on high-altitude balloons, rockets, and even- is a region of nonuniform index of in tually extraterrestrial satellites. These methods the earth's atmosphere. Whereas in the case of might eliminate our seeing problems, but before optical scintillation the principal effects appar- adopting them we should know, for example, ently are caused in the troposphere and lower how high the instrument platform must be stratosphere, the majority of the effects on radio in order to be safely above the troublesome layers of the atmosphere. In the case of a 1 "Scintillation" is used here to mean only the changes with time of the brightness of the telescopic image of a star or other source. "Seeing" , the answer is likely to be of is construed to mean the degree of change in the position, shape, or size the order of hundreds of . There are of a stellar linage or of a detail of an extended object. 1 Perkins Observatory, Delaware, Ohio. obvious difficulties with such an approach 9 10 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. 1 when we attempt to apply it to large telescopes, ably by electro-optical methods, compensating so that it will probably be many years or even adjustments in tie optical system in the tele- centuries before the bulk of astronomical scope. Thus, if the air in a portion of the light observations can be made at extra-atmospheric path is unusually warm, the corresponding index stations. of refraction will be abnormally low and the Having picked a site where the disturbances velocity of light abnormally high. As a result, caused by intrinsic seeing and scintillation are a wave front tends to advance farther than it at a minimum, the is next interested should as it passes through this region. The in knowing how best to design his telescope compensating device attempts, in effect, tempo- and observatory so that thermal disturbances rarily to move the surface of the primary mirror in the air caused by their structure may be farther in the area where this portion of the minimized. A number of theoretical and ob- wave front is reflected, so that after the reflec- servational studies of the behavior of scintil- tion of the wave front the irregularities have lation have been made at the Perkins Observa- been ironed out. Methods of this type have tory that seem to point more and more to the certain drawbacks. For example, we must ob- conclusion that the local thermal irregularities serve objects that are sufficiently bright to give of the air may be the predominant cause of enough photons per second to activate the bad image distortion (as distinguished from device which controls the compensator; yet, bad scintillation). The work of the astronomers enough photons must be left over to make at Paris and Haute-Provence seems to have possible the primary astronomical observation. led them to similar conclusions, so that the In the preceding paragraphs we have looked new 193-cm telescope for the observatory at at the problems of astronomical seeing from the Haute-Provence is being very carefully de- point of view of the typical astronomer, to signed to minimize the temperature irregular- whom seeing is an unmitigated evil that must ities in the air in the tube and just outside be avoided or removed by whatever means come the dome (see , 1953). Should these to hand. To the atmospheric physicist and efforts be successful, a major improvement in meteorologist, the situation may be quite the the performance of large telescopes may result. reverse. The fluctuations in the light of a The importance of minimizing irregular bright star, known as scintillation, are known to refraction in the gas inside of auxiliary astro- be caused by the motion across the telescope nomical equipment has been well demonstrated objective of a complex pattern of and by the experience of observers with the new darks, known as the pattern or shadow 52-foot vacuum spectrograph at the McMath- bands. This shadow pattern is caused by the Hulbert Observatory (McMath, 1955). The passage of starlight through atmospheric irregu- steadiness and sharpness of the spectral lines larities which must occur at a considerable were greatly increased as the spectrograph was height. These irregularities diffract the light evacuated. Doubtless there are many other and cause interference and reinforcement of the pieces of auxiliary equipment whose perform- wave front at various points along the ground. ances could be greatly improved by a similar A fairly complete theory of the relation between procedure. the atmospheric irregularities and the pattern of lights and darks in the shadow pattern has Although the writer knows of no successful been developed by van Isacker (1953). With a instruments, so far devised, to compensate for reliable means of studying the structure of the (rather than eliminate) the effects of seeing, shadow pattern, much may be said about the some have been envisioned and might well be pattern of the density variations in the air itself. investigated (, 1953). Any device which A suitable method for observing the shadow attempts to compensate for seeing must de- pattern has been developed by the author pend, first, on knowledge of what sort of optical (1955) and by Protheroe (1955b). There seems irregularities exist in the air; this requires study to be considerable promise that by the use and of the distortion in the wave fronts in starlight extension of this method we might be able to which has just passed through a layer of irregu- obtain considerable information about the larities. The device must then make, presum- NEW HORIZONS IN ASTRONOMT 11

nature of clear air turbulence, its motions, its References height, and how these quantities vary with BABCOCK, H. W. meteorological conditions. At the same time, 1953. Publ. Astron. Soc. Pacific, vol. 65, p. 229. more empirical studies of the relationship BAUM, WILLIAM A. between scintillation and velocity are 1955. Sky and Telescope, vol. 14, pp. 264, 330. being made (Gifford and Mikesell, 1953; COUDER, ANDRE. 1953. Publ. TObe. Haute-Provence, vol. 2, No. 53. Gifford, 1955; Protheroe, 1955a; and Mikesell, GIFFORD, FRANK, JR. 1955). 1955. Bull. Amer. Meteorol. Soc., vol. 36, p. 35. Studies of the scintillation of radio objects GIFFORD, F., AND MIKESELL, A. H. are also being made, and much information 1953. Weather, vol. 8, p. 195. about the structure of the ionosphere is being KELLER, G. obtained (Pawsey, 1955). 1955. Journ. Opt. Soc. America, vol. 45, p. 845. Also, Contr. Perkins Obs., ser. 2, No. 5. To summarize, the study of astronomical MCMATH, R. seeing and scintillation is of vital interest to 1955. Sky and Telescope, vol. 14, p. 372. the astronomer in the same sense that studies MIKESELL, A. H. of optics and radio technology are of importance 1955. Publ. U. S. Naval Obs., ser. 2, vol. 17, pt. 4. to him. There is much to be learned about the MIKESELL, A. H., HOAG, A. A., AND HALL, J. S. subject, and there is considerable promise that 1951. Journ. Opt. Soc. America, vol. 41, p. 689. with more understanding will come ways and PAWSEY, J. L. means of greatly improving the performance of 1955. scintillation due to ionosphere focusing. Physics of the ionosphere. telescopes. Although seeing may be a nuisance The Physical Society, London. to the astronomer, it also is a tool for making PROTHEROE, W. M. both fundamental and routine studies of the 1955a. Contr. Perkins Obs., ser. 2, No. 4. earth's atmosphere. 1955b. Journ. Opt. Soc. America, vol. 45, p. 851. A few observatories in this country and Also, Contr. Perkins Obs., ser. 2, No. 6. abroad are making studies of both kinds. RTLE, M., AND HBWISH, A. 1950. Monthly Notices Roy. Astron. Soc. Lon- Much exciting work remains to be done and it don, vol. 110, p. 381. is very much to be hoped that additional VAN ISACKER, J. astronomers can be persuaded to take an 1953. Publ. Inst. Roy. Me'teorol. Belgique, ser. B, interest in the subject. No. 8.

Spectroscopy By Charlotte E. Moore

The outlook for future research in spec- spectral range. This is a taxing but vital troscopy is highly promising. The rocket program. spectroscopy of the solar spectrum, for example, Our knowledge of stellar spectra is far from extends our horizon to the shortwave region, complete. Any study of abundances requires and thus provides a far more complete picture correct identifications of the lines, together with of solar spectroscopy than ever before. The a knowledge of the excitation and ionization O vi solar emission lines and other observations of the separate atoms and . of equally far-reaching significance demonstrate P. W. Merrill has recently reported that in the the astrophysical importance of further study of Se star, R And, there are some 275 unidentified high-ionization spectra in the laboratory (John- absorption lines between 4000 A and 6900 A. son, Malitson, et al., 1955). Ultraviolet stand- J. Greenstein has observed several hundred ards of wavelength must be established in the unidentified lines in v Sgr. (see Merrill, 1956). shortwave region, and more term values must In the solar spectrum about 30 percent of the be worked out, particularly in spectra of the lines are still of unexplained origin. These and more abundant elements. Fe iv is an outstand- other similar examples demonstrate how much ing example of a neglected spectrum urgently work is yet to be done in spectroscopy. This needed by the astrophysicist. The coronal need is reflected in the persistent demand, spectrum, the nebular spectra, and now the among astronomers, for a new, large, revised spectra of distant galaxies, whose lines are multiplet table that extends from the far shifted further and further to longer waves, all ultraviolet to the extreme infrared. show that further study of ultraviolet spectra is These demands impose upon the laboratory extremely important. spectroscopist the task of observing and In the longwave region, beyond the photo- analyzing more atomic spectra, and also of graphic range, the development of suitable heat selecting spectra to be studied that anticipate detectors has extended the astrophysical , future astrophjsical needs. The use of modern as well as that of the laboratory, to the far sources, such as electrodeless discharges and infrared, and spectroscopy has infrared detectors, is one of the most promising extended it still further. With the astrophysi- approaches to the subject. Earlier descriptions cal range from Lyman alpha (1215 A) to the of atomic spectra are not only lacking in faint 21-cm wavelength of , our studies can lines but are incomplete, also, because of mask- no longer be limited to selected regions of the ing by oxide bands. The electrodeless discharge "visible" spectrum. has the advantage in atomic work that it can Among high-dispersion stellar spectra the be controlled in such a way as to suppress sunspot spectrum is one of the most challenging. molecular lines. More than 1,200 Fe i lines Hundreds of spot lines have not yet been have been identified in the solar spectrum by measured; many of them are doubtless of prediction from the known atomic levels. molecular origin. Furthermore, no homo- A suitable laboratory source of excitation would geneous set of sunspot spectrograms exists. doubtless excite these and many more faint During the coming sunspot maximum, high- lines. Kiess has already observed a number dispersion sunspot spectrograms should be of the predicted Fe i lines in selected regions. taken with the same spot, over the whole EdleVs (1953) observations of CA n in the infrared have resulted in the identification of 1 National Bureau of Standards, Washington, D. C. 13 14 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS four new multiplets among solar lines. For "Atomic Energy Levels." In solar and spot many familiar spectra there exists no complete spectra alone this subject is far from completed or homogeneous set of observations. Modern and promises interesting results in future equipment and techniques could be used to astrophysical research. good advantage in improving this situation. Spectroscopy embodies, also, the study of The spectra of the rare are the most line intensities, equivalent widths, and curves conspicuous group that needs study. Within of growth—all rewarding subjects for future the next 3 years it is planned to enlist the research, particularly with high-dispersion stel- services of all available experts in attacking lar spectra. Just as in the case of spectrum this difficult problem. Volume 4 of "Atomic analysis, measured laboratory intensities are Energy Levels" will be devoted entirely to required to increase our knowledge of abund- rare-earth spectra. This program should pro- ances of elements. This phase of the work vide the astrophysicist with a wealth of new must be carried on for years to come. data. In order to disentangle the intricacies The door is open and the opportunity is rich of these spectra, the assistance of theoretical for the student interested in spectra. The investigators to predict the positions of the more unusual examples of astrophysical identi- terms of various configurations is needed. fications—nebular lines, coronal lines, solar This problem to use of digital computers emission lines of high ionization, and technetium in solving large matrices. Here is a useful field lines in the S-stars—all challenge the astro- of activity in theoretical problems that parallels physicist and in themselves more than justify the laboratory and astrophysical research. active financialsuppor t of spec troscopic research The comments just given apply chiefly to in the laboratory, in theoretical fields, and in atomic spectra, but a wide vista appears, also, the observatory. in molecular spectroscopy. For example, the use of controlled flames as a source makes it References possible to bring out certain band spectra and EDL£N, B. suppress others. Precise laboratory measure- 1953. Accad. Naz. Lincei Conv. Sci. Fis. Mat. ments and analyses, particularly for diatomic Nat., vol. 11, p. 56. molecules of the more abundant elements, are JOHNSON, F. S.; MALITSON, H. H.; PUBCELL, J. D.; sadly needed. The writer is continually receiv- AND TOTJSET, R. ing requests for a critical compendium of 1955. Astron. Journ., vol. 60, p. 165. molecular data paralleling the volumes on MERRILL, P. W. 1956. Carnegie Inst. Washington, Publ. 610. Radial Velocities By R. M. Petrie1

The measurement of the line-of-sight compo- shift is always small and great care is necessary nent of stellar motion ranks among the early to avoid spurious instrumental shifts such as, achievements of stellar spectroscopy. The for example, those introduced by mechanical principle was announced by in the flexure, temperature variations, and faulty middle of the 19th century and was extended optical components. The development of re- and elaborated by a number of astronomers and liable instruments such as those at Potsdam, physicists soon after. The first applications Mount Hamilton, or may be studied were made by visual methods and are now of in an article by Eberhard (1933). The main historical interest only. The subject did not problems of design have now been solved, and attain full stature until the introduction of there exists a considerable number of instru- photographic methods during the closing years ments capable of yielding accurate radial of the century, but it then became at once a velocities. highly successful and valuable technique of Some problems of instrumentation remain as astronomy. Although radial-velocity work is a challenge to the student. The following are thus about 60 years old, it is still of prime im- examples of these. portance and promises to remain an indispens- Gratings.—The development of the "blazed" able tool in many branches of astronomy. grating of high efficiency opens a field for ap- The pursuit of this work is not successful with- plication to radial-velocity work in the visual out a considerable expenditure of patient region. Spectral types of F 5 and later may be effort, but the student may look to it to bring studied effectively in this way, which offers two substantial rewards for his labors. He may, important advantages. First, the spectra are with profit, reflect upon the prophetic words of simpler in the visual region, so that one can : "It would scarcely be possible, with- minimize the troublesome problem of using out the appearance of great exaggeration, to "blended" lines; and, second, many stars are attempt to sketch out even in broad outline the brighter in the visual than in the usual photo- many glorious achievements which doubtless lie graphic region. A successful grating spectro- before this method of research . . . . " graph has been constructed and used by Merrill In this article attention will be directed to (1931), and of course the powerful coud6 present activities and some probable future spectrographs of the largest telescopes employ needs. The development of our subject from gratings as the dispersive units. There is still the beginning, outlined above, is described in room, however, for experiments in the most the first part of 's "Stellar Motions," effective use of gratings in radial-velocity work published in 1913. with moderate dispersion. Long-focus instruments.—The great power of Spectrographs modern coude spectrographs lends itself well to After the introduction of photography the main radial-velocity investigations requiring the high- question to be answered was how accurately the est precision. It is not yet known with cer- spectrograph could record the positions of tainty whether the long-focus collimator and stellar absorption lines with reference to some wide-slit spectrograph fed by a combination of terrestrial comparison source. The Doppler three to five mirrors realizes the potential ac- curacy of the very high dispersion. We want 1 Director, Dominion Astrophysical Observatory, Victoria, British Columbia, Canada. to know, on the one hand, whether the wide 15 16 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS slit (relative to the stellar image) has limited of an objective prism free from distortion, and the accuracy because of the introduction of studies by Treanor (1948) and Schalen (1954) guiding errors; on the other hand, we want to have verified and extended this result. A know whether the multimirror system has pre- wholesale accumulation of radial velocities of served full "identity of source" between stellar faint stars is now in prospect; continued experi- and comparison spectra. Experimental work mentation and application are sure to bring to answer these questions needs to be under- rich rewards. Already the "mean errors" have taken as a preliminary to radial-velocity work been reduced to about ±6 km/sec, and this is of the highest precision. As an example of the sufficient for application to many problems in inherent possibilities, one may refer to ' galactic motions and structure. (1941) spectroscopic determination of the solar Image convertors.—Another promising de- . velopment is the image convertor, which is Comparison spectra.—Present practice gen- capable of high efficiency and may ultimately erally is to use an arc, or spark, in air, with the replace the use of photographic plates in re- most convenient metals being or titanium. cording stellar spectra (Argyle, 1955). Con- These sources are fairly satisfactory but are not siderable experimentation is now going on to entirely constant unless a number of precautions improve this technique and to apply it to stellar are taken. It is possible that a better source spectroscopy. The subject is well worth the can be found in some modification of a discharge attention of the student who wishes to make tube such as the hollow cathode type recently a contribution to work. described (Edlen, 1955). The greatest obstacle to the study of stellar Measurement motions through radial velocities is the slow Doppler shifts are, generally speaking, very rate at which spectrograms and measures ac- small and their successful measurement calls cumulate. The slit spectrograph, while accu- for painstaking care and experience. The rate, is very wasteful of starlight and ordinarily labor of measuring the stellar and comparison utilizes only a few percent of the incident light. lines on several hundred spectrograms is con- As a result we are, for example, almost totally siderable and often hampers the progress of a ignorant of the radial velocities of stars of program. 10 and fainter. Fifty The measurements were made originally years of work have yielded velocities for some with the aid of a direct-vision 15,000 stars, but this sample is much too small mounted above a traveling micrometer, or, in for galactic studies, and current progress is slow. the Hartmann comparator, spectra were A major effort and a new attack are obviously matched against suitable "standards," again required if the study of stellar dynamics is not with the aid of a microscope. The fatigue to be hampered by lack of observations. accompanying this process was lessened, some Fortunately there is now some hope of im- years ago, by the introduction of projection proving matters through two different current methods (Petrie, 1937; Redman, 1939). A developments. further improvement was made at Victoria Slitless spectrographs.—The idea of removing when a new type of projection instrument the spectrograph slit and collimator and thus (Petrie and Girling, 1948) allowed one to photographing with one exposure the spectra of measure the Doppler shift directly upon the of a field of stars is an old one. An account of projected image and, by virtually eliminating the early experiments with objective prisms is the "reduction," gave a substantial increase in given by Millman (1930). efficiency. Indeed, so useful is this projection Until recently the objective-prism work fell comparator that the measurement and reduc- short of its goal because of the presence of field tion of spectrograms is no longer a serious distortions and instrumental errors exceeding obstacle to the completion of a heavy radial- many times the Doppler shift. However a velocity program. series of important experiments by Fehrenbach The obvious advance to be made now is the (1947, 1948) have resulted in the development development of an automatic machine that NO. 1 NEW HORIZONS IN ASTRONOMY 17 will remove the subjective errors and allow an 1952). The problems must be faced anew impersonal definition of the precise position of with each new spectrograph and vigilance is a stellar line relative to the comparison spectrum. necessary to ensure that the deduced radial A machine of this sort has been built by Johnson velocities are, in fact, motions and not spurious (1949), and another of different type by Hos- displacements arising from uncertain, or in- sack (1953). In spite of these ingenious correct, wavelengths. machines there is still room for experimenta- In spite of the work already done on this tion and development. The technical aids subject, more is required. We do not, for already exist and it is fairly certain that the example, have a satisfactory set of standard present generation is the last that will be re- wavelengths for O-type spectra and cannot, quired to measure Doppler shifts with hand therefore, be entirely certain of their motions and eye. A real triumph awaits some student in star "associations" nor of their apparent equipped with a sound knowledge of modern " shift" as members of galactic clusters. electronic aids. And, at the other end of the temperature se- A word of caution may be permitted here. quence, we lack effective wavelengths for low- An automatic machine does only half the job dispersion spectra of N stars. This is of some if it merely presents one with the measured importance since we need to know the positions of stellar and comparison lines. The motions of a larger number of these (apparently) calculation of a radial velocity from the engine faint objects. readings requires nearly as much time as the original measurement. The ideal machine, Applications therefore, will measure the shift of the spectral The application of radial velocities leads us into line from its zero-velocity position and so allow the subjects of stellar dynamics, galactic struc- the rapid deduction of the radial velocity. ture, binary stars, etc. These topics are being discussed elsewhere in this compilation Wavelengths and the writer will therefore merely point out The radial velocities deduced from the measured some of the possibilities awaiting the astronomer positions of spectral lines are, of course, very who devotes himself to radial velocities. The sensitive to the accuracy of our wavelengths. worker who undertakes to supply the funda- We recall that in the photographic region an mental observations of stellar motions can be error of 0.1 A is equivalent to 7 km/sec. A sure that his contribution is of permanent value firm knowledge of wavelengths is a prerequisite to astronomy. He must therefore be patient to a significant knowledge of motions. and content to see his work gradually increase The practical difficulties are considerable. as his programs and plans are completed. Radial velocity programs usually force one to The general problems of stellar motion re- employ a spectroscope of low dispersion and main with us: solar motion, star streaming, purity; with such an instrument all spectral galactic rotation, and mean velocities in differ- features in types later than A 0 are blends of ent parts of the galaxy. The need is always two or more lines. The effective zero-velocity for more observations and observations of more wavelength of a feature depends upon the distant stars. For a general description of relative intensities of the blending lines, the present-day need one should refer to the report spectrographic resolution, and the amount of of the International Astronomical Union Sym- and other line-widening in- posium on "Co-ordination of Galactic Re- fluences. Even in early-type spectra difficulties search" (Blaauw, 1955). sometimes arise because of blends, stellar The measurement of galactic rotation is of rotation, and Stark effect. prime importance. We require a more accurate At Victoria a good deal of attention has been knowledge of Oort's constant A in the neigh- devoted to the matter of effective wavelengths borhood of the sun. Current programs of B and it is believed that satisfactory methods and stars at the Radcliffe and Dominion Astro- values have been established (Petrie, 1946, physical Observatories will supply about 1,000 1947, 1948, 1953: McDonald, 1948; , new velocities, mostly within a distance of 18 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 1,500 of the sun. These, together with clusters and associations is sure to yield im- existing observations, should give a good portant and exciting information. determination of the coefficient of differential Stellar atmospheres.—Recent work on super- rotation in our part of the galaxy. giant stars and special eclipsing stars, such as The more distant regions are quite im- and f Aurigae, shows how radial- perfectly studied. With the recognition of velocity measures reveal the existence of atmos- spiral arms about 2,000 parsecs away, it pheric motions. This is a little-explored field becomes an exciting task to discover what and requires study with highest possible disper- velocity effects exist at the edges of the arms sion. Comparisons between giant and main- and between them. This means we must sequence stars of the later spectral types will observe large numbers of spiral-arm objects— be valuable in promoting knowledge of the early B stars, Cepheids, and supergiants of stability of stellar atmospheres. types B 8 to A 3. At the same time we must Variable stars.—Experience has shown that measure the speeds of the interarm population variable stars should be studied with simul- which will probably consist of stars of Popula- taneous spectroscopic and photometric meas- tion II together with a number of stars that ures. The observations of Sanford (1952) and have wandered out of the spiral arms. Ob- Abt and Sanford (1954) on the servations of the early-type stars will also and of Struve (1954) on BW reveal the motions of the interstellar matter. Vulpeculae are good examples. The work to A problem of considerable interest is the be done here is limitless, considering the great stellar velocity distribution perpendicular to the range of stellar variation. To be most fruitful, . This may be found from radial the highest possible dispersion must be used. velocities of stars near the galactic poles. The Binary stars.—The problems of binary stars observation of the brighter stars of types A 0-F 5 are legion, but perhaps no other field offers more within 10° of the Galactic Pole is now near- challenges calculated to advance our knowledge ing completion at Victoria. A companion program of stars. of fainter stars is proceeding at Mount Hamilton. Visual binaries should be observed continu- In this branch of stellar motion work much re- ously so that, ultimately, the combination of mains to be done. Extensive observations are spectroscopic and micrometric data will tell us required among the later-type stars of the disk the dimensions and of the compo- population and, on the other hand, of the faint nents. Similarly, eclipsing binaries should be stars of (probably) Population II. From observed assiduously, especially those with the velocities of the disk population we can determinate light elements and those showing determine the density of matter in the galactic the spectra of both components. Our knowl- plane region, in the neighborhood of the sun. edge of stellar , radii, and luminosities, The above proposals involve observations of as well as further extensions and verification of many stars and are rather long-term projects, the mass- relation, must come from but much may be done in other directions this work. Radial velocity observers will find where shorter programs will bring valuable it a rewarding task. results. Some of these are suggested very The bewildering array of spectroscopic anom- briefly in the following paragraphs. alies exhibited by the close binaries has been Star clusters and associations.—These fascinat- brought to our attention by Struve (1950). ing groups are of first importance in studies of The observation and interpretation of radial- stellar luminosity and evolution. Radial-veloc- velocity curves are an endless challenge. One suspects that the most potent observational ity observations are needed to verify cluster aids will be required to advance our under- membership and to calculate space motions. standing of these spectacular phenomena. Careful and extensive observations of the Stellar satellites.—The discovery of planetary "aggregates" and expanding associations must systems is probably just within the powers of be made before we understand the kinematical the most precise radial-velocity measures that properties of these groups. The study of can be made with existing equipment. This NEW HORIZONS IN ASTRONOMY 19 has been shown by Struve (1952), who proposed MERRILL, P. W. that a search be made. It may be that a major 1931. Astrophys. Journ., vol. 74, p. 188. astronomical discovery will result from the MlLLMAN, P. 1930. Circ. Harvard Astron. Obs., No. 357. meticulous observation of a number of our PETRIE, R. M. nearby and bright stars. 1937. Journ. Roy. Astron. Soc. Canada, vol. 31, p. 289. References 1946. Journ. Roy. Astron. Soc. Canada, vol. 40, p. 325. ABT, H. A. 1954. Astrophys. Journ., Supplement, No. 3. 1947. Journ. Roy. Astron. Soc. Canada, vol. 41, p. 311. ADAMS, W. S. 1941. Astrophys. Journ., vol. 93, p. 11. 1948. Journ. Roy. Astron. Soc. Canada, vol. 42, p. 213. ARGYLE, P. E. 1955. Journ. Roy. Astron. Soc. Canada, vol. 49, 1953. Publ. Dominion Astrophys. Obs., vol. 9, p. 19. p. 297. BLAATJW, A. PETRIE, R. M., AND GIRLING, S. S. 1955. Int. Astron. Union Symposium No. 1. 1948. Journ. Roy. Astron. Soc. Canada, vol. 42, Cambridge University Press. p. 226. EBEBHARD, G. REDMAN, R. O. 1933. Handbuch der Astrophysik, vol. 1, pt. 1, 1939. Monthly Notices Roy. Astron. Soc. Lon p. 299. don, vol. 99, p. 686. EDLEN, M. B. SANFORD, R. F. 1955. In Draft reports of the International Astro- nomical Union Dublin meeting, August- 1952. Astrophys. Journ., vol. 116, p. 331. September 1955, p. 104. Cambridge SCHALEN, C. University Press. 1954. Arkiv fur Astronomi, vol. 1, p. 545. FEHRENBACH, C. STRUVE, O. 1947. Ann. d'Astrophys., Paris, vol. 10, p. 306. 1950. , chap. 3. Princeton Uni\ 1948. Ann. d'Astrophys., Paris, vol. 11, p. 35. Press. HOSSACK, W. R. 1952. Observatory, vol. 72, p. 199. 1953. Journ. Roy. Astron. Soc. Canada, vol. 47, 1954. Publ. Astron. Soc. Pacific, vol. 66, p. 329. p. 195. TREANOR, P. J. JOHNSON, H. L. 1948. Monthly Notices Roy. Astron. Soc. Lon- 1949. Astron. Journ., vol. 54, p. 190. don, vol. 108, p. 189. MCDONALD, J. K. WRIGHT, K. O. 1948. Journ. Roy. Astron. Soc. Canada, vol. 42, 1952. Publ. Dominion Astrophys. Obs., vol. 9, p. 220. p. 167.

Astrometry By K. Aa. Strand*

Astrometry deals with the space-time be- complexity the problems range from setting up havior of the stars, and thus belongs to the an accurate reference system of stars covering classical field of astronomical studies. While the entire sky to determining the parallax of an the early investigations were directed mainly individual star. towards establishing a suitable frame of refer- A series of conferences held during the past ence for determining the complex motions of the two years at Groningen, Evanston, Pulkovo, planets, studies of the positions and motions of and Brussels has resulted in general agreement individual stars, as well as of the various stellar that, to solve many present-day problems, as- systems, gradually developed. tronomers must undertake an international pro- If one examines the observational data accu- gram of meridian circle observations. The pro- mulated in astrometry, the record is indeed im- gram, starting in 1956, will consist of observa- pressive. During the past 150 years more than tions of nearly 21,000 stars brighter than 9.1 370 catalogs of meridian observations of stars magnitude north of —5°. The total have been published. We have monumental number of observations, including those re- astrographic catalogs like the Carte du Ciel, the quired of the fundamental stars, will amount to Astronomische Gesellschaft catalog (AGK2), 260,000, since each star is to be observed 10 and the Yale zone catalogs, to mention only the times. Twelve observatories will participate, largest. The Yale General Catalogue of Stellar and by using timesaving devices in both the lists the distances for nearly 6,000 observational and reduction phases of the pro- stars measured since the beginning of this cen- gram, they expect to complete it in 5 years. tury, and the and American astronomers should note with some catalogs collect more than half a million microm- pride that the U. S. Naval Observatory will eter measures, the result of a century's work. carry out a substantial part of the observations Considering this impressive observational and will also supply, on request, the reductions record, it is not surprising that a person who to the day and to the equinox 1950 for all the has not worked in astrometry usually assumes, stars observed. The U. S. Naval Observatory mistakenly, that most of the work has been will also compile the final catalog of positions for completed. Our knowledge of such properties these stars. of the stars as their masses, luminosities, and What information will this large undertaking motions depends directly or indirectly upon the give us ? It will provide a reference system of available astrometric data, and any improve- stars with greater density and precision than ments to be made in these parameters will de- any previous system; and it will make possible pend upon further work in the field. the solution of problems that could not be What future work is needed in astrometry to solved without it, some of which may be sum- broaden our general knowledge of astronomy? marized: A whole series of problems confronts us. They 1. The program will provide a reference sys- range in size from those that can be investigated tem for the AGK3, the second photographic ob- only by close cooperation among many observ- servation of the stars in the Astronomische Ge- atories to those that can be solved by a few sellschaft Katalogs, which when compared with astronomers in a^single observatory; and in the AGK2 will to accurate proper motions for approximately 180,000 stars brighter than '^Director, Dearborn Observatory, Northwestern University, Evans- ton, HI. 11.5 . Dr. Vyssotsky 21 381014—56 3 22 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS of the Leander McCormick Observatory and his until 1967. How soon after that date the pro- associates will supply spectra for approximately gram will be completed depends upon the meth- 60,000 stars not previously classified. ods of measurement and of reduction. It is 2. About 11,500 of the stars in the reference quite clear, however, that the program is of such system will be used in the Russian program on magnitude that it requires the utilization of the stellar proper motions relative to extragalactic latest developments in automatic or semiauto- nebulae. matic equipment to complete it without pro- 3. The proper motions determined from the longed delay. AGK2 and AGK3 will provide the systematic Since the value of the Lick program would corrections that are needed to make use of be much enhanced if it could be extended to many of the older star catalogs. the southern sky which is not observable at 4. The results from the AGK3 will provide the , serious consideration the necessary data to reduce to absolute the should be given to extending the program to proper motions obtained from a reobservation of the Southern Hemisphere. In view of the the Carte du Ciel plates. small number of observatories of the 5. The stars from the AGK3 could be used as equator and the magnitude of the problems reference stars for the Lick Observatory pro- that already confront them, this program prob- gram, which expects to determine proper mo- ably should be undertaken by an observatory tions of stars with respect to extragalactic in the . It remains to objects. be investigated whether a telescope identical It is gratifying to note that it is still possible with the Lick 20-inch Carnegie astrograph to undertake large-scale programs requiring should be constructed or whether, as suggested careful international cooperation, for which as- by Dr. Luyten at the Meeting of the Inter- tronomy is conspicuous among the sciences. national Astronomical Union in Dublin, a However, it should be observed that the burden limited number of plates taken with the of American participation in the AGK3 rests telescope would serve the purpose. solely upon the U. S. Naval Observatory. Photographic astrometry with long-focus Because the Lick program of deriving proper refractors is traditionally a field in which motions of stars with respect to extragalactic American astronomers have held the leader- objects will yield important scientific results, it ship; here, also, we face a series of problems should receive the support necessary for its com- which must be solved for further progress in pletion. The Lick program will provide abso- astronomy. lute proper motions of stars down to the 17.5 It has long been known that the trigonometric photographic magnitude, resulting in data that parallaxes are affected by systematic errors; can be expected to yield corrections to the whether further statistical investigations of the adopted values of precession and lead to new existing observations will provide additional determinations of solar motion and galactic ro- information on this problem remains doubtful. tation. The most distant stars in the Lick The systematic errors have, of course, the most survey will have distances of more than 2 kilo- serious effect upon the small parallaxes; but parsecs and thus reach beyond the region of the the absolute magnitudes of and sub- maximum circular velocity, as indicated by the dwarfs, which are scarce in the solar neighbor- Oort model. This study should settle the pres- hood, are therefore derived from a relatively ent discrepancy between the data on galactic small number of small parallaxes and thus are rotation, as obtained from proper motions and also affected by large systematic errors. To from radial velocities. improve this situation it is essential that many If the Lick survey is to attain the needed of the small parallaxes be repeated, and that minimum accuracy, a time interval of approxi- the observational series for the individual mately 20 years must elapse between first- and stars be increased from the 20 plates or less second- plates. This means that the com- used in the very early work to 50 plates or plete repetition of the Lick plates will not begin even more. These determinations would add NEW HORIZONS IN ASTRONOMY 23 to our knowledge of the of task; nevertheless, many of them must be ob- the stars just mentioned, and would also serve served if we wish reliable data on such questions as important statistical material for further as statistical distribution of the orbital elements investigation of the systematic errors in trig- in binaries and combination of spectra in phys- onometric parallaxes. ical pairs. These are only two of the questions The stars of low luminosity in the solar that have a bearing on stellar evolution and neighborhood, discovered in the Luyten proper which we cannot answer by studying the small motion survey, are another group for which group of known binaries within 20 parsecs of we need determinations of trigonometric paral- the sun. laxes. The great majority of these stars, Problems of membership and internal motions late-type M-dwarfs, subdwarfs, and white in clusters will become even more important as dwarfs, are considerably fainter than any we the time interval between early- and late-epoch can observe with the telescopes now being used plates increases the accuracy. The early Carte for parallax work. These telescopes cannot du Ciel plates and those taken with the large reach stars much fainter than the 12 th mag- refractors shortly after the beginning of this nitude, but most of the known white dwarfs century are especially important in that are fainter than the 14th. For this reason, respect. the commission on parallaxes and proper One of the most recent additions to astrom- motions of the International Astronomical etry, the study of the age of the O-B associ- Union recommended at the recent meeting in ations from their expansion, deserves high pri- Dublin that observatories possessing large ority, and it is to be hoped that the large merid- reflectors seriously consider a program of ian programs in connection with the AGK3 will determination of trigonometric parallaxes of still permit the stars selected for this study to faint stars. be observed with utmost precision. A similar Other aspects of the use of long-focus re- program on Cepheid variables, to trace the po- fractors are the positional study of the stars sitions of the spiral arms in our galactic system, within 10 parsecs by an extended series of also deserves special attention. plates to discover perturbations for "single Lack of manpower to cany out the many im- stars," and the subsequent orbital analysis portant problems just outlined remains a prob- of these perturbations. Quite recently such a lem. We may summarize briefly what we re- study led to the visual discovery of a binary quire in instrumental developments to solve this () in which the companion has only question. The staff of the U. S. Naval Observ- 8 percent of the sun's mass. atory has, over a period of years, introduced When we examine the relation of astrometry timesaving devices that have increased the to the double stars, we find a series of problems speed as well as the accuracy with which the which deserve the attention of modern astron- stars are observed, and because of the use of omy. For the double stars within 20 parsecs, IBM machines the reductions of the observa- it is possible to derive stellar masses with an tions are no longer lagging behind as in the past. accuracy of 25 percent or better; to attain this In photographic astrometry the use of auto- accuracy, however, visual and, whenever pos- matic devices in guiding, plate transport, and sible, photographic observations of the relative timing has resulted in increased accuracy in motions are needed as well as extended series of certain problems, but the real problem here is to plates from which the mass ratios and parallaxes can be derived with great precision. Much in- develop equipment that will effectively shorten formation in regard to stellar masses has been the time required for measurement and reduc- obtained in this way over the past 15 years, but tion of the photographic plates. One measuring much more is needed. This group of binaries machine, allowing semiautomatic measurements constitutes only about l/1000th of the approxi- and operated in connection with IBM machines, mately 40,000 known binary systems. To ob- is already in use. Since the operating costs of serve them all would clearly be an impossible this system are prohibitively high for most 24 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.! American observatories, it seems desirable to tention. However, those outlined show that in look into other recent technical developments spite of, or rather because of, the huge amount which might aid in reducing the time spent at of work done in the past, there are still many the measuring machine without requiring a interesting and important problems on which highly expensive computing system. progress in astronomy depends; therefore, they In this brief report it is not possible to men- deserve the attention of the active research tion all phases of astrometry that deserve at- workers in astronomy. Photoelectric Astronomy By A. E. Whitford l The past two decades have seen increasing The final in the information chain, the use of photoelectric methods in all branches of radiation detector at the focus of the telescope, astronomy where quantitative measurement of is therefore the element which must be scruti- radiation is important. This trend seems nized for possible lack of efficiency. If the destined to continue. In the years ahead the efficiency is low, improvements can be counted photoelectric cathode may become the most as equivalent to increasing the aperture of widely used radiation detector in the astronom- existing telescopes. ical tool kit; indeed, it may supplant the photo- The various practical detectors in actual use graphic plate in the same way that the plate can be evaluated by comparing them with a supplanted the human eye 50 to 60 years ago. hypothetical ideal detector that recorded the A brief review of the links in the information arrival of each quantum, and did it with 100- chain along which knowledge about the external percent efficiency for all accessible wavelengths universe must pass will show the reasons for of the astrophysical spectrum. The photo- believing that photoelectric methods may graphic plate cannot be given a very high mark assume a dominant place. The light beam from in such a comparison. Since it is a saturable, a distant luminous source, which carries all the nonlinear device, no unique quantum efficiency information the astronomer will ever have can be assigned. But in the range of densities about that object, arrives at the surface of the favorable to contrast-judgments by the human earth after some losses in the interstellar eye, only about 1 quantum in 1,000 is recorded medium and in the earth's atmosphere. The by producing a developable plate grain. Its atmosphere adds an unwanted background ability to integrate a long exposure is the reason glow, and, at times of bad seeing, diffuses the for its superiority over the eye in detecting beam somewhat. The potential information faint stars and recording low ; in the light beam suffers a loss and degradation the eye makes its decision in about % second. before it enters the astronomer's instrument Within that % second the eye, as shown by Rose that are for the most part beyond his control. (1948), is a better detector than the photo- Some day the astronomer may get above the graphic plate. The eye cannot, however, atmosphere, but he will not get beyond the deliver the plate's objective record of thousands . Meanwhile, he must of details or the extremely precise positional make sure that his earthbound instruments do information needed for astrometry. the best possible job of extracting all of the The cesium-antimony cathode is the one most information remaining in the light beam. widely used in astronomical photometry. It The astronomical telescope has long since has a useful range of 3100A to 6000A. The reached a state of perfection where the tran- average quantum efficiency at the peak near quility of the atmosphere, rather than the 4400A is 12 percent, and exceptional tubes go quality of the optics, determines the image as high as 25 percent. There is thus a superi- quality. The astronomer may, in theory, ority of a factor of the order of 100 over the gather more information by increasing the photographic plate in the initial information aperture, but the engineering difficulties and the made available in a given time from a given cost both rise steeply for any considerable in- angular area on the sky. As presently used, crease in size over that of existing telescopes. only one chosen element or area may be meas- Wasbborn Observatory, University of Wisconsin, Madison, WJs ured at a time; if photometric information on 25 26 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 many objects in one field is needed the photo- Photographic interpolation between photo- graphic plate is a great timesaver. But the electrically established standards in the field superior quantum efficiency of the photo- of a star group such as a electric cathode is the basic reason why it (e. g., as by Arp, Baum, and Sandage, 1953) seems destined to play an increasing role in will continue to be used where photoelectric measurements of astronomical radiation. measurements of one faint star at a time would In order for the favorable quantum yield to be too time-consuming. But for transfer of be useful, there must be no instrumental back- standards from Selected Areas, photoelectric ground or amplifier noise large enough to methods appear to be needed for good accuracy. degrade appreciably the information contained The time is rapidly approaching when every in the released photoelectrons. Existing tech- observatory which does any photometry at niques succeed quite well in approaching this all will think of the plateholder and the photo- ideal and further efforts can be expected to electric photometer as routine interchangeable produce only small improvement. There is not attachments to the telescope, each to be used much to choose between any of the three as occasion demands. currently used methods of registering the output Since photoelectric methods require the use of a phototube: current measurement with a of the telescope for only one object at a time, load resistor, charge integration with a storage and the reductions are also quite time-consum- condenser, and charge integration by pulse ing, magnitude measurements have thus far counting. At the lowest intensities integration been restricted to specific classes of stars, such methods will probably find increasing use. as B-stars and Cepheid variables, and to the The future uses of the photoelectric method establishment of standards in Selected Areas in astronomical photometry are of course not or star groups under special study. A deter- entirely predictable. The only really safe mined effort might provide instrumental and prediction is that new and ingenious applica- computing aids which would make mass pro- tions of the method, not now foreseen, will duction of magnitudes more economical. These surely be devised. A projection of the pres- ently successful techniques suggests activity might include better coordinate mechanisms in three main areas. In general, these exploit for rapid identification and centering of stars, the basic advantages of the photoelectric automatic subtraction of sky readings, auto- process: high quantum yield, linear response, matic computation of and correction and wide spectral range. of the star signal for it, a logarithmic amplifier to give direct reading magnitudes, and direct Intensity comparisons: Magnitude systems processing of telescope output onto data storage The range of intensity between the sun and records such as punched cards. Then it might the faintest stars yet detected is of the order not be too enormous a task to do what Irwin of 1020. Important astronomical conclusions, (1955) suggests, and undertake to provide such as the cosmic distance scale, depend on accurate photometric data for all the stars, the precision with which this range is covered. say, in the Catalogue. A massive The superior accuracy of the photoelectric attack by one observatory could accomplish method rests in part on its higher quantum it; it would be a public-spirited endeavor, to yield, in part on its linearity. Elimination of be viewed as a repayment of the scientific debt the step calibration needed by the nonlinear owed to previous generations for the patient detectors such as the photographic plate not determination and compilation of star positions, only saves a great deal of time but also removes spectral types, radial velocities, and proper a fertile source of accumulated error. An motions. It would at least correct the present intensity range of 10* to 106 can be covered incongruous situation where the simplest and with one phototube and a single optical system. most fundamental datum about a star, its The task of calibrating magnitude systems brightness, is often the least reliable entry in is already largely a photoelectric province. the catalog. NEW HORIZONS IN ASTRONOMY 27

Color systems and spectrophotometry parison of these "normal" stars with Type II The advantage of the photoelectric cathode stars and peculiar stars by multicolor or scan- in studies stems partly from its wide ning methods should be most informative. Ex- spectral range, partly from its linearity. The amples would be subdwarfs, metallic line stars, simple ratio of the intensity through two color and pulsating stars of all kinds. Differences of filters gives a good without the need composition or structure may be expected to of establishing two magnitude scales as in the leave clues in the continuous spectrum; accu- photographic method. Because the photo- rate photometry will be needed. electric color index comes from a differential Composite sources are another fruitful field measurement, the internal accuracy is higher for multicolor studies. Globular clusters and than for the magnitudes themselves. A two- galaxies are examples. The resolving power in color index provides the minimum amount of distinguishing between various syntheses of the information about a star. The wide spectral over-all energy distribution from hypothetical range of a photoelectric cathode permits three mixtures of known stellar types may never be or even five narrow-range with a cesium- very high. But a combination of spectra, of antimony cathode. The range may be extended counts of brightest stars of identifiable types, to 10,000A or more with a cesium-oxide cathode and of over-all energy distribution may lead to and the number of color filters extended to six fairly definite conclusions. The diffuse nature or seven. The next step is to scan the whole of these objects precludes anything but low range with an exit slit on a spectrograph. The resolution spectrophotometry by the scanning observing time per star goes up because of the method, and for faint objects multicolor filter limitation that only one element of the spectrum photometry may be the best that can be done. can be observed at one time. High-resolution Their low contrast with the sky can be handled photoelectric scans of short portions of the solar cleanly by the differential subtraction methods spectrum have been made, and similar scans will so easily applied in photoelectric photometry; undoubtedly be attempted for the brightest and photoelectric accuracy will be needed. stars. Over the whole range, however, a re- Photometry of distant galaxies for diameter and solving power of about 350 gives a reasonable magnitude (as suggested by Baum, 1953) and observing time per star and adequate delineation for energy distribution (Whitford, 1953) will of the main features of the continuous spectrum. provide evidence of the state of affairs at a much earlier era in the evolution of the universe, The color of a star is already recognized as a and will thus be of interest to those concerned necessary part of a photometric determination with . of its distance in order to evaluate the inter- stellar reddening. For early-type stars, at least, Differential measurements three colors on the U-B-V system of Johnson and Morgan can give a measure of reddening Light curves of eclipsing binaries provided when the spectrum is not known. Extension of one of the early demonstrations of the precision this powerful method to later-type stars, and attainable by photoelectric methods. The to the longer wavelength portions of the spec- accuracy was due partly to the linear response, trum where the interstellar absorption is less, and partly to the differential comparison would be a great help in galactic and extra- of variable and comparison star, thus minimiz- galactic distance determinations. ing variable factors in sky and apparatus. The fundamental information about stars Determination of the absolute energy distri- derivable from precision light curves of eclipsing bution in normal Type I stars will be worthy of variables will continue to offer attractive pos- considerably more work over the whole range sibilities to the photoelectric observer with a that can be covered with existing photoelectric small- or medium-sized telescope. For recon- cathodes. From such studies can come better naissance purposes other methods will continue model atmospheres, and from these, in turn, to be of value, but the precision needed for higher accuracy in quantitative studies of nu- deriving properties like limb darkening requires clear abundances from line strengths. Com- photoelectric accuracy. 2S SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 Linear response is a great aid in all manner Some reduction in observing time can be of differential measurements because simple realized by using two or more phototubes, fed subtraction is possible. Alternation of fields by beam-splitters, or by several different out- once a minute or oftener is one way of doing puts of a spectrograph. The most exciting this. The flicker or a. c. method is another, development in photoelectric astronomy, how- and it may give a better elimination of unwanted ever, would be an image tube that would extraneous factors. One example is the meas- preserve the high quantum yield of a cathode, urement of very faint stars or galaxies which while responding simultaneously yet separately add only a few percent to the sky radiation to all the bits of information in the various that comes through the smallest usable focal- area elements of the image plane. Such a plane diaphragm. Nebulosity of low-surface development seems to be a definite possibility. brightness would be another example. Ex- The techniques being explored, as explained tension of flicker methods to this type of by Baum, have been reported by Struve measurement would be a logical development. (1955). Work is actively in progress in Wherever it is possible to "put a tag" on France, England, and the United States. the wanted signal, it can usually be lifted out In its simplest form, that first introduced by of an overwhelming background by differential Lallemand, the image tube consists of an elec- methods, often involving a. c. amplification. tron lens system for focusing the accelerated One example would be the of photoelectrons in an imago on a photographic starlight (Hiltner, 1949, and Hall, 1950); plate at the end of the tube opposite to the another would be measurement of solar magnetic cathode. Every electron sensitizes a plate fields (Babcock, 1953) where the modulating grain, thus preserving the high quantum effi- element again involves polarization. These ciency of the cathode. The same number of flicker methods have been highly developed in developable grains is achieved in a much shorter radio astronomy where the signal may be 100 time than would be needed in direct exposure of to 1,000 times smaller than the unavoidable the plate to light. A fine-grain plate may be noise. Further use of the technique in photo- used without loss of speed, thus increasing the electric astronomy seems likely. number of bits of information in a given image area before saturation sets in. As yet, exposure Possible advances in technique times adequate for showing an intensity level If photoelectric astronomy continues under as low as the have not been possible, the restriction of one picture element at a time, and the anticipated gain in threshold detection what developments might extend its range of level has therefore not been realized. usefulness ? At the head of the list would come A proposed alternative scheme involves a more sensitive cathode for the red and infra- storage of the released charge on a mosaic, as red where the quantum efficiency is now 20 in the image orthicon. This is later wiped off times poorer than for the blue. This region by a scanning electron beam, and the effect of the spectrum may assume increasing im- on the beam is the signal from that element. portance for two reasons: Interstellar absorp- While a great deal of hard work remains, tion is greatly reduced, and discovery level there appear to be no insuperable difficulties or penetrating power correspondingly increased; in the way of final success. If such a tube and at very large red shifts, extragalactic systems comes, what will be the first applications? carry not only their energy maximum but also Reduction of prohibitively long exposure times all the parts of the spectrum ordinarily studied in photographic spectroscopy would be one of into the red region. A high-efficiency photo- the first. All-night exposures would be re- electric detector would greatly facilitate in- duced to a few minutes. Spectra of faint vestigations in these areas. A new tri-alkali galaxies, so important in considering cos- cathode (Sommer, 1955) promises considerable mological questions, would become fairly rou- improvement in the red as far as 7500A. tine. On direct pictures of the sky a detection Germanium phototransistors may be developed limit 2.5 magnitudes fainter should result, sufficiently to fill the need as far as 16,O0OA. because 100 times as many countable events NEW HORIZONS IN ASTRONOMY 29 go into the decision as to whether one picture low efficiency to something approaching the element (say the size of the seeing tremor disk) ideal state where every quantum is counted. has received a statistically significant excess It would then resemble the present state of of photons from a star during the exposure optics as an astronomical tool: no revolutionary •time. improvements could be expected. But even While high quantum efficiency and accurate then, as now, astronomical advances will imaging can be realized in an image tube, it depend quite as much on the skill and ingenuity may be doubted whether the linearity and other of the astronomer as on the tools he has at photometric advantages of a single phototube his disposal. can be preserved through the whole process. Nevertheless a successful image tube would References bring many observations, now utterly unat- ABP, H. C, BATTM, W. A., AND SAN DAG E, A. R. tainable, into the realm of the possible. In 1953. Astron. Journ., vol. 58, p. 4. the rich harvest that would follow its intro- BABCOCK, H. W. 1953. Astrophys. Journ., vol. 118, p. 387. duction, there would almost certainly be some BAUM, W. A. unexpected discoveries. 1953. Astron. Journ., vol. 58, p. 211 (abstract). HALL, J. S. Conclusion 1950. Publ. U. S. Naval Obs., No. 17, pt. 1. As one looks ahead, it seems clear that the HlLTNER, W. A. 1949. Astrophys. Journ., vol. 109, p. 471. next decade will see increasing use of the photo- IBWIN, J. B. electric process in astronomical photometry. 1955. Sky and Telescope, vol. 14, p. 132. Wherever two quantities of radiation have to ROSE, A. be compared, the photoelectric method will be 1948. In Marton, ed., Advances in electronics, the preferred method if time permits. Should ch. 3. Academic Press, New York. the image tube come into common use, a revolu- SOMMER, A. H. tion in astronomical observing methods may 1955. Rev. Sci. Instr., vol. 26, p. 725. STBUVE, O. ensue. It is not impossible that the whole 1955. Sky and Telescope, vol. 14, p. 224. field of the detection and measurement of WHITFOBD, A. E. radiation may pass from the present state of 1953. Astron. Journ., vol. 58, p. 49 (abstract).

881014—56 4

Interferometers in Radio Astronomy By B. F. Burke * and M. A. Tuve x

Only a faint glimmer of an idea yet exists as would still cover an area 5 minutes of arc in to how astronomical objects can radiate, appar- diameter. Arrays of great size and interfer- ently with a fair degree of conversion efficiency, ometer spacings of as much as a are simple electromagnetic waves some tens of centimeters by comparison to such a dish. or even meters in length. The discovery phase It should not be surprising if a new branch of is probably over; only a few such objects are science raises difficulties in its early days, and known, but these are enough to face us un- radio astronomy has not been disappointing in equivocally with the problem. "Plasma oscil- this respect. In the past few years, this young lation" is an easy phrase, and where the electron branch of astronomy has provided a large densities are high, as in the solar atmosphere, variety of both physical and astronomical these words seem to offer a plausible explana- problems. A striking feature of the earliest tion, but when one is confronted with the investigations has been that most of the in- problems of boundary conditions and the mech- trinsically strong sources of radio noise are anism for escape of the radiation to free space relatively insignificant optical objects, a circum- the plausibility disappears. Additional difficul- stance that adds to the difficulty of examining ties arise for the low apparent electron densities in detail their physical behavior. The handful in gas . Strong and perhaps turbulent of optically identified discrete sources all exhibit magnetic fields combined with violent stream- one feature in common, however, which pro- ing motions, perhaps even with shock waves, vides some indication of the nature (and diffi- may resuscitate the hypothesis, but theory has culty) of the problem, for in all cases we seem yet to rescue the observer in this field. to be dealing with highly turbulent gas whose Progress toward solution of this problem may constituents are in high states of ionization. still lie in the hands of the observer, however, Other general or special requirements may as there exist only a dozen or so sources which exist, and they will have an important influence have been specifically located and for which on any theory of how such a tenuous but "color" information (in the radio spectrum), for violently agitated gas becomes a relatively effi- example, has begun to be available. When at cient emitter of radio noise. For example, the least several hundred sources, thermal and non- polarization of the light from the thermal in their radiation characteristics, have (pi. 1) provides evidence for the existence of a been cataloged and their radiation properties and large quantities of highly described in some detail, the puzzles may di- energetic electrons, which also could be efficient minish. Large arrays of antennas, mainly radiators in the radio spectrum, but whether dipoles or helices, connected and used as inter- this is a general characteristic of all radio ferometric devices, are the principal technical sources or a peculiarity of this one object only basis for obtaining observations at the longer the future will tell. It appears certain, how- wavelengths. Parabolic reflectors of any rea- ever, that the necessary highly turbulent gas sonable size cannot be expected to give the can be produced in a variety of ways. In our necessary precision of> location in the sky; at own galaxy, cataclysms such as 20-cm wavelength even a 500-foot parabola explosions are known means of providing a turbulent medium that is capable of strong 1 Department of Terrestrial Magnetism, Carnegie Institution of radio emission; but although all discrete galactic Washington, Washington, D. C. 31 32 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 sources may possibly be supernova remnants, All the sources we have been considering are many of them may also represent an entirely characterized by their nonthermal nature, since new class of objects. nonequilibrium processes appear to be involved. The extragalactic sources, which are intrin- An equivalent body in order to exhibit an sically more powerful emitters of radio noise, equal surface "brightness" in the radio region require a correspondingly more violent origin. not only would have to be at an extraordinarily The collision of two galaxies seems to be a high temperature but also would have to have sufficiently drastic occasion, but there are pe- an equivalent temperature that rises with in- culiar galaxies such as M87 which evidently re- creasing wavelength. In the meter-wave re- quire some other means of excitation. gion, where many sources are sufficiently in- The existence of two different populations of tense to be seen with relatively simple equip- radio sources raises the important question of ment, they appear superimposed on a smooth how to distinguish between them, and how to background of radio noise, most of it evidently establish, if possible, a distance scale. The of galactic origin. only reliable method at present seems to be iden- Interferometric techniques are particularly tification with a visual object, but as time goes valuable in removing the effects of the back- on this method must prove increasingly difficult. ground, "labeling," as it were, the signals origi- Even now it is clear that if a pair of colliding nating in discrete sources. Figure 1 illustrates galaxies as violent as the source exists, the principles of the phase-switching interferom- but 10 times farther away, we have little hope eter first described by Ryle (1952). In the of observing it visually even though it would upper drawing, two antennas are arranged still be an observable radio source. Similarly, symmetrically to form the radio analog of many galactic objects, conspicuous in the radio the Michelson interferometer. If a radio source region, are undoubtedly so obscured by is directly overhead, or is off-axis by an amount that optical observation will be difficult. In sufficient to result in a path difference of both cases, however, a distance scale would cer- an integral number of wavelengths, the two tainly bring to light interesting new problems. signals constructively interfere. Therefore, by For example, the distribution in space of extra- switching the phase of one antenna at a regular galactic radio sources may provide a clue to im- rate the signal from a discrete radio source is portant ! problems, since it is prob- modulated, while the slowly varying back- able that many observable radio sources are ground is not, and we1 can distinguish the noise situated at great distances, approaching the from such sources by looking for a modulated limits of the visible universe. So far, only sta- component in the total signal received. The tistical methods have been employed, and at system has other advantages as well. If a present there appears to be a serious discrepancy source is at a half-power point of the inter- between observations. The results of the Cam- ference pattern, the signal received will be the bridge interferometer radio survey indicate an same whether a half wavelength has been in- unusually large number of faint sources, imply- serted or not, and consequently at these po- ing an actual increase in apparent spatial density sitions the signal from a discrete source is un- of radio sources in the universe at great dis- modulated. As a result, position measurements tances. On the other hand, preliminary obser- can be made by a null method, with much vations by Australian radio astronomers show greater precision than is possible with a pencil a number-magnitude relation that implies a beam. Furthermore, the relative amplitude of nearly uniform space density of radio sources. the interferometer signal as a function of an- The resolution of this direct observational con- tenna spacing gives a measure of the angular flict is certainly one of the most pressing prob- size of the source, effectively constructing the lems in radio astronomy today. The solution transform, just as in the optical Michel- may lie in an understanding of the techniques son interferometer. It is possible, therefore, to involved, for both surveys utilized interfero- obtain information with relatively small an- metric techniques but in radically different tennas that otherwise could be obtained only forms. with a single antenna of prohibitively large size. BURKE AND TUVE PLATE 1

The Crab Nebula. Photograph from the Mount Wilson and Palomar Observatories (200-inch photograph).

NEW HORIZONS IN ASTRONOMY 33

I I I I •>v

ANTENNA CABLE CABLE ANTENNA A IRECEIVERI B DIRECTIONAL RESPONSE WITH EQUAL CABLE LENGTHS; SIGNALS FROM OVERHEAD SOURCE ADD

11< • 1111 L_ at o I RECEIVER! -g- X

HALF WAVE DIFFERENCE IN CABLE LENGTHS: SIGNALS FROM OVERHEAD SOURCE CANCEL FIGURE 1.—Phase-switching interferometer.

There is danger in the method, however, The study of the sun is an interesting special which must be guarded against; if more than case, for the radio emission at longer wave- one source is being received, the interferometer lengths provides a useful tool for probing the will measure the combined effect of the two, corona; here interferometers have already and, unless one reconstructs the Fourier trans- proved their value, providing us with a qual- form by going to sufficiently large spacing, a itative picture of the sun at a number of differ- single source will appear to be observed where ent wavelengths. The detailed appearance of in reality there are two. The effective gain of the sun, particularly at the longer wavelengths the interferometer is the geometric mean of (greater than 4 m) needs much more study, the gains of the two antennas, and at first sight and there is little doubt that interferometric it would appear that if we simply make the techniques must prove necessary. antennas larger, and thus see fewer sources per It was remarked earlier that the effective beam width, this difficulty would be circum- sensitivity pattern of a pair of antennas used vented. It is true that the confusion of stronger as an interferometer is equal to the geometric sources may be resolved by this technique, but mean of the gains of the individual antennas. the increased antenna gain then makes it possible This fact is utilized in the special case of the Mills to detect a still larger number of sources per Cross (Mills and Little, 1953), which, in reality, beam width, and since one naturally wishes is a zero-spacing, phase-switched interferometer to examine these, also, multiple spacings are utilizing two linear arrays at right angles. Fig- still required. At the longer wavelengths, ure 2 gives an illustration of the principle. particularly, where a pencil beam of the order Each array has a fan-shaped sensitivity pattern, of a minute of arc or less requires an array of but when connected as a phase-switched inter- prohibitively large size, interferometric tech- ferometer the effective sensitivity is the geo- niques are indispensable, but they must be metric mean of the two antenna patterns, which applied with caution and full knowledge of the is a pencil-beam whose resolving power is equiv- difficulties involved. alent to a square array containing many more 34 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.! PRINCIPAL LOBE WHEN PATTERNS ARE MULTIPLIED

MAIN LOBE MAIN LOBE OF X ^ OF "B"

ARRAY 7C ARRAY OUTPUTS TO RECEIVER FIGURE 2.—Antenna diagram of the Mills Cross interferometer. elements. At the longer wavelengths this pro- meter or more, have been emphasized since vides us with a technique for producing a pencil- the observation of a large number of sources beam with simplified arrays, and the future will at these wavelengths can probably be accom- undoubtedly see arrays based on this principle plished more conveniently by interferometric but of far larger dimensions than those in use techniques than by an equivalent array, while today, since a pencil-beam is unquestionably the measurement of angular size of small the best way to investigate extended sources. sources is probably possible by no other means. As in most experimental and observational The phase-switching technique has the further fields, the advantages of this method are ac- advantage that the noise originating in discrete companied by disadvantages as well, for this sources is "labeled" by modulation, and if, in system is troubled by spurious responses aris- the future, it is possible to investigate the ing from its unusual diffraction pattern. It intrinsic properties of source noise this feature is possible for a strong source, passing well may be invaluable in distinguishing the desired outside of the main pencil beam, to appear as signal from the galactic noise background. a weak source, and great care must be exercised All present interferometric systems have weak- to avoid, if possible, or at least to identify such nesses, however, which must be appreciated spurious responses. (See fig. 3.) if the important work on weak radio sources The longer wavelengths, of the order of a is to have reliable significance. NEW HORIZONS IN ASTRONOMY 35

PATH OF STAR

FIGURE 3.—Detailed antenna pattern of a Mills Cross interferometer. 36 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. 1 In our concern with equipment it is important essential? The answers to these questions, that we not lose sight of the larger issues. The and others as well, depend on an increase in cosmological problem has been outlined, but other the information concerning discrete radio important questions also exist. What is the sources. Larger interferometers will certainly cause of radiation in the case of those galaxies play a role in the gathering of this information. that are not, apparently, colliding with any- References thing? What is the significance of the dis- tribution of radio sources in our own galaxy? MILLS, B. Y., AND LITTLE, A. G. And, what are the physical conditions necessary 1953. Australian Journ. Phya., vol. 6, p. 272. RTLB, M. for the emission of radio noise from large, 1952. Proc. Roy. Soc. London, ser. A, vol. 211, turbulent gas clouds? Are magnetic fields p. 351. Related Sciences

Rocketry By R. Tousey *

For many years astronomers the world over Improvements needed in rocketry have dreamed that sometime it might become Although the V-2 rocket was capable of possible to discover the nature of the extreme carrying a tremendous payload to an altitude ultraviolet and X-ray spectrum of the sun and of 100 miles, which is adequate for many the stars. Research in this field became experiments, it was extremely complicated, possible at the conclusion of World War II difficult to launch, expensive, and more often as a result of the development in Germany of than not it misfired. Two American rockets the V-2 rocket, which was capable of carrying were designed to replace the V-2 for upper scientific equipment to altitudes far above the atmosphere research: the Aerobee, a small, main absorbing regions of the earth's atmos- comparatively inexpensive rocket with a peak phere. The V-2 rocket and the American altitude of about 120 km; and the Viking, Viking and Aerobee rockets which followed similar to the V-2 hi complexity and size, have opened the field of rocket astronomy. capable of 200-km altitude, and stabilized The cost of this research is large, but the stakes during . A number of Vikings were are very high, for it seems probable that the flown with successful experiments, but the key to many unsolved astrophysical problems trend toward economy has forced scientists may be found in the nature of the spectrum of to use and improve the Aerobee rocket for the radiation of very short wavelengths emitted research. by the sun and the stars. The Aerobee uses an acid-aniline fuel, is The difficulties of conducting astronomical launched with a booster from a tower, and research from rockets are many. The space costs at present about $30,000. This rocket is available for instrumentation in all but the quite satisfactory for many experiments. It largest rockets is rarely more than 10 cubic can be launched at sea, though not as easily feet, and the payload is of the order of only a as is desirable. The early type of Aerobee did few hundred pounds. The time available for not reach altitudes sufficient for many experi- observation is 2 to 5 minutes. It is difficult ments. The Aerobee-Hi, now in production, to keep instruments pointed at the celestial will ascend to 200 to 300 km with payloads be- object under study since rockets roll and turn tween 150 and250 lbs.; thus,it has performance in an unpredictable fashion. Flight prepara- characteristics that are satisfactory for almost tions are time consuming and complicated, so all types of research. Although it is a much that the firing time is often determined by simpler and cheaper rocket than the V-2 or circumstances other than those that are opti- Viking, the Aerobee is still too expensive to be mum for the experiment. used as frequently as would be desired in In spite of the vicissitudes of research from astronomical research. rockets, during the past several years great The preparation of a rocket and the launch- advances have been made toward the solution ing are complicated. Furthermore, at land of the problems involved in rocketry; yet, there bases such as the White Sands Proving Grounds, is much still to be desired. Astronomers are the scientist is limited to firing at a particular most anxious that the work be carried ahead as tune, when the range is ready and clear and the vigorously as possible. are favorable, even though he may wish 1 U. S. Naval Research Laboratory, Washington, D. C. to fire, say, during the onset of a . 39 40 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 The improvement most needed in rocketry, high, therefore safe recovery is mandatory if therefore, is a really inexpensive rocket that many experiments are to be conducted. can be made ready and flown without elaborate Further development of pointing controls in preparations and precautions. Recent ad- the direction of lower cost and greater load- vances encourage hope that such a rocket, carrying ability and simplicity is very important employing a solid fuel , can be devel- for rocket astronomy. With such controls the oped in the next few years. A rocket that will study of the bright stars from rockets may carry 100 pounds to 300 km and can be launched become possible. This will probably require a by a few persons, perhaps by the scientists pointing control that operates from the moon, themselves, at a cost of perhaps $5,000 no longer or the sun, and the earth's magnetic field, or seems unattainable. Such a rocket might best possibly from simply the moon and the sun; be launched at sea, or at least flown out to sea the star itself may provide the final signal for from shore. This would avoid complicated precise pointing. and restricting safety requirements. Sea re- covery methods would have to be worked out, The fields of rocket astronomy since most astronomical experiments are photo- The experiments already conducted from rock- graphic. This involves the sealing and flota- ets cover many fields of research in astron- tion of the section containing the instrumenta- omy and the physics of the upper atmosphere. tion and the development of a system for locat- The work already accomplished has been most ing the floating section. recently summarized by a collection of papers For over land, recovery of photographic presented at the Gassiot Committee meetings film may be achieved simply through the use (Boyd and Seaton, 1954). Many fascinating of a strong casette, but ordinarily the equip- new experiments have been proposed. In this ment is completely ruined. Recent advances section an attempt is made to cover some of the in parachute techniques, however, make it most interesting. possible now to salvage the equipment for The solar spectrum.—The sun is certainly the further study or reuse. The cost of the para- most important and the least difficult celestial chute assembly must be greatly reduced, object to study. From the rocket work con- perhaps to the neighborhood of $500 to $1,000, ducted by the Naval Research Laboratory before its general use will be warranted econom- (Tousey, 1953a, 1953b; Wilson et al., 1954; ically. Johnson et al., 1955), by the University of One of the most difficult and important Colorado (Rense, 1953), and by the Air Force problems in rocket astronomy is that of keeping Cambridge Research Center (Jursa et al., 1955), the equipment accurately pointed at the celes- its ultraviolet spectrum is now known in con- tial object being studied. Although it may be siderable detail to 977 A. Friedman (1955) and possible to stabilize a rocket in flight, the preci- coworkers (Havens et al., 1955; Friedman and sion of equipment stabilization required for Chubb, 1955; Byram et al., 1955), using photon many experiments can be far more easily and counters, have measured the sun's X-ray cheaply provided by stabilizing only the emission in several wavelength bands between instrument. So far, attempts at pointing have 100 A and 5 A, and also have obtained much been limited to the sun. The most successful information about the Lyman alpha line of of several "sun-followers" is the "Biaxial hydrogen and the dissociation of in the Pointing Control" developed by the University upper atmosphere. of Colorado under Air Force sponsorship and Down to 2085 A the ultraviolet solar spec- described by Stacey, Stith, Nidey, and Pietenpol trum resembles, generally, the visible spectrum (1954). This device kept a spectrograph flown in having many deep lines. The in a U. S. Navy Aerobee on February 22, 1955, intensity continues to fall below that of a 6,000° pointed at the sun within an average error of K , and at 2100 A has reached a ± 1 minute of arc during a total exposure time brightness temperature of approximately 5000° of 2 minutes. The cost of this equipment is K. The photon counter measurements indicate NEW HORIZONS IN ASTRONOMY

4,050° K at 1450 A and 3,900° K at 1250 A. Special studies of the sun.—The solar spectrum At these wavelengths the continuous spectrum is most easily studied by using the light coming must come from the very coolest layers of the from a considerable area of the solar disk. sun's atmosphere. At 2085 A strong continu- It is important, however, to learn exactly ous absorption of unknown origin sets in; where on the disk the various emissions orig- although nearly obliterated, the Fraunhofer inate, and how their intensities depend on the lines can still be seen, the lowest so far identified height in the solar atmosphere. Several ap- being at 1700 A. The continuum itself has proaches to this problem have been proposed. been photographed to 1550 A. Some information can be obtained with a Below 1650 A the solar spectrum consists, in stigmatic spectrograph on whose slit the sun's the mam, of emission lines. So far, 32 emission image is held stationary to within 1 minute lines have been identified. These include Ly- of arc with a pointing control. Spectra man alpha and beta of hydrogen, the two obtained in this fashion by Johnson et al. resonance lines of O VI at 1032 A and 1038 A, (1955) show, for example, that the O VI line one resonance line of N V at 1239 A, the line at 1032 A has a greater ratio of limb to center of He II at 1640 A corresponding to Ha, and of the disk intensity than does Lyman beta at various (mostly) resonance lines of and 1026 A, as would be expected. More measure- up to three times ionized. The intensity ments of this sort would make it possible to of Lyman alpha has been measured on several improve greatly the model of the chromos- occasions, both photographically and with phere and corona. photon counters, with results varying from 0.1 Ideally one would like to fly a spectro- to 1 or more ergs/cm2/sec, suggesting that the heliograph and to measure the intensity dis- emission in this line may be quite variable, tribution over the disk and corona of all lines in depending on solar conditions. The soft X-ray the vacuum ultraviolet. For the Lyman alpha intensities measured with photon counters line, which has very high intensity, such an appear to be roughly those expected from a instrument is feasible. A possible design was corona at 750,000° K. They have been shown published by Behring et al. (1954). Several by Havens, Friedman, and Hulburt (1955) to Lyman alpha spectroheliographs or cameras be quite sufficient to produce the E-layer of the will be flown in the next or two. Following ionosphere. this, weaker lines of higher excitation energy There remains a great deal of fascinating re- should be investigated, including the ultimate search on the sun to be done from rockets. lines of He I and II at 584 A and 304 A, In this field only the surface has been scratched. respectively. While the data already obtained bear closely Another method of studying the distribution on the structure of the sun, much more needs to of Lyman alpha over the disk, suggested by be learned to make possible the formulation of Friedman, employs a photon counter of a a complete picture of the solar atmosphere. narrow field of view, sensitive to this line only. The solar spectrum from 977 A to 100 A is still By using a television scan of about 20 lines completely unknown, and from 100 A to 5 A it resolution, a crude picture could be transmitted has been measured only in broad chunks. In to ground. this spectral range lie the resonance lines of most Once these experiments have been successful of the highly stripped atoms present in the it becomes of immediate interest to study the corona. It is now possible to study this re- spectrum and the distribution of the spectrum gion. The requirements are a good pointing over the disk under disturbed solar conditions control and a rocket that will spend several and during an intense flare. The solid propel- minutes above 200 km. A scanning spectro- lant rocket and sea recovery are almost a graph with a photon counter detector and prime requisite for this purpose because they telemetering may be the simplest approach, at make it possible to fire at almost the instant a least to obtain information without great flare takes place. Though this may also be resolution. In the end, however, spectrum possible with Aerobee-Hi, it will be difficult photographs should be obtained. indeed, owing to the safety requirements 42 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 associated with launching from ground and the chromosphere and corona. This experi- because of the great complications surrounding ment would be much better performed with a the use of an acid-aniline fuel. solid propellant rocket, since the balloon may A first approach to the question of the change drift out of the path prior to totality. of ultraviolet and X-ray emission during a The stars.—It would, of course, be of tre- flare is planned by Friedman and his coworkers. mendous interest to use rockets to study the They will instrument a number of "Rockoons." short-wavelength spectrum of the stars. Though These small rockets are carried aloft at sea with their faintness in the visible makes the problem balloons, and launched from an altitude of appear extremely difficult, certain of the O and 80,000 feet; they reach approximately 80 km, B stars are so very hot that their extreme ultra- which is sufficient for studies of Lyman alpha violet and X-ray spectrum may well be suffi- and hard X-rays. When they are launched at ciently intense. Exploratory investigations can sea they can be watched during the entire day be accomplished with photon counters, and here and fired at the moment a flare appears. Each again Friedman has planned several experi- rocket will carry a photon counter sensitive to ments, both by day and by night. For the Lyman alpha only, another to detect the first experiments the sky will be scanned, as the possible presence of 1-2 A X-rays, and a third rocket rolls and yaws, by a Lyman alpha photon device to search for a burst of 7-ray Brem- counter of high sensitivity and also by certain strahlung that may be produced in a shock other ultraviolet and X-ray counters. In later wave in the solar atmosphere associated with experiments the counter will be directed with a the flare. It is also possible to have several pointing control. By day the situation is com- Rockoons ready to launch in sequence, a few plicated by the expected presence of scattered minutes apart, to study the time history of Lyman alpha radiation from the sun, which in the radiation from the flare. Such experi- itself is worth studying. At night, however, ments can be done far better from a simple Lyman alpha and X-rays may be expected from solid propellant rocket, since this would carry stars, nebulae, and other galaxies, and there is the instruments much higher than the Rockoon, even the possibility of detecting these radiations to altitudes where soft X-rays and regions of in the zodiacal light. The relation between the the ultraviolet other than Lyman alpha can F-corona and the zodiacal light could be studied also be studied. Furthermore, if no flare by visible or ultraviolet light photometry of the should appear during the day the rockets would sky from a rocket at altitudes where the Ray- not be launched, and so would not be lost, as leigh sky background is very dim. are Rockoons. Many experiments can be imagined if an Still another interesting experiment would be accurate stellar pointing control were developed. to study the extreme ultraviolet and X-ray solar For example, A. Boggess and J. E. Milligan have spectrum during a . It is not out suggested photographing the stars in the ultra- of the question to do this spectrographically violet with a Schmidt camera and objective with solid propellant rockets launched at sea. A prism. The energy distribution of ultraviolet further development would be required to enable spectra, worked out from these spectra, would a pointing control to track accurately in spite of be of great interest in connection with stellar the enormous change in solar light. It would atmosphere studies. It might also be possible also be quite possible to photograph the flash to extend knowledge of the wavelength depend- spectrum, and the results would be most im- ence of interstellar reddening into the ultra- portant for deducing the correct model of the violet. Ultraviolet spectroscopy of the planets chromosphere. might well give information concerning possible Friedman has proposed studying the sun absorbing atmospheric constituents. Perhaps with photon counters and Rockoons during an most exciting of all, however, is the possibility eclipse. Here one would hope to determine the that the spectra of many of the stars could be distribution of intensity in the corona of radia- photographed in the vacuum ultraviolet. tion in several bands of the X-ray region, and .—There are other fields of astronom- the distribution of Lyman alpha with height in ical research which can be pursued by means of NEW HORIZONS IN ASTRONOMY 43 rockets. One of these is the study of micro- looked for, since it may be expected from re- meteorites. These tiny particles are thought to combining protons and electrons. bombard the outer atmosphere in such abun- Outside the auroral zone spectacular displays dance that it may be possible to detect acous- are infrequent, but there is still present the night tically their impacts on a rocket, or even to airglow, an emission of light from the upper at- collect a sample and parachute it to earth for mosphere. The origin of the various emissions study. Experiments of this sort are being in the airglow and their possible connection with conducted by both the Navy and the Air Force. the aurora are not well understood; and opinions Another field is the analysis of the primary differ concerning the altitudes at which the cosmic radiation. This is difficult at low alti- various spectral lines are emitted. Experi- tudes because of the modifications and showers ments from rockets designed to study directly that take place due to the stopping effect of the these emissions are underway; they should be air. At rocket peak altitudes, however, it continued until the airglow is well understood. should be possible to study the real primary For all these experiments the development of radiation. the solid fuel propellant rocket is most impor- Upper atmosphere physics.—Finally, there is tant. Since many experiments will be required, the large field of research on the physics of the at special times and locations, a cheap and earth's upper atmosphere, a field claimed alike simple rocket is essential. by astronomy, geophysics, and meteorology. Rocket work has already solved the problem of the origin of the E-layer, and when higher A satellite is simply a special type of flying rockets are available the radiation respon- rocket that is projected at a velocity proper to sible for the F-layers will certainly be dis- place it in a stationary . It is especially covered. Direct measurements of the composi- suited to experiments that require a long ob- tion of the upper atmosphere have been made, serving time. In the foreseeable future, how- and the density of ions of various sorts has been ever, the space and weight limitations on instru- measured with mass spectrometers. The elec- mentation are very severe, and there can be no tron density through the E and Fi layer has thought of recovery. already been measured from rockets by a radio There have been proposed several important propagation method by Seddon (1954) and experiments that cannot be done with ordinary coworkers (Seddon et al., 1954). This work rockets and may be feasible from a satellite. should be carried on at many points over the Perhaps the simplest experiment is the study earth and under varying ionospheric condi- of solar ultraviolet and X-ray radiation over a tions. The pressure and density of the atmos- long period of time in order to determine the phere have been measured from rockets and type of intensity variations that may occur. the work should be continued at many places Such variations are considered to be a cause of over the earth and at different seasons. It sudden changes in the state of the ionosphere may be noted that the density measurements and radio transmission, and may also have an of Friedman (1955), made by studying the important effect on weather. Photon counters absorption of X-rays by the atmosphere, do not would be used. agree with measurements made close to the Friedman and coworkers have proposed a rocket with pressure gages. This interesting search of the sky for radiation hot spots by conflict of results must be resolved. means of detectors in a satellite. The Lyman Finally, it is hoped that the origin of aurorae alpha line would be the radiation chosen for can be explained by means of rockets. It is the first experiments because of its probable expected that during the International Geo- high intensity and possible correlation with physical Year rockets can be launched within known regions of high concentration of the auroral zone and flown into actual aurorae. hydrogen. The nature of the incoming particles will be The earth's magnetic field could very well studied. Emission of Lyman alpha will also be be studied from a satellite. Since the satellite's 44 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.]

orbit would cover a large fraction of the earth BOYD, R. L., AND SEATON, M. J., EDS. during a day, it would be possible to study the 1954. Rocket of the upper atmos- phere. Interscience, New York. variation of the earth's magnetic field with BYRAM, E. T., CHUBB, T. A., AND FRIEDMAN, H. latitude and longitude, and, if the orbit were 1955. Phys. Rev., vol. 98, p. 1594. elliptical, also with altitude. Experiments of FRIEDMAN, H. this sort are important in connection with ttfe 1955. Ann. de Geophys., vol. 11, p. 174. -St0rmer ring current. An actual FRIEDMAN, H., AND CHUBB, T. A. 1955. In Report of the 1954 Cambridge Con- experiment from a satellite to determine the ference, pp. 58-62. The Physical Soci- magnitude of the ringcurren t has been proposed ety, London. by J. P. Heppner. HAVENS, R. J., FRIEDMAN, H., AND HULBURT, E. 0. The satellite would, of course, be ideally 1955. In Report of the 1954 Cambridge Con- ference, pp. 58-62. The Physical Soci- suited to the study of so small ety, London. that they do not penetrate far into the earth's JOHNSON, F. S.; MALITSON, H. H.; PURCELL, J. D.; AND atmosphere but which probably ionize the TOUSEY, R. upper atmosphere in process of being stopped. 1955. Astron. Journ., vol. 60, p. 165. Even more important, however, may be the JURSA, A. S., LEBLANC, F. J., AND TANAKA, Y. study of primary cosmic rays of energies so 1955. Journ. Opt. Soc. America, vol. 45, p. 1085. RENSE, W. A. low that they do not penetrate far into the 1953. Phys. Rev., vol. 91, p. 299. atmosphere. Particles of this sort are believed SEDDON, J. C. to exist and, in fact, many of them may come 1954. Journ. Geophys. Res., vol. 59, p. 463. from the sun. The satellite also offers the SEDDON, J. C, PICKAR, A. D., AND JACKSON, J. E. possibility of studying the distribution in 1954. Journ. Geophys. Res., vol. 59, p. 513. STACEY, D. S.; STITH, G. A.; NIDEY, R. A.; AND PIETEN- space from which cosmic radiation arrives, and POL, W. B. perhaps of providing clews to the origin of this 1954. Electronics, vol. 27, p. 149. mysterious radiation. TOUSEY, R. 1953a. In Kuiper, ed., The sun, p. 658. Univer- sity of Chicago Press. References 1953b. Journ. Opt. Soc. America, vol. 43, p. 245. BEHRING, W. E.; JACKSON, J. M.; MILLER, S. C, JB.; WILSON, N. L.; TOUSEY, R.; PURCELL, J. D.; JOHNSON, AND RENSE, W. A. F. S.; AND MOORE, C. E. 1954. Journ. Opt. Soc. America, vol. 44, p. 229. 1954. Astrophys. Journ., vol. 119, p. 590. Computing Machines in Astronomy By Paul Hergetl

The impact of computing machines upon meteor studies currently being conducted by scientific research in recent years is not fully Whipple. appreciated outside of a small group of actual In astrophysics, one large group of problems machine users. When an electronic computa- concerns stellar models. The equations of tion center is installed in a scientific or technical state and equilibrium conditions for stellar institution, almost invariably a workload interiors form a system of ordinary differential quickly develops that requires more than the equations in one independent variable, the available running time on the computer. This radius. Similar mathematical but different is a "snowballing" situation in which the dem- physical conditions exist in stellar atmospheres. onstration of what can be done on one problem Recently Henyey has elaborated on the prob- soon leads to the exploration of similar pos- lem of stellar interiors by taking into account sibilities in related problems. the evolutionary changes within the star. The next effect is that nearly all nondefense Again, if one deals with pulsating stars, not and pure research projects are crowded out, for only the stellar radius but also the time must lack of priority. The National Science Founda- be used as an independent variable. Then tion has recognized this situation and called there are such specialized problems as the search an ad hoc committee in February 1955 to con- for a mechanism that can explain the occurrence sider the need for computing facilities in of spicules on the sun. university research centers in general. Astron- Another new field of computational problems omy was represented by Schwarzschild and is now arising from considerations of magneto- Herget. More recently other groups have hydrodynamics. These are even more difficult recognized the need to stimulate the develop- because they involve partial differential equa- ment of computation centers in universities. tions. The whole field of "wave equations" The need for electronic computing machines is of this same type. Much work has been done in classical astronomy is mentioned in another by L. C. , L. H. Thomas, and others, but paper in this series, "The Minor Planets." it is always necessary to restrict the equations As additional examples, we may note that somewhat because even the largest computers , Clemence, and , using these in existence are still inadequate to deal with machines, have completely recomputed the this problem in its entirety. This problem is motions of the outer planets and Herget has of interest not only for the light astronomical prepared a uniform system of coordinates for atoms but more or less for the whole periodic Earth and . Clemence has recomputed table up to the fissionable and heaviest elements. a new general theory for which is nearing Examples of problems in stellar atmospheres completion. In studying the secular variations may be found in the Proceedings of the National of the elements of the major planets, it is now Science Foundation Conference held at Indiana feasible to compute second order effects simply University in October 1954. Other typical by recomputing the coefficients under the problems are given in the report of the Joint appropriately varying conditions at intervals Meeting of the American Astronomical Society of, say, 1,000 years. An extensive computa- and the Association for Computing Machinery tional program is also a significant part of the held at Ann Arbor, Mich., June 1954. One interesting problem was the computation of the 1 Department of Astronomy, University of Cincinnati, Cincinnati, Ohio. history of a globular cluster (in two dimensions) 48 46 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 in which the individual stellar were The design of radio antennas for special carried forward simultaneously until, eventu- purposes is another new computational problem ally, two out of a hundred stars evaporated in astronomy. from the cluster. It is significant that, in nearly every one of Optical design, lens computations, and ray- these astronomical problems, the researcher tracing have been revolutionized by the devel- has been able to develop his own facility in opment of high-speed computers. Thus the the use of an electronic computer as a research cost of experimenting with a novel design is tool even though he has not been primarily a greatly reduced, since all the effects can now computer. The use of computing machines in be so readily computed. Highly efficient meth- astronomy will undoubtedly grow exponentially ods have been developed, especially by Baker, as each successful solution breeds more prob- but there is still room for further study as the lems, and each new experience inspires more capabilities of newer machines are improved. ideas. Turbulence By S. Chandrasekhar1

An aspect of theoretical investigations in as- of turbulence is no greater than was our under- tronomy and geophysics that is not often recog- standing of the dynamical theory of gases, be- nized is that matter and motions, in the large, fore Maxwell and . may exhibit patterns of behavior that one might Since a deductive fundamental theory of never suspect if one restricted oneself to matter turbulence is a prerequisite for a rational com- and motions in the small. While several ex- prehension of many astrophysical and geo- amples can be cited to illustrate this, the most physical phenomena, an attempt to follow cur- striking is that which requires all motions in rent developments in the theory of turbulence systems with large linear dimensions to be tur- should be an integral part of astronomical edu- bulent; for turbulence will be the natural state cation. The need is the greater in view of the of motions if the number is sufficient- increasing importance of magnetism in astro- ly large. And since the Reynolds number con- nomical thinking; we have only to recall in this tains the linear dimension of the system as a connection the problems of interstellar polari- factor, it is clear that hydrodynamical motions zation and magnetic fields, the synchrotron which occur in astrophysics and geophysics must radiation from the Crab Nebula, tie galactic necessarily be turbulent. Indeed, "turbulence" corona, and the still unsolved problem of the is the refuge under which astronomers most earth's magnetic field. commonly seek protection whenever they do not Hydrodynamics and hydromagnetics are understand a particular phenomenon. At pres- bound to play central roles in future astronom- ent, however, our understanding of the nature ical developments and the part of turbulence is « , Williams Bay, Wls. only one aspect of this. 47

Astroballistics By Richard N. Thomas 1

Over the postwar years a small group of pletely understood all the processes that occur astronomers have found themselves acting as when a solid object moves through a gaseous midwives at the birth of a new subject, astro- atmosphere, we would have a fine method for ballistics, conceived on the common borders inferring the corresponding properties of each of astronomy, aerodynamics, ballistics, and meteor observed. We do not yet have such chemical -kinetics. Its multiple origin an understanding, however, and the science of provides an excellent example of the preoccu- astroballistics has arisen from our attempts to pation with crossfield research in contemporary remedy this lack. Our present knowledge of astronomy. Its evolution exhibits the trend, meteors, as a part of the interplanetary medium, beginning with spectroscopy, which is changing is discussed elsewhere in these papers. Here astronomy from a wholly observational and we consider the meteor only as it represents a theoretical science to one demanding extensive solid object moving through a gaseous atmos- laboratory investigations. Its future depends phere, under one variety of conditions. on our success in establishing this crossfield When a solid moves through a gas, several work; that is, on our continuing to produce a things may happen. The solid loses momen- group of investigators who feel at home in at tum to the gas and is decelerated. The amount least the overlapping parts of the four areas— of deceleration depends critically upon the astronomy, aerodynamics, ballistics, and chem- mass, shape, and velocity of the body, and upon ical reaction-kinetics. the composition and thermodynamic state of the gas. The solid loses energy as it deceler- Origin and scope ates, and the energy contributes heat to both The motion of planets and planetoids was for body and gas. The partition of heat between years a conventional part of study in astronomy. the two depends critically upon the velocity, In other papers of this group appear descriptions shape, composition, surface structure, and of some of the exciting new data and theories thermodynamic state of the body, and upon the which have deflected our interest in these composition and thermodynamic state of the objects from their kinematic to their physical gas. The heat transfer to the solid may result properties. We recognize, however, that we can in and (either by hardly claim a satisfactory knowledge of our melting or by vaporization) from the surface, as until we can also specify well as in a simple temperature rise throughout the properties of those objects so small that the interior. The heat transfer to the gas may they cannot be observed individually. In result in excitation and ionization, and thus in general, our information on these objects is radiation, if the temperature rises sufficiently. wholly statistical, coming from studies of their Moreover, the material ablated from the solid collective effect on sunlight, action may interact chemically with the gaseous at- as a resisting medium, etc. One notable excep- mosphere, serving either as a heat source or tion is the study of meteors; for these, the earth sink, depending upon the reaction. serves as a sampling box, sweeping up the The astronomer can make at least three types interplanetary material, with the earth's at- of measures on a meteor moving through the mosphere as a fluorescing medium, responding atmosphere: (1) He can measure directly time to the entrance of each meteor. If we com- and position, and so obtain velocity and 1 Harvard College Observatory, Cambridge, Mass. deceleration; (2) he can measure the total radi- 48 50 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 ation, and possibly its spectral distribution; and turn to luminosity and ionization studies; and (3) by radar, he can infer the ionization pro- here, the cupboard is essentially bare. duced by the meteor. Ideally, he would like to Ballistic and aerodynamic studies of objects have available a set of tables which he could moving at such high speeds that ablation occurs enter with these data, and so read off the phys- are virtually nonexistent, at least so far as we ical characteristics of a single body which alone can infer from the unclassified literature. A could have produced such a disturbance. At few simple free-flight experiments have been worst, he might expect to obtain a range of such made, with velocities reaching ~6 km sec"1 at bodies, and this range would represent his un- maximum. We find an equally small number certainty of the properties of the particular me- of wind-tunnel studies, involving always sub- teor observed. These properties he could com- stances requiring but low heat transfer for pare with those of rare meteors that were large ablation. From a theoretical standpoint, the enough and slow enough for the bodies to sur- situation is equally bad. The study of the vive passage through the atmosphere, and be aerodynamic flow pattern about a body that is recovered as meteorites. losing mass at its leading surface has hardly Unfortunately, when we attempt to find or been discussed. Indeed, the most satisfactory even to prepare such a set of tables, we find a theoretical work on any aspect of the field disheartening situation. Meteors impact on the probably lies in a few studies, made in collab- earth's outer atmosphere at speeds ranging be- oration by astronomers and chemists, relating tween 10 and 75 km sec"1. The ballistician and to ablation in the meteoric regime—high aerodynamicist, to whom we might expect to velocity and very low gas density. turn for information, have just since the war When we consider the data available on become really conscious of the velocity range excitation and ionization caused by the ablated that is of interest to the meteor astronomer. material ejected into the gaseous atmosphere, They, in turn, have looked to the astronomer, we find the bleakest situation of all. In the late hoping for insight into just those problems we 1920's and early 1930's theoretical atomic pose to them. Only during the last war did the physics was striking out into the areas opened aerodynamicist develop a keen interest in speeds by the new , and experi- as high as that of sound, ~% km sec"1. While mental physics was entering those areas opened the ballistician customarily dealt with speeds by improved instrumentation. Problems of several times that of sound, his knowledge was radiative transition probabilities, while not primarily empirical, and his empirical investi- completely solved, were at least understood. gations were limited by bis accelerating devices. On the other hand, problems of collisional The experimental investigations reported in the excitation were vague, expecially those in- unclassified literature do not deal with objects volving exchange reactions, and particularly of known geometry whose speed exceeds ~2 km atom-atom inelastic collisions. A number of sec"1. At these speeds, interest lies in air- preliminary experimental investigations were resistance; aerodynamic heating of the body is reported in the energy range of interest to the negligible. By analogue devices such as the astronomer, < 1 kev. Unfortunately, from our wind tunnel, deceleration properties of bodies of present standpoint, the early 1930's were also known shape at equivalent airspeeds of ~3 km the era of the crucial experiments in the then sec"1 have been studied. Theoretical investiga- infant science of . The focus of tions have made rapid progress in interpreting interest rapidly shifted, and concern with this and correlating these empirical studies. On the area of low-energy atom-atom collisions di- whole, we may say that the deceleration prob- minished. An extensive base of data upon lem is in a most satisfactory state, not only in which to draw never developed, and it is only the range covered by the experiments but, for in recent years, among relatively small groups purposes of extrapolation, considerably beyond of physicists and chemists, that interest in these that range. But deceleration measures alone do problems for their own sake has reawakened. not suffice to permit an inference of the physical The astronomer's interest arises largely from properties of the meteor. For that, one must his meteor studies, and occasionally from NEW HORIZONS IN ASTRONOMY 51 phenomena in stellar atmospheres where, how- up to speeds of 6 km sec"1. Microwave equip- ever, his chief concern lies in electron-atom ment facilitates studies of the ionization inelastic collisions. As the instrumental range caused by the same type of artificial meteor. increases, particularly in studies of very strong Theoretical studies already begun, based on Shockwaves, the aerodynamicist also is in- the insight developed in these first investiga- creasingly preoccupied with the details of these tions, will help guide subsequent research. atomic collision processes, excitation, ioniza- Thus, we are laying a solid foundation from tion, and thus radiation. which to study the problems described earlier To categorize and collate the variety of prob- as largely unsolved. We should, however, lems just described, the science of astrobal- stress one aspect of all these investigations: listics developed. Its province includes the the results of these first attempts have been study of those phenomena arising out of the invariably crude and preliminary, and progress motion of a solid through a gas at such speeds in extending them has been dishearteningly that ablation occurs. Interest in and data on slow. We will return to this point. At various of these phenomena arises in astronomy, present, the one satisfying feature the results aerodynamics, ballistics, and . have in common is the simple fact that they Investigations that combine the points of view exist. of the several sciences involved will have a These first experimental investigations had better perspective and are more likely to solve one very stimulating outcome. The first the various problems than a study based on spectral observations, though crude, demon- any individual science alone. strated that it is possible to observe details of the chemical reactions involving ablated Evolution of the science material and atmospheric atoms. Meteor ob- The field of astroballistics, as such, developed servations have never provided such data, when astronomers recognized that it would primarily because of the unfavorable conditions probably be necessary to set up a systematic of (1) high velocity, which inhibits the ex- program that linked astronomical observations othermic chemical reactions expected when a of meteors with laboratory studies in ballistics metallic spray hits the atmosphere; and (2) and aerodynamics. The immediate postwar low atmospheric density, which permits con- period found ballisticians and aerodynamicists siderable diffusion before conditions are such still preoccupied mainly with measures of the that the reaction may occur. In the laboratory, system of aerodynamic on a or such inhibition and diffusion effects may now and, to a lesser extent, with problems be studied. Thus, a very fertile method for of strong Shockwaves. A number of advances studying chemical reaction kinetics appears. in experimental techniques, however, made it Even more, all these occur in a high-speed possible to begin, at least, to bridge the gaps stream of gas, so that we have an approach to between meteor observations and the data an aerodynamic flow problem of exciting and theory of ballistics and aerodynamics. possibilities. For the astronomer, who is just A certain amount of development of new now acquiring a deep interest in gas-dynamical launching devices, and considerable improve- problems in stellar atmospheres, such observa- ment of the old, plus the use of low melting- tions are invaluable in providing a physical point materials, together have made it possible insight otherwise inaccessible. In the lab- to simulate ablation from meteors. Newly oratory, it is hard to provide gas streams developed free-flight firing ranges permit studies having sufficient internal energy that atomic under varying conditions of the atmosphere and molecular excitation change appreciably through which the solid moves. Advances and observably. Consequently, such data as in high-speed photography allow us to make the above become highly desirable. The re- inferences of the spectral distribution of lumi- mark made above, however, should be re- nosity along the axis and trail of an artificial peated; these studies are chiefly preliminary. meteor, which we have been able to accelerate The field has just been opened. 52 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 Needs and future of astroballistics may be under investigation at a dozen places The gross problems of astroballistics may be throughout the country. But in how many clearly delineated. A very great deal of of these places is there an astronomer working experimental work must take place, both to as a member of the research group, both con- provide the empirical constants required in any tributing and gaining information ? How much theoretical or interpretive work and to estab- of the information gained is accessible in the lish the basic physical feeling for the problems open literature? We should recognize two without which the theory cannot proceed. At problems here. First, we have need for a present, facilities for such work are largely national astrophysical laboratory that can oper- nonexistent or inadequate in comparison with ate with considerable latitude to investigate our needs. Consequently, the field can develop problems such as those in the field of astrobal- only slowly. listics. Its staff should comprise astronomers For a long time the chief characteristic of as well as scientists from "related" fields, and it American astronomy has been the size, quality, should include both permanent members and and quantity of its telescopes. Astrophysics in those on a temporary basis. Second, we must America has grown largely through the output urge our graduate schools to help produce the of observational data. Several spectroscopy staff for such an institution by emphasizing in laboratories are contributing invaluable work their program the experimental aspects of on wavelengths and transition probabilities. astronomy as much as the observational. Aside from these, however, we have no cor- Astroballistics is hardly unique in requiring responding astronomical laboratories devoted these supports; it is, after all, but a small field to the study of astroballistics, of conditions in astronomy. The needs of astroballistics, underlying growth in the interstellar however, are a prototype of those existing in medium, and of magnetohydrodynamic proper- contemporary astronomy in general, and the ties of plasmas and the like. It is of course future progress of the field may serve as a true that some phase of each of these subjects significant example. Electrodynamics of Fluids and Plasmas ByMaxKrook1

Sufficient evidence has accumulated to es- literature in any comprehensive way. In a tablish the widespread occurrence of large- discussion of this kind, it is not possible to do scale electrodynamic phenomena in astro- justice to the individual contributions of the physical contexts. To understand such phe- many authors in these fields, but some published nomena and to interpret them quantitatively reviews which include extensive bibliographies we must investigate the origin, nature, and are listed at the end of this article. behavior of cosmic electromagnetic fields and their interaction with cosmic matter. These Fundamental equations tasks naturally involve considerations of fluid In the kinetic theory of plasmas, the state of a electrodynamics and, from a more general and system is specified by the following: fundamental point of view, considerations of 2 (1) State-variables for the electromagnetic field. the kinetic theory of plasmas. Since these These are the usual parameters of the Maxwell-Lorentz latter are still in a very theory; i. e., electric intensity E, magnetic intensity H, rudimentary stage of development, there is considerable scope for research not only in charge density p«, current density J. (2) State-variables for the matter. These are astrophysical applications but also in the parameters which give a statistical description of the fundamental physical disciplines themselves. microscopic state of the plasma. An appreciable fraction of cosmic material With any dynamical process in a plasma is consists of ionized, and hence electrically con- associated some characteristic length L, and ducting, gas. The dynamical behavior of such some characteristic time T. If L and r are conducting fluids is determined in part by forces both "sufficiently large," the plasma may be exerted by the electromagnetic field on electrical treated as a continuum whose dynamic and charge and current distributions in the fluid. electrodynamic properties are specified by a But equally, the dynamical behavior of the set of material constants; e. g., dielectric con- electromagnetic field is determined in part by stant K, magnetic susceptibility /x, kinematic the fluid motions. In general therefore, con- viscosity v, electrical conductivity

An ionized gas or plasma possesses two presence of magnetic fields is variously referred obvious general classes of dynamical modes as to as hydromagnetics, magnetohydrodynamics, a consequence of its being a mixture of two and fluid electrodynamics. electrically coupled gases. In the electron gas the velocity of sound is much higher than in Hydromagnetics the other constituent, the ion gas, so that if The importance of hydromagnetics in astro- the plasma is deformed sufficiently rapidly, the physical phenomena may be seen from the ion gas will lag behind the electron gas; this fact that much of the region within the galaxy, will result in separation of charge and in large including stellar atmospheres and interiors, is electrostatic forces that tend to restore the occupied by gas in various degrees of ionization initial relative configuration. The resulting and readily transports electric charge. Both oscillations of this "electric" mode are called direct observation and theoretical inference plasma oscillations. In the absence of magnetic now generally support the belief that much of fields, the oscillations have a fundamental interstellar space, together with a large fraction angular frequency of the stars, is permeated by magnetic fields

> with an energy density comparable to the &>,=(4wieTi/m) *, (1) density of many of the gas where e is the charge of an electron in electro- motions. Thus, the magnetic field stresses static units, m is the electron mass in grams, are comparable to the Reynolds stresses. The and n is the number of electrons per cm8. The high conductivity of the matter indicates that frequency defined in equation (1) is called the it is strongly coupled to the pervading field, plasma frequency, and we might add that in so that in treating most fluid motions of astro- the presence of magnetic fields there are three physical interest, instead of the hydrodynamic possible frequencies that are algebraic combina- equation, tions of wp with the electron gyrofrequency Be/me (Westfold, 1949). (2) If the disturbances initiated in a plasma are slow compared to the plasma frequency, then one must use the hydromagnetic equations, obviously the electron and ion gases will move together. In this case the plasma is well approximated by a conducting classical fluid (Elsasser, 1954, 1955a). Since the motions dv (VXB)XB. (4) of an electrically conducting fluid interact df with magnetic fields, and in astronomical phenomena are so generally associated with Here c represents the conductivity of the fluid magnetic fields, this plasma mode may be and M the permeability in mks units. termed the "magnetic" mode. The study of The complexity of the nonlinear hydro- the dynamics of conducting fluids in the magnetic equations indicates that the general problem must be divided into the several basic 1 Assisted In part by the Office of Scientific Research and the Oeo- Phystcs Research Directorate, Air Force Cambridge Research Center, physical processes; consequently, the study of Air Research and Development Command, V. 8. Air Force. hydromagnetics involves the investigation of 'Enrico Fermi Institute for Nuclear Studies, The , Chicago, m. individual effects, one at a time. Instead of a 60 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 formal general solution of equations (3) and lines are analogous to stretched strings, allowing (4), we will ultimately achieve a general under- the propagation of transverse waves with a standing based on quantitative understanding velocity whose square is equal to the stress of the many basic concepts (Elsasser, 1955b). density B2\\i divided by the material density To mention but a few of the physical prin- (Alfv6n, 1950). These hydromagnetic waves ciples, let us consider first a fluid of very high are transverse, and for this reason are generated conductivity. Then equation (3) reduces to more copiously by the transverse fluid motions of subsonic hydrodynamics than are the f=VX(vXB). (5) longitudinal acoustical waves. As the mag- netic energy density becomes large compared to the kinetic energy density of the fluid, all From equation (5) it is readily shown that the hydromagnetic motions reduce to hydromag- lines of force move exactly with the fluid. netic waves (Parker, 1955b). It is interesting Formally, if C represents a moving with to note, however, that the presence of the the fluid, and S the area enclosed by the contour, magnetic field may enhance the generation of then acoustical waves (Kulsrud, 1955). The fundamental problem of the convective (6) stability of a rotating conducting fluid in the presence of a field has been solved by Chan- The lines of force retain their identity, since any drasekhar (1953b; see also 1952, 1953a, 1954a) two elements of fluid threaded by the same line with remarkable results. Simple Benard cells of force will remain so connected. Equation do not occur. Instead, competition develops (5) is the same as the Helmholtz vorticity between the rotational and magnetic effects equation for an inviscid fluid. 's in- and results in two unstable regimes. One tegration gives B as an algebraic function of the might say, very roughly, that the magnetic derivatives of the Lagrangian coordinates of the field tends to stiffen the fluid, producing fluid particles (Brand, 1947). stability, and the forces tend to pro- If the conductivity is not essentially infinite, duce instability. The results of Chandrasek- then of course the lines of force slip relative to har's analytical formulation of this very dif- the fluid. Besides the slipping, however, they ficult problem have been given experimental lose their identity by continually changing confirmation (Nakagawa, 1955). their pattern of connection. Thus, for instance, Chandrasekhar (1955) has also treated the a loop in a long tube of may detach itself problem of hydromagnetic turbulence and finds from the tube, giving a magnetic ring and the that there are two statistical modes, with either original, more or less straight tube. This the fluid motions or the magnetic field dominat- detaching process forms one of the basic steps ing at large wave numbers. The analysis in the regenerative hydromagnetic dynamo shows that there is not a close approach to discussed below. equipartition of energy between the fluid mo- The forces exerted on the fluid by the mag- tion and the magnetic field, but, on the other netic field are derivable from the magnetic 2 2 hand, it does indicate that K pv and B /2n will stress tensor, BtB}/n, (i, j=l, 2, 3), representing tension along the lines of force, and the isotropic not differ by an order of magnitude. magnetic pressure, B2/2fi (Cowling, 1953), in The origin of the stationary dipole field of mks units. Thus the total pressure isp-\-B2/2pi. earth and that of the migrating solar fields Within a region of magnetic field the gas pres- resulting in sunspot formation are beginning sure p will tend to be less than it is outside by to be understood on the basis of hydromagnetic the amount B2/2n. The gas density tends to dynamo effects whereby two magnetic fields in be less within the field, resulting in the phe- a region are each generated from the other by nomenon of magnetic buoyancy (Parker, 1955a). fluid motions. The so-called toroidal field is The tension in the lines of force and the mass produced by simple shearing from the poloidal of the fluid clinging to them indicate that the field; the latter is then regenerated by C37clonic NEW HORIZONS TN ASTRONOMY 61 motions in the fluid, forming loops from the by moving matter. The acceleration is pre- toroidal field (Parker, 1955c; Elsasser, 1955b). sumed to take place in the galaxy and solar Presumably a sunspot is formed from the solar flares. One consequence of the high conduc- toroidal field through a combination of mag- tivity of most matter is that the in netic buoyancy, (Par- the frame of reference moving with the fluid is ker, 1955a), and the observed cooling of very small. Thus, it is readily shown that the interior of the spot, though most of the details remain to be treated quantitatively. E=-YXB, (7) Probably nowhere does one find more phe- nomena whose origin so obviously cannot be where v is the fluid velocity, and there can be explained on hydrodynamic grounds than in no large electrostatic acceleration effects. The the violent phenomena of solar activity, such general solution of the motion of a charged as spicules, active prominences, flares and particle in slowing varying hydromagnetic surge prominences, the ejection of ion clouds, fields, as well as calculations of several cases of and cosmic ray bursts. The origins of these abruptly varying fields, indicate that there are phenomena are today not at all understood, no hydromagnetic accelerating mechanisms but one presumes that at least some of them other than Fermi's mechanism (Fermi, 1949) admit of a hydromagnetic origin. The ter- and the betatron effect (Swann, 1953; Riddiford restrial consequences make solar activity a and Butler, 1952). Since both mechanisms particularly pressing challenge to hydromag- require that the particles initially have energies netic research. of the order of 0.5 Bev per nucleon to avoid the Understanding the terrestrial effects is im- enormous ionization losses at low energies, one portant, but it has often been suggested that presumes that there is some more vigorous solar activity may prove also to be the "little mechanism capable of accelerating particles brother" to some of the grander phenomena from thermal energies in spite of the tremendous that result in the Wolf-Rayet stars, flare stars, energy losses. Finding such a mechanism is and Babcock's magnetic variables where prom- obvious^ of fundamental importance. inence activity, flares, and sunspots have devel- The assumption that cosmic rays are accel- oped to pathological proportions. erated throughout the galaxy (Morrison, Olbert, Solar activity apparently ejects field-bearing and Rossi, 1954; Unsdld, 1955; Parker, 1955b; clouds of gas into interplanetary space. At the Elsasser, 1955a) presents energetic difficulties. earth the clouds are somehow responsible for Davis (1955b) has suggested that the gas ejected the aurorae, magnetic storms, certain iono- from the sun forms a cavity in the magnetic spheric disturbances, the configuration of the field of the galaxy capable of trapping particles terrestrial weather pattern, and the observed in the solar system, though the mechanism interplanetary gas density of as much as 500 apparently cannot account for the cosmic ray hydrogen atoms per cm3 (Storey, 1953; Sieden- cutoff with the attendant correlation with the topf, Behr, and Elsasser, 1953; Behr and sunspot cycle. A galactic origin of cosmic rays Siedentopf, 1953). Further, cosmic ray in- and the sweeping away of galactic cosmic rays tensity variations not directly due to the by outgoing magnetic interplanetary clouds immediate emission of new particles by the seems to give a suitable cutoff, but so little is sun are now generally regarded (Neher, 1955) known about the nature of interplanetary clouds as being due to the antics of interplanetary and Davis' cavity that the details of the mech- clouds. It becomes important, therefore, to anism are entirely speculative. investigate the dynamics of such clouds. If at Besides the evidence from the polarization of the same time the dynamics of ejection from starlight (Davis and Greenstein, 1951), it has the sun can be expounded, we will be able to been shown (Chandrasekhar and Fermi, 1953) check our ideas by both solar and terrestrial that the observed properties of the spiral arms observations. of the galaxy require a magnetic flux density of Cosmic rays are presumed to be accelerated about 10"5 gauss along each spiral arm. A by interaction with the magnetic fields carried magnetic field of this strength permeating the 381014—56 « 62 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. 1 interstellar medium will be important in deter- absence of magnetic fields. The plasma fre- mining the characteristics of stars formed by quency decreases outward so that escaping of the interstellar medium, radio waves of low frequency can only have for, although it does not directly affect ' originated in the outer corona. Higher fre- instability criterion (Chandrasekhar, 1954b), quencies may originate farther into the atmos- the field will strongly influence the nature of the phere. The observed cutoff frequency at the fluid motions against which work the gravita- low end of the frequency spectrum affords an tional forces of collapse. Hydromagnetics evi- indication of the level of the source in the solar dently plays an important role in the formation, atmosphere; the time rate of change of the cut- structure, and evolution of galaxies, though off frequency is an indication of the velocity of little has been done on the integrated problem. the source. But, at the same time, the re- The pulsations of some variable stars suggest fraction of the waves in the solar atmosphere (Chandrasekhar and Fermi, 1953; Chandrase- complicates the problem of determining the khar and Limber, 1954; Ferraro, 1954) that heliographic position of the source. the general magnetic fields of the individual The presence of a magnetic field further stars play an important role in determining the complicates the propagation, producing, among form and period of the oscillations of the stellar other things, a tendency of the transverse body. The tremendous general stellar field electromagnetic waves to propagate parallel to that is required implies a hydromagnetic dynamo the magnetic field, as a consequence of the of immense power. electrons' circling around the field. This effect appears in the terrestrial phenomena of atmos- Plasma problem pheric whistlers investigated by Storey (1953). The extraordinary number of modes of oscil- The propagation depends upon the interplane- lation of the electron gas in a plasma is exhibited tary gas density near the earth and so gives in Bailey's (1950, 1951) general electromagneto- information on the interplanetary medium. ionic theory. Starting with a plasma in the The anomalous sources of radio noise in the presence of a magnetic field, and using Max- sun and elsewhere are often associated with high well's equations, the equation of continuity, velocity gas or particle streams. The rate of etc., neglecting damping by collisions, and con- decrease with time of the observed frequency of sidering small sinusoidal deviations from uni- a given solar burst suggests that the excitation formity, Bailey finds a number of steady state might be due to clouds that are moving out modes with frequencies that are simple alge- through the solar atmosphere with velocities of braic combinations of the plasma frequency uP the order of 600 km/sec (Pawsey and Smerd, and the gyrofrequency

RlDDIFOBD, L., AND BtJTLEB, S. T. SWANN, W. F. G. 1952. Phuos. Mag., vol. 43, p. 447. 1933. Phys. Rev., vol. 43, p. 217. SlBDENTOPF, H., BBHB, A., AND ELSASSEB, H. Twi88 R Q 1953. Nature, vol. 171, p. 1066. ^ ph R L g4 44g SHKLOVSKY, I. S. A 1953. Dokl. Akad. Nauk S.S.S.R., vol. 90, p. 983. UNSOLD, A. 9TOBET. L. R. O. 1955- Zeitschr. Phys., vol. 141, p. 70. 1953. Philos. Trans. Roy. Soc. London, ser. A, WBSTFOLD, K. C. vol. 246, p. 113. 1949. Australian Journ. Sci. Res., A., vol. 2, p. 169. Solar System

Jiartti s Magnetism

By Walter M. Elsasser 1

An explanation of the phenomena of geo- investigations of A. , that phenomena magnetism is of interest to theoretical astro- equivalent to sunspots, but sometimes of much physics not just because the earth happens to larger extent, occur in many stars. Perhaps it be a celestial body, but mainly because to is useful to recall that the magnetic field is the reproduce such phenomena in a conventional dynamically essential feature of a sunspot, as laboratory is extremely difficult if not nearly evidenced by the fact that the mechanical impossible. Physical processes in the universe forces exerted by the magnetic field exceed all clearly fall into two classes: those that have a other forces acting in a sunspot. readily accessible laboratory equivalent (e. g., The main interest in cosmic magnetic fields spectroscopy, nuclear reactions), and those that centers on two problems: first, the way in do not (e. g., radiative viscosity, large-scale which they influence observable astrophysical turbulence). The phenomena we choose to phenomena, stellar atmospheres, interstellar designate as hydromagnetic, of which geomag- gas motions, radio noise, the acceleration of netism is the prime example, belong to the cosmic rays, and others; and second, the manner second class. The nearest hydromagnetic phe- of their generation. The first class of problems nomena beyond the confines of the earth are lies outside the domain of this paper, but is those on the sun, particularly in sunspots, but dealt with elsewhere in this volume. The their study has shown them to be exceedingly earth does furnish us with a prime example of complex. the mechanism of generation of large-scale Babcock's now classical investigations of magnetic fields; indeed, there is reason to be- stellar magnetic fields started as a search for lieve that geomagnetism lends itself somewhat magnetic fields in early-type stars. These were more readily to the construction of a consistent chosen because many of them are known to theory than do the phenomena in the solar rotate rapidly and a connection between rota- which give rise to sunspot and tion and magnetic fields was suspected. The solar magnetism, not to speak of the magnetic conditions under which stellar magnetic fields fields of stars whose details are barely, if at all, can be ascertained are extremely stringent: the accessible to observation. stars must be bright enough (>7th mag.), and The ubiquitous presence of magnetic fields the average magnetic field over the visible face in the universe leads one to accept a model of of the star must be at least several hundred the generation of the earth's magnetic field gauss. Of the stars investigated, 65 had to be which demonstrates that magnetic fields can be rejected as spectroscopically inadequate for the generated, and maintained, by inductive am- analysis, 35 yielded measurable magnetic fields, plification in an electrically conducting fluid of and in another 20 magnetic fields are suspected sufficiently large dimensions. This is the but not yet assured (as of 1953, the date of dynamo theory for which a great deal of ob- Babcock's review). As Babcock points out, servational evidence now exists. In keeping this is an extraordinarily large yield even from with the nature of this volume we shall not a purely statistical viewpoint, and the presump- dwell on the evidence in detail, but shall tion is justified that very many stars possess emphasize the basic dynamical features and magnetic fields of appreciable magnitude. their implications. There is also reason for the belief, based on the The geomagnetic dynamo is located in the > Department of Physics, University of Utah, Salt Lake City, Utah. earth's liquid core (diameter 0.55 of that of the 67 68 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS earth). There seems little doubt that this core, and the solid part of the earth, the , which seismological evidence has shown to be surrounding it. fluid, consists of molten iron. It may be shown In order to explain the observations it is by an analysis of the observed field that the necessary to assume that the mantle has a small sources of the geomagnetic are electrical conductivity. This is plausible con- located near the boundary of the core. The sidering the temperatures of perhaps 2,000- geomagnetic secular variation is a very remark- 4,000° prevailing in the lower mantle. Some of able phenomenon; it is the only effect originat- the magnetic field generated in the core on ing in the interior of the earth whose spectrum penetrating into the mantle sets up a mechan- has periods of order of some tens up to some ical force there. As a result of the action of hundreds of years, whereas all geological proc- this force, there is a small difference in angular esses are of course measured in millions, at the velocity between mantle and core, the mantle best in hundreds of thousands of years. rotating slightly faster than the top layers of The earth's magnetic field (if we ignore the the core. In the stationary state a balance is very small component that originates in the achieved between the accelerating force just ionosphere) consists of a dipole inclined by mentioned and an opposing eddy friction, the 11K° relative to the earth's axis, and of an latter probably also of a magnetic nature. The irregular part made up of all the higher terms has been observed and of a spherical-harmonic analysis. If the time analyzed; it is known as the westerly drift dependence of the irregular field is considered, because for an observer stationed at the earth's it is found that in all probability this irregular surface, the individual features of the irregular field does not have a component constant in magnetic field seem to undergo a slow average time. The dipole axis also changes in longitude motion toward the west. It has also been and latitude, but much more slowly than the shown that this motion is suffering fluctuations higher spherical harmonics. There is at present of moderate size over periods of 10-20 years. a good deal of evidence from paleomagnetism These fluctuations appear correlated with the (the study of the remnant magnetization of fluctuations in the earth's (more precisely, the rocks) to the effect that the present angle mantle's) rate of rotation which are directly between the geographical axis and the magnetic observed in positional astronomy of the moon dipole axis is larger than it was in the (r. m. s.) and planets. Theoretical analysis has led to average over past geological periods, and that the result that these fluctuations in the earth's the statistical distribution of these angles is angular velocity can be satisfactorily explained centered about the geographical axis. This in terms of a slightly variable magnetic coupling result indicates the reality of a close connection between core and mantle. The required fluc- between the earth's rotation and its magnetic tuations of the magnetic fields near the bound- field which the dynamo theory postulates. ary of the core are quite in keeping with what The secular variation (which is essentially we might expect on the basis of our general the time derivative of the nondipole component knowledge of the earth's magnetic field. The of the geomagnetic field) originates in a rela- fluctuations may be either in the over-all field tively thin layer at the top of the core. The producing a torque on the mantle or may be reason for this is that a layer of metal acts as a due to changes in the (magnetic) eddy friction screen for time-dependent fields, and the periods which counterbalances the accelerating torque. of the secular variation are such that they would In any event, this explanation of an important be effectively screened off by a layer, say, 100 astronomical phenomenon seems now to rest on km thick. Thus the secular variation cannot a sound geophysical basis. give us information about the interior of the As we go into the interior of the core there core where the essential part of the dynamo appears a type of magnetic field which has no mechanism is located, but it does teach us counterpart on the outside, the so-called toroidal something about the outer layers of the core field. In a pure toroidal field the field lines and, incidentally, about the relation of the core would be circles about the earth's axis (the NEW HORIZONS IN ASTRONOMY 69 actual field in the earth is a superposition of be sheared or twisted locally by the fluidmotio n more than one type and hence more compli- so that amplification results. It remains to be cated). The toroidal field, since it cannot be demonstrated that amplificatory processes can observed directly, had to be inferred from so- be organized in such a way that they can main- lutions of the electromagnetic field equations, tain the magnetic fields observed on the earth, but its existence does not underlie any serious sun, and stars. This turns out to be an exceed- doubt since its existence is mathematically nec- ingly difficult mathematical problem. Those essary. The toroidal field apparently is appreci- who seek a formal solution after the manner of ably larger than the dipole field in the interior existence proofs for the solutions of differential of the core, being perhaps of order 20-40 gauss. equations are likely to remain disappointed. Its counterpart also exists in the solar convec- The problem is essentially nonlinear (clearly, no tion zone, and there is reason to believe that it linear system can exhibit self-sustained ampli- provides the field from which the sunspot fields fication) and our understanding of the behavior are ultimately formed. of nonlinear dynamical systems is still in an ex- Some 20 years ago C. T. Cowling showed that tremely rudimentary state. Those making an a fluidi n which the motion is confined to merid- analysis must, for the time being, be content ional planes cannot form a dynamo. This re- with inadequate rigor and must rely to some sult may be generalized to the statement that extent on intuition, keeping in mind that there no fluid motion of rotational symmetry about are definite phenomena waiting for explanation. an axis can maintain a dynamo. There are good The present writer finds this situation rather theoretical reasons, furthermore, to believe that challenging to the theoretician, but he realizes no two-dimensional motion, that is, no motion that it does not always appeal to the more in which the fluid particles are confined to sur- mathematically inclined analyst. faces, is capable of providing dynamo action. Equation (1) applies only to a medium of Hence dynamos have an essentially three-di- infinite conductivity. In such a fluid we could mensional geometry; this makes the structure not have a steadily operating dynamo because and dynamics of even the simplest models so the continued stretching and twisting of lines complex that we prefer not to expound them of force would (as Bondi and have here in any detail. Instead, we shall confine remarked) result only in a progressive confusion ourselves to a discussion of the essential in- of magnetic fields, not in a stable pattern. gredients of dynamos, the general theorems on In a fluid of finite electrical conductivity, the which their operation is based and which make right-hand side of (1) is no longer zero. The their existence plausible. So far as we can judge, ratio of the right-hand to the left-hand side these principles are most likely to be the same is then measured by a nondimensional quantity for all hydromagnetic dynamos, the terrestrial known as the "magnetic Reynolds number," one as well as the solar and stellar ones. T» T TT (n\ The simplest formal conditions appear in the limit of infinite electrical conductivity. One can where L and V are length and velocity, used in then show that the manner known from the conventional hydrodynamic Reynolds number, and 0- is ItJ Bn d

Time By William Markowitz1 The problems of chief interest concerning year to year for the interval 1951-1954. Short- time are the precise nature of the variations in period terms due to tidal action of the moon the speed of rotation of the earth and the re- were also found. lation between the gravitational and atomic It was formerly believed that sudden changes time scales. in speed of rotation as large as several milli- The first problem deals with the basic proper- seconds per day occurred at irregular intervals. ties of the earth. The second deals with a fund- Brouwer (1952), however, has been led to the amental physical property of the universe. Al- hypothesis that the irregular changes are caused though we do know that variations in the speed by small, cumulative, random changes in speed of rotation of the earth occur, we are not certain of rotation and that the large sudden changes of the details. In fact, within the last few years do not occur. many of the ideas on the subject have been dras- Tidal friction tends to slowly diminish the tically revised. As to the second problem, we speed of rotation. Holmberg (1952) has recent- have no idea at all, at present, whether or not ly pointed out that atmospheric may in- the gravitational and atomic time scales are crease the speed of rotation, an idea due to the same; this must be decided by experiment. Lord Kelvin. Hence, we do not know whether It has been known for some time that the the earth is slowing down or speeding up. speed of rotation of the earth is subject to The quartz-crystal clock, which has replaced changes, which may be classified as periodic, ir- the pendulum clock in precise timekeeping, had regular, and secular. These variations made been developed by 1950 to the point where it the mean solar second unsatisfactory as the could reliably be used to determine the periodic fundamental unit of time for scientific and tech- terms in the speed of rotation of the earth. In nical purposes. Accordingly, in September 1955 1955 the announcement was made that an at Dublin, the International Astronomical Un- atomic standard of high precision had been con- ion redefined the unit of time as a specified structed which can be compared with astro- fraction of the tropical year for 1900.0. The nomical observation (Essen and Parry, 1955). new unit is identical with the second of Ephem- A program is now under way to obtain the Time, and is obtained in practice from ob- frequency of the atomic standard in terms of the servations of the moon. second as determined at the U. S. I shall briefly describe recent developments Naval Observatory with the dual-rate moon po- concerning studies of the rotation of the earth sition camera. If the frequency changes in the and the instruments and clocks used in deter- course of time, the gravitational and atomic time mining time. scales are not alike. An answer may be ob- The yearly and semiyearly periodic terms tained within 10 years. were found by Stoyko at Paris in 1937 and were The atomic standard will also be used to confirmed at Greenwich. Later, large changes study variations in the speed of rotation of the in these terms from year to year were indicated. earth by comparing Universal Time with atomic However, a homogeneous solution based on ob- time. servations made with the photographic zenith Thus, we shall shortly be comparing atomic tubes (PZT's) of the U. S. Naval Observatory time with astronomical time—Ephemeris and (Markowitz, 1955) showed little change from Universal—to provide information on the prob- 1 Director Time Service, T7. S. Naval Observatory, Washington, D. C. lems cited. The atomic standard apparently 73 74 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 reproduces the behavior of the cesium atom to instruction is not given at any school or observa- one part in 109, and higher accuracies will tory in time determination or meridian probably be attained. At present the fre- astronomy as currently practiced. The end quencies that can be obtained from astronomical result is that, because of the general shortage observation have precisions of several parts in of scientists and of astronomers as a whole, 109. The probable error of the differences there is a particularly acute shortage of as- between atomic time and astronomical times tronomers entering these fields. will be due ahnost entirely to uncertainties in For the past few years the Naval Observatory the latter. Hence, it is the astronomical deter- has not been able to fill vacancies in the Time mination of time which must be improved. An Service and Transit Circle Divisions with as- artificial satellite which does not move within tronomers, and the outlook for the future is not the earth's atmosphere could provide another encouraging. What is desired is that enough determination of Ephemeris Time. We must astronomers be trained so that work in these wait and see, however, what particular satellites fields can be actively pursued at the Naval are created and how well the position can be Observatory, and so that, in addition, research— determined. In general, we must look forward at least in theoretical aspects of the problems— to continuing and improving the present tech- can be pursued at some other observatories or niques in time determination. schools. We shall now consider what needs to be done. From time to time foreign astronomers and The Committee on Needs in Astronomy asks students have come to the Naval Observatory where financial aid might prove most beneficial for study and research. Such visits could also in aiding progress. Dr. Struve (1955), in his be made, with considerable benefit, by American article "The General Needs in Astronomy," astronomers and students. The minimum use- asks a more basic question; namely, should ful period would be about 2 weeks, while for more attention be given to the type of positional research and thesis purposes at least a semester astronomy carried out at the U. S. Naval would be required. Observatory, and should more research be done Hence, if money is to be spent to promote in this country with respect to the earth as a advances in these fields, it can best take the body? form of grants for study at the Naval Observa- I should like to answer these questions for tory. Grants could be made to teachers and the fields of meridian astronomy and for time writers of textbooks in astronomy and even to determination. Both types of observation pro- a few undergraduates. At the present time vide basic astronomical data which can not be money is available for research grants, but few obtained in any other way. Hence, research requests for research in these special fields will in both fields should be encouraged in order be received until there are more astronomers that results of the highest possible precision able to use the grants. may be obtained. The determinations of time and of funda- References mental positions have become highly specialized BBOUWBH, D. fields, and are carried on in the United States 1952. Astron. Journ., vol. 57, p. 125. ESSEN, L., AND PARRT, J. V. L. only at the Naval Observatory. The observing 1955. Nature, vol. 176, p. 280. techniques have been improved over the years, HOLMBEKG, E. R. R. and other astronomers in the United States 1952. Monthly Notices, Geophys. Suppl., vol. 6, have been relieved of the necessity of doing No. 6. similar work. On the other hand, these as- MABKOWITZ, W. 1955. Astron. Journ., vol. 60, p. 171. tronomers have become isolated from develop- STRUVE, O. ments in these fields. In the United States 1955. Publ. Astron. Soc. Pacific, vol. 67, p. 214. Geographical Positions

By John A. O'Keefe *

The determination of position on the surface Because he cannot measure vertical angles of the earth is the subject of two branches of with anything like the precision with which astronomy; namely, field astronomy and ge- horizontal angles are measured, the geodesist odesy. The latter is not always considered as in a mathematical world which is exactly a branch of astronomy, but Simon like the Flatland of Abbot's famous book, so in the Encyclopedia Britannica has given us a long as it is a question of geographical positions. precedent for doing so; and it seems very He knows the lengths of his lines and the reasonable on historical and logical grounds. angles between them, but he does not know the It is the function of field astronomy to pro- space configuration of the surface upon which vide the initial orientation and position for the he works—the sea-level surface and its pro- surveys which determine geographical position. longation un4er the land. The maneuvers by Geodesy then extrapolates these positions by which he succeeds in avoiding the unanswerable means of measurements on the ground to cover questions about the space situation bear a a whole country, or even a whole , remarkable resemblance to the squirming of without further reference to the stars, except the relativists when asked how their curved for orientation. The role of extrapolation in space would look from the outside. The re- geodesy is unique, so far as I know, in science. semblance is a family one, since C. F. Gauss is If it were physically possible, the geodesists the father of both systems of thought. would certainly prefer to have only a single Of course there are techniques for finding astronomical determination for the whole world heights, but these are only with relation to and to rely on geodesy for all other determina- this sea-level surface of unknown form. tions. The reason for this is, of course, that The principal application of the techniques because of the irregularity of the earth's , of field astronomy in geodesy is therefore in the direction of the vertical at any point is not determining the form of the sea-level surface of exactly what would have been expected on the the earth, for each measurement of an astron- basis of a perfectly ellipsoidal earth. So long omical position determines the slope of the sea- as we are thinking only of navigation at sea level surface at a point. The difference be- these errors, which very rarely amount to as tween the astronomically measured position and much as one nautical mile, are insignificant. the geographical position determined geodeti- But when we deal with precise land surveys, cally is, in fact, the slope of the sea-level surface we can extrapolate relative positions halfway at that point (if we assume that the slope at round the world from the measurements of the initial point was zero). From a large num- triangulation more precisely than they can be ber of such determinations of slope, the form found from on-the-spot measurements by field of the sea-level surface can be found by astronomy. integration. As a consequence, vast nets of triangulation Finally, we may also seek to find the form have developed which must, for the best results, of the sea-level surface of the earth from the be adjusted simultaneously. The feat of adjust- measurements of . There is a famous ing systems of equations that involve several theorem, called ' theorem, which con- thousand unknowns has been accomplished nects the form of the sea-level surface of the repeatedly in the last few years. earth with the measured values of the intensity 1 U. 8. Army Map Service, Washington, D. O. of gravity. 76 76 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 Thus the classical approach was to deal in determination of place on the earth. There terms of known angles and known distances on are several ways of using this fact; each of a largely unknown surface, the determination them deals in its own way with the twin of whose form was left to a small group of difficulties of the remoteness of the moon and geophysicists, since the mathematicians had the irregularity of its surface. found ways to bypass the requirement for it. The most perspicuous method is that of By extrapolation, vaster and vaster systems Markowitz, who simply photographs the moon of triangulation have been constructed, until in a star-field and measures the distances from now we have most of the Western Hemisphere points of the moon's limb to the stars. Simple on one system, and most of the Eastern Hemis- as the theory is, its practice was not mastered phere on another. until recently. The light of the moon must be reduced by a filter so that it can be observed on The new techniques the same photograph as an eighth magnitude The geodimeter is rapidly replacing the invar star, 100 million times fainter. The motion of tape for the measurement of baselines. The the moon during the time required to obtain an geodimeter is essentially the result of reversing image of an eighth magnitude star amounts to the method of getting the velocity of several miles; this must be allowed for by a light so that, from the known value of the special optical device which, in effect, arrests velocity, the distance is found. The toothed the motion of the moon relative to the stars for wheel is replaced by a Kerr cell, modulated a few seconds. The measurements are made by ultra high-frequency voltages. The new with great care with a precise measuring engine, technique has proved remarkably successful and the effects of limb irregularities are taken in trials over the last few years, since it measures out by averaging about 30-odd points. The directly the triangulation sides. The old invar result of a pair of plates has a probable error of tape method could be employed only on rela- 0.15 second of arc in each coordinate, which tively short lines, which then had to be related amounts to about 900 feet on the ground be- to the long lines of the regular chains by an cause of the moon's remoteness. To obtain elaborate base-expansion net. more precise results on relative positions, it is The employment of radar for the measure- now planned to obtain a long series of obser- ment of great distances is rapidly expanding. vations. The most useful tool at the present time is the A second method, which aims to obtain modification of the wartime Shoran, which is greater precision in the individual observations, known as Hiran. The principal new feature makes use of . Near the moments of of Hiran is the attempt to reduce the signal second and third contact, the relative positions intensity correction by the use of gain control. of the sun and moon may be determined from The Mediterranean has been bridged by this photometric measurements of the flash spec- method from Crete to Libya; the North Sea trum or of the beads. Since the limbs of from Scotland to Norway; and the Straits of the sun and moon are alined in space, there is Florida from Key West to Cuba. no disturbance from atmospheric refraction of For yet longer lines, radar techniques do this relation. In addition, the resolution not seem to work. The methods that go over theoretically obtainable far exceeds the resolu- the horizon depend on the ground wave; and tion of the best telescopes. the velocity of the ground wave depends, it The weakness of this method is that the seems, on the nature of the ground. Repeat- photometric observations are disturbed by ability is rather easily secured; but the inter- scintillations of the beads. This is a serious pretation of the results offers considerable matter because the diminution of brightness at difficulty at the present. the sun's limb is not instantaneous; hence it is The moon is now being employed for meas- necessary to fix an arbitrary level of brightness ures over the horizon. The parallax of the as corresponding to the limb. The precise moon, unlike that of any other celestial object, determination of the moment when this bright- is large enough to furnish a method for the ness is reached is troubled by the scintillation. NEW HORIZONS IN ASTRONOMY 77 The precision of a tenth of a second in arc which on geodetic problems. In aerial photography is required for the success of these methods is as a method of mapmaking, we have already thus difficult to reach even for an individual seen the entry of the three-dimensional ap- bead. For the elimination of the irregularities proach into the actual preparation of the of the lunar limb, reliance is placed on the topographic map. With the development of averaging of large numbers of points. aerial triangulation, we are witnessing in our A very similar technique starts from obser- times the gradual but steady superseding of the vations of the of stars by the moon. lower order triangulation. At present, in ad- For precise results it is not sufficient to rely on vanced mapping projects, only the first-order visual observations because of the relatively triangulation is felt to be required, unless long and uncertain reaction time of the human there is a question of cadastral mapping. . Photographic observations Even for cadastral projects, there are reports are likewise excluded because of the relatively that the necessary accuracy has been attained low efficiency of the photographic plate in by aerial methods. detecting faint stars in a short time. The only The most important change in the geodesy practical method has turned out to be photo- of the future will come, perhaps, from the electric. application of the satellite. Its use will pre- As compared with eclipses, are sent a genuinely three-dimensional problem, so relatively sharp and well-defined phenomena, that the determination of height with respect consequently there is relatively little disturb- to the center of the earth will be on a par with ance from scintillation. The precision with the determination of ground position. which the angles can be measured is of the The satellite can be employed like the air- order of 0.005 second of arc, which partly com- craft of the Hiran and flare observations. pensates for the moon's remoteness. On the That is to say, its position can be fixed from other hand, there is no chance to allow for the one set of ground stations while another set of lunar irregularities by averaging, since a single positions in a new area is fixed from the satel- occultation refers to a single point of the moon's lite. This procedure parallels the procedure limb. To remove the effects of the irregu- in flare triangulation, where the aircraft posi- larities, it has been found necessary to place tion is fixed by one group of stations by inter- the observers in such a way that all observers section while the positions of the unknown of a given occultation see the star disappear stations are fixed by resection (three-point fix) behind the same point of the moon's limb. from the aircraft. However, there is one This means that the telescopes must be port- important difference. In speaking of positions able and that special computations must be in the flare problem, we mean only the hori- made to determine the points at which they zontal coordinates; in fact, the vertical coordi- are to be placed. Not more than two occul- nate is not normally measured. In the satel- tations can be observed per month by a single lite case, the vertical angle will be measured astronomical party because of the problems of along with the horizontal angle, since the alti- movement and survey. tude above the horizon will be sufficient to The lunar methods of determining position permit this; and hence we shall have what the do not fit into the conventional picture of the photogrammetrists call the problems of space old triangulation. They do not fix either the intersection and space resection. direction or the distance between the points There is another difference, however, which involved; rather, they yield condition equa- arises from the fact that the satellite position tions involving both the horizontal and the can be successfully extrapolated forward, or vertical positions. They represent the ap- proach of a new point of view in geodesy. interpolated, so that the observations of inter- section and resection do not have to be simulta- The geodesy of the future neous, as they must be in the case of aircraft. A development that will most certainly come This greatly increases the flexibility and range with the future is the three-dimensional attack of the method. 78 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. a In addition, the satellite can be used to deter- expressed in latitudes and longitudes. Al- mine more precisely the parameters that fix the though this question could be avoided by the size and shape of the earth. This subject, which use of rectangular coordinates, as a practical belongs under another heading in this book, is matter it is not; and hence we can expect that yet germane to the discussion of geographical the establishment of definitive values of the positions because the assumed numerical values earth's semimajor axis and of flattening will of the size and shape of the earth are formally materially assist in the international coordina- involved in all surveys or maps which are tion of survey and map material. By G. M. Clemence1

The unsolved problems in celestial mechanics The heliocentric motion of a planet under the are very numerous; in fact, relatively few prob- disturbing action of another planet is a very lems have been solved with all the rigor, gen- different case, since the corresponding ratio is erality, and elegance that may be desired. so large in general that expansions in its powers The problems are of widely varying degrees are not practicable. In the methods so far used, of generality. In the most general case we may the difficulty has been overcome by allowing the study the character of the motions of any sys- time and its positive integral powers, reckoned tem of bodies, real or imagined, under any initial from an arbitrarily chosen initial epoch, to ap- conditions whatever. In the least general case, pear as factors of the trigonometric terms, out- we must be able to calculate the actual orbit of side of the arguments. Thus the terms lose their a particular body for .a limited time and to strictly periodic character. As the time in- compare it with observed positions of the body; creases without limit so do the coefficients, and our purpose in this case may be either to find the series necessarily loses its physical signifi- the body on a future occasion or to study the cance at times very remote from the initial discrepancies between theory and observation. epoch. In the applications of the method so far It probably is not possible to say that one made, the practical value is limited to a few such problem is more important than another, thousand years at most. in an absolute sense. I have selected for dis- The ability to construct a planetary theory cussion here two that seem especially interesting having the same generality as the solar' portion because of the many practical applications that of the lunar theory has long been recognized as could immediately be made, once the solutions very desirable, and several unsuccessful at- become known, and also because fast calculating tempts have been made to devise a practicable machines will be of no assistance in solving method. The failures should not deter anyone them. Thus I may disprove the common belief from taking up the problem again. The quali- that such machines can supply all the answers to fications needed are a sound basic knowledge of problems in celestial mechanics. the principles of celestial mechanics and of the The first may be called the problem of the art of practical computation, acquaintance with generalized planetary theory. The solar portion the power and limitations of electronic calcu- of the lunar theory has been brought to a high lating machines, and a readiness to discard the degree of formal perfection; the coordinates of traditional methods for new and unconventional the moon relative to the earth, as affected by ones. the disturbing action of the sun, are represented The second problem to be mentioned is that as the sum of trigonometric series, each term of of adapting the conventional planetary theory which is strictly periodic with a constant co- to automatic electronic calculating machines. efficient. Thus the series may be evaluated for By conventional planetary theory is meant the any value of the time whatever. In the develop- kind already mentioned, where the time and its ment of the series, advantage is taken of the powers appear as factors of the periodic terms. smallness of the ratio of the distances of the In spite of its limitations the theory is of great moon from the earth and sun, with expansions value and deserves many more applications being made in powers of this ratio. than it has had. The principal difficulty in adapting it to automatic calculation is that 1 Director, Nautical Almanac, TJ. 8. Naval Observatory, Washington, D.C. the most discriminating judgment is required 79 80 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOUl at all stages of the work if the amount of the sort of program that could be laid out in a arithmetic is to be kept within reasonable year the machine would still work for some bounds. For example, the actual calculations thousands of hours. Another possibility is to involved in Hill's theory of and , start and stop the machine a thousand times which occupied him for 8 years, could be per- in the course of the work in order to give the formed in a very few hours with a modern astronomer an opportunity of exercising his machine if it were practicable to incorporate judgment. But the most attractive method all of Hill's judgments in the automatic pro- would be to recast the form of the developments gram. But to devise such an elaborate pro- completely so as to avoid the numerous multi- gram might well require 8 years of detailed plications of pairs of trigonometric series, planning, with even more liability to error than which have to be integrated term by term, and was the case with Hill's own work. The to substitute for them a more easily controlled obvious alternative, to simplify the program numerical representation of the disturbing at the cost of letting the machine do unneces- forces. It may reasonably be expected, al- sary calculations, is not attractive; it appears though it is not certain, that the new method that with the most simple program conceivable would involve more numerical work than the the machine would do about a million times as old; we would willingly accept any increase up much arithmetic as is necessary, and that with to, say, a hundredfold. Meteorites By John S. Rinehart1

Meteorites are bits and pieces of extrater- The classification and description of mete- restrial matter which have survived the rigors orites is in good order. Meteorites are named of their passage through the earth's atmosphere. according to the locality in which they are Small ones and some quite large ones are found. They may be classed roughly as , recovered. They are our only ponderable stony-irons, and stones. The irons contain extraterrestrial material. As such, they are a three main constituents in varying proportions: storehouse of information to the astronomer. kamacite, which corresponds to the a-phase in The science of meteoritics embraces a wide the pure iron-nickel alloys (up to 6.2 percent diversity of technologies: , geology, nickel by weight); taenite, corresponding to the physics, chemistry, aerodynamics, astronomy, pure 7-phase; and plessite, which may consist astrophysics, mineralogy, and metallurgy. of a microscopic octahedral arrangement of The meteoriticist carries out a variety of kamacite and taenite or may be of a granular activities. He may simply collect, catalog, nature. If the contains more than and display meteorites. He may study their 6 percent nickel, kamacite and taenite bands metallurgical and crystallographic structures appear interlaced in an octahedral pattern, and cogitate on the cosmological significance known as Widmanstatten figures. Occasionally of his findings; he may measure the relative an iron will have inclusions of iron sulfide, iron abundances of chemical elements in meteorites phosphide, diamonds, graphite, and other and try to relate them to the universe as a minerals. The stony irons are of two kinds, whole; he may investigate the craters that those in which the iron is distributed as nodules have been formed by meteorites striking the and those in which veins of iron are intertwined earth, or the debris that a meteorite spewed with siliceous material. The stones consist about as it hurtled through the air. largely of siliceous material, through which In contrast to many other branches of are interspersed only flakes or fine veins of iron. astronomy, meteoritics, in a broad sense, has Meteorites have been described in a variety been seriously neglected by scientific investiga- of ways. X-ray studies have established the tors. Although much has been written about crystal structures of kamacite and taenite meteoritics in recent years, we still lack a well and the crystallography of the Widmanstatten integrated concept of the science. The goal figures. Neumann bands have been identified of future studies must be to effect this integra- as mechanically twinned regions. Gross com- tion. Before this can be accomplished, a positions and densities of a large number of multitude of specific problems will have to be meteorites are known. The relative abun- defined and solved. dances of elements, particularly the rarer ones, Where do they come from, these metallic have been found by the use of neutron activa- and stony calling cards from ? tion and dilution techniques. What are they like? How many are there? The classical method of fixing the age of a The answers to these questions will tell us a meteorite is to determine the relative amounts great deal about the origin of the solar system, of , uranium, and thorium it contains. the structure of the major and minor planets, and Originally, investigators assumed that all of the relative cosmic abundances of the elements. the helium was derived from radioactive disintegration of uranium and thorium. Re- ' Astrophysics! Observatory. Smithsonian Institution, Cambridge, MM. cently, the existence of helium generated by 81 82 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS cosmic rays has been demonstrated and allowed diagram of the iron-nickel system is not known for in age calculations. An -potassium nor have the diffusion coefficients of nickel in method, which is limited in application to iron been established. The segregation of stony meteorites, an isotopic lead composition elements on a microscale is a completely un- method, and a strontium-rubidium method are touched though most fertile field. all in use. The apparent ages of the stony Neumann bands, slickensides, distorted lay- phase of meteorites determined by each method ers, crushed nodules, cleaved surfaces—each are in substantial agreement, approximately has its story to tell if we can but read it. 4X109 yr. Iron meteorites are much younger, Mechanical metallurgy is almost completely 3X10* yr, according to the uranium-helium unexplored. method; but the tritium measures by Fireman At least 10 authenticated large meteorite on recently fallen irons indicate that some craters exist on the earth. The moon contains billions of years would have been required for some 2,000 craters believed to be of meteoritic cosmic rays to produce the observed He8. origin. These craters are important geological These age discrepancies need resolution. Is it and selenographical features. They are, more- possible, as Urey has suggested, that radio- over, positive evidence of encounters between active He escapes while cosmic-ray He remains substantial chunks of fast-moving planetary —a consequence of radioactive substances and interplanetary material. Study of the crystallizing only on grain surface? craters gives us much insight into the cosmo- Physical metallurgy has contributed sig- logical role of such physical encounters. It is nificantly, though not extensively, to our only recently that such craters have been knowledge of the origin and age of meteorites. recognized for what they are, and it is still The approach here is indirect. Consideration difficult in most instances to state unambiguously is given to the combined pressure and tempera- the velocity and mass of the meteorite that ture conditions, aging, and constraints that made the crater. Indeed, it is frequently could have produced the observed metal- difficult to know for certain that the crater lurgical structures. Especially important are is of meteoritic origin. If we are ever to the Widmanstatten figures, the local concentra- know the rate of of matter by the tion of nickel within elements of the figures, earth, we must first solve the riddles that and phase transitions indicative of pressure reside in and about the craters. We must changes. Research in this area has been ex- examine the geologic havoc that has been ploratory and fragmentary. The need for more wrought, we must collect meteoritic debris work both on terrestrial alloys and meteorites from within and without the crater, and we must is great. The influence of pressure on the phase conduct high-speed impact tests of our own. Meteors By Fred L. Whipple *

In these days of expanded research poten- their origin will remain conspicuously in- tialities, the study of meteors has matured as a complete. research area and many of the major problems In the area of two-station direct photography of yesterday are now solved; for example, the of meteors for velocities, trajectories, decelera- controversial problem of hyperbolic meteors. tion and light curves, the J. B. Baker Super- Both the precise photographic measures of Schmidt cameras in the Harvard Meteor Pro- brighter meteors and the radar measures of the gram have largely completed the basic observa- fainter meteors have failed to prove the exist- tional research in terms of their present-day ence of hyperbolic meteors. If such meteors capabilities. Analysis is being carried along exist, they represent a minority below the 1- rapidly and effectively by L. G. Jacchia, R. E. percent level. Furthermore, the photographic McCrosky, and A. F. . I must point out, and visual meteors are clearly of cometary however, that no serious effort has yet been origin except for a small fraction not exceeding devoted to the use of the television image tube 10 percent that may be of asteroidal origin. in optical meteor methods. There is every Most photographic and radio meteors, how- reason to believe that the sensitivity can be in- ever, move in direct of relatively low creased more than an order of magnitude, since inclination to the ; their aphelia tend to the transient character of the meteoric radi- lie beyond the belt and usually beyond ation is highly suited for the present-day photo- Jupiter's orbit, but many of the fainter radio electric techniques. Utilization of the image meteors exhibit quite small aphelion distances. tube, on the other hand, requires a major effort The cometary meteoric mass is proved to con- since the telescope and the receiver must be sist of extremely fragile, crumbly material that integrated into a new and specialized form. is probably of very low density. The spectra by Such a technique shows promise of reaching to P. M. Millman indicate a composition some- the ninth magnitude visually for very slow thing like that of stony meteorites, but it is un- meteors, and to the sixth apparent magnitude likely that any sizable pieces have reached the for the fastest meteors. Applied as in the Har- ground. Micrometeorites, on the other hand, vard Program, the image tube may accomplish may well represent the cometary solids. a corresponding improvement (some four mag- Although photographic and radio techniques nitudes) in the registration of persistent meteor have been exploited to an amazing degree in the trains shortly after the meteors have passed. past several years, serious gaps remain to be Thus, measurement of upper winds as a function bridged before we can reach a satisfactory of altitude in the range from 80 to 112 kilome- understanding of meteoric phenomena and of ters could be systematic and highly precise. the origin of these tiny bodies from interplane- The train type of observation is important tary space. I should like to mention briefly not only for the understanding of upper atmos- certain of these difficult observational problems pheric turbulence but also for solving the basic and devote most of my attention to the almost problem of meteoric masses. The "coasting untouched areas in laboratory and theoretical motion" in meteor trains presents to date the work. As long as these areas are unexplored, only direct method of measuring meteoric our knowledge of meteors, their nature, and masses, via their momentum transfer to the ambient air. Either photographic or image- 1 Harvard College Observatory, Cambridge, Mass., and Director, Astrophyslca) Observatory, Smithsonian Institution. tube methods should be applied with larger base 83 84 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 lines than those of the present stations, about nature, decay, and diffusion of the electron 50 rather than 20 miles, to clarify the question clouds produced by . as to the average masses associated with mete- There is every reason to hope that the ors of a given brightness and, from a more planned Earth Satellite Program of the United fundamental point of view, the question of the States and the International Geophysical Year densities of cometary solids. A single example of the National Academy of Sciences will leads to a density of only 0.05 gm/cm3! provide direct above-the-atmosphere measures It has not yet been possible to utilize the of small meteoroids in space, so far undetectable full power of the Baker Super-Schmidt meteor by any of the meteoritic methods. It is cameras for the photography of faint meteor extremely important that such observations spectra. Relatively coarse transmission rep- be made in order to clarify the relationships lica-gratings with a high concentration of light among comets, meteors, and the zodiacal light in one first order have not yet become available and the effect of corpuscular radiation on small for use on fast cameras of large aperture. particles in interplanetary space. Coupled with image tubes, such coarse gratings Although sound observational techniques for might lead to remarkable progress in the under- the study of meteors are being utilized vigor- standing of the spectra of fainter meteors and ously, the opposite can be said of the theoretical of train and wake phenomena for brighter and laboratory research which should now be meteors and might make it possible to de- underway to consolidate these observational termine the composition of cometary solids. gains. The basic meteoric phenomenon still D. W. R. McKinley at Ottawa, Canada, and remains unexplained in terms of even an A. C. B. Lovell with his colleagues at Jodrell approximate physical theory. This deplorable Bank, England, have made remarkable strides situation exists not only because the problem in measuring the velocities of faint meteors by is difficult, with many facets, but because no radio techniques. J. G. Davis has now devel- serious effort has been directed towards the oped a three-station method for measuring laboratory determinations of the fundamental radiants, and, therefore, orbits as well. Only a data required in such a theory. Desperately major effort utilizing this technique can pro- needed are measures of atomic and molecular vide the many data critical to our understand- cross sections for dissociation, excitation, ioniza- ing of the orbits and origins of meteors, as well tion, attachment, and recombination for me- teoritic and atmospheric elements in the energy as their interaction with the atmosphere. In 3 particular, it is essential that such a system range up to 10 ev. These quantities are need- be used with a number of receiving stations in ed for such elements as silicon, iron, nickel, order that observations may be made along the magnesium, sodium, calcium, manganese, and others in air, or, better, in pure atmospheres full length of meteor trails to determine ioniza- of or oxygen. Such laboratory meas- tion as a function of distance and height, dif- ures coupled with the remarkable progress of fusion effects at great altitudes, deceleration in recent years in shock-tube research and ultra the atmosphere, precise correction to outer high-velocity phenomena should make it pos- atmospheric velocities and orbits, heights, winds, sible to utilize the observations of meteors to and other important meteoric observables. establish the beginning, at least, of a satisfactory Such a program is being undertaken at Harvard basic theory for meteors. The needs in the under the technical direction of G. S. Hawkins. area of astroballistics are elaborated in a paper This technique should be carried to its prac- of this volume by R. N. Thomas. tical limit. Artificial meteors, either in evacuated ranges Fortunately, many radio groups are studying or at moderate altitudes in the atmosphere, the properties of forward-scatter by meteors could provide extremely valuable information because of its importance in communication. regarding the meteoric phenomena under con- In the near future, such information needs to ditions in which the nature and mass of the be collated, summarized, and interpreted with meteoric body would be known. Such re- a view to increasing our understanding of the search, started under the direction of J- S. NEW HORIZONS IN ASTRONOMY 85 Rinehart at the Naval Ordnance Test Station, physical quantities as the heat transfer coef- Inyokern, Calif., should be pursued vigorously ficient and the luminous efficiency. These to provide the confirmation of theories based quantities are, respectively, the percentage of upon the laboratory measures. the available translational energy of the me- In quite a different area of meteoritic studies, dium that is transferred to the surface of the laboratory measures would also be of extreme meteoric body and the percentage energy of value. Corpuscular radiation from the sun the ablated material that is transferred into almost certainty plays an important role in the observable radiation. A corresponding quan- disintegration of meteoroids in space. We tity is the ionization efficiency, measuring the require laboratory measurements of the sputter- number of electrons freed in the meteoric ing phenomena produced by protons in the process by each atom of the . These energy range 10*-10*ev on common meteoritic quantities are, naturally, functions of the surfaces—stony minerals as well as nickel-iron. velocity and very likely are functions of the air A knowledge of the rate of disintegration under density and composition of the meteoroid. these circumstances would greatly increase our The physics of the meteoric process involves understanding of the cometary meteoroid. the integration of physical data, physical Once such a body has been ejected from a phenomena, and meteoric phenomena into a , it is subjected to corpuscular radiation coherent whole. The development of such a which causes it to spiral inward towards the theory has obviously been the goal of much sun and also to disintegrate. There seems research. In one area, there is hope that such a little doubt that cometary particles are the theory can be developed without so much sub- major source of scattering and diffraction of sidiary information, viz, the theory of the per- sunlight in the solar Fraunhofer corona, in the sistent meteor train- The evidence points zodiacal light, and in the Gegenschein. We strongly to the conclusion that meteoric energy cannot integrate our information concerning is transferred rapidly to some retentive agency comets, meteors, the corona, the zodiacal light, in the atmosphere, such as active nitrogen, and and the Gegenschein until we know the effect is transformed slowly into radiation after the of corpuscular radiation on small meteoritic meteoric phenomenon is ended. Such a theory particles. will be closely integrated with the problems of Similar laboratory measures of atmospheric the physics of the upper atmosphere but does components impinging on meteoritic material not seem to be hopeless in terms of our knowl- with energies in the range 10-103ev would edge at the moment. provide a sound foundation for an understand- A number of theoretical problems fall in ing of the interaction of extremely small the realm of celestial mechanics. The problem meteoroids with the upper atmosphere. If the of the "Jupiter barrier" requires much more sputtering effects produced under these condi- attention. The Jupiter barrier arises from the tions are sufficiently small, then "micromete- perturbational disturbances by a large planet orites" can indeed pass through the upper which tend to prevent a particle spiraling in atmosphere without destruction; by black-body towards the sun from reducing its aphelion radiation at temperatures below the melting distance below the perihelion distance of the point they can radiate away the heat derived planet. Perturbations by close approaches from atmospheric encounter. Such sputtering to the planet necessarily throw the particle information would be of extreme value in into an orbit that will return near the point of interpreting the observations of meteoritic dust encounter, and thus an orbit with aphelion in the atmosphere and in deep sea oozes. It beyond the planet's orbit. Many of the would also help clarify the question of whether perturbational problems of interactions between micrometeorites provide condensation nuclei to the planets and small particles of the solar initiate heavy rainfall at extremely great system take on a stochastic character and, in altitudes. fact, are soluble by modern stochastic theory. In the laboratory-ballistic area of measure- An interesting problem combining stochastic ment, particularly needed are the values of such theory and celestial mechanics concerns E. 3S1O14 — 30 7 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1

Opik's suggestion that perturbations of aster- of problems relating to the orbits of comets, oidal material by Mars are responsible for the meteors, meteorites, and . existence today of asteroidal material crossing The author hopes that this report, incomplete the earth's orbit. Here we require a more though it is, will at least tend to eliminate comprehensive theory of the perturbative complacency with regard to the present com- effects of a smaller planet upon microparticles pleteness of meteoric research. in the solar system. This enumeration of problems in the field References of meteors is far from complete; in particular KAISBH, T. R., BD. it omits a number of important questions in 1955. Meteors. Pergamon Press, Ltd., London. the category of correlation-interpretation- LOTBLL, A. C. B. 1954. Meteor astronomy. Oxford at the Clar- theory. These questions involve the relation- endon Press, New York. ship between the orbits of comets and meteors, POBTEB, J. G. the distribution of particle orbits to produce the 1952. Comets and meteor streams. John Wiley observed zodiacal light, the nature of the Ge- & Sons, New York. genschein, the development of meteor streams, RlNEHABT, J. 8. 1954. Behavior of metals under impulsive loads. identification criteria for meteor streams of American Society for Metals, Cleveland, great diffusion or low population, and a number Ohio. Minor Planets By PaulHerget1 The techniques and methods of observing the the minor planets and to provide a basis for in- minor planets have not improved substantially vestigating theories of their origin. A very for more than half a century. Within the last important concomitant result was a nearly com- decade, however, the development of electronic plete survey of all positions, which computing machines has increased the possi- proved to be a very valuable standard of refer- bilities for computational work by a factor of ence whenever doubtful cases occurred in the nearly a thousand. Advances in technology improvement of orbits. If the computations have improved photographic emulsions so that were based on actual observations it should be we can observe planets much fainter than those possible to locate the doubtful planet on the observed formerly, but there is no immediate survey plates from its computed position and, prospect of improving methods of blinking, thereby, provide still more observational data, measuring, and reductions based on stellar com- unless the planet was below the magnitude limit parison positions from star catalogs. Minor of the plate. On the other hand, if the supposed planets will continue to be discovered more or observations were spurious it would be impos- less at random, and each one must be followed sible to locate an object at the computed po- almost individually for several months in order sition on the survey plates. The value of such to secure its orbit and to predict the following a survey is so great that it should be repeated at . least once every decade. For the ordinary mi- One of the most difficult problems for the nor planet orbits, with sufficient observations in Minor Planet Center occurs when the observer the past and with good perturbation computa- assigns an erroneous identification to the object tions, this should be sufficient to assure accu- whose position is determined. As fainter and rate ephemerides; and it may have the effect of fainter objects are observed, the likelihood of displacing at least half of the ordinary minor making such misidentifications increases greatly planet observing that now goes on. and can be avoided only by securing a confirm- There is room for a special type of observing ing observation from two to four weeks later. In program which could be undertaken by someone very doubtful cases even a third observation is who has a sufficiently large telescope, who likes needed so that an independent orbit can be to sleep through the night, and who is willing to computed, if necessary. For this same reason, work low on the horizon. The purpose would it is now increasingly important to provide be to increase the length of the observed arc of much more accurate ephemeris and perturbation newly discovered minor planets, both at the end computations than in the past. of the first opposition in which they are dis- In 1950 Kuiper launched a large-scale ob- covered and at the beginning of the next follow- serving project in which the minor planet zone ing opposition. A moderately large telescope, of the sky, about 23° on each side of the ecliptic, something like a 24-inch reflector, would be was observed in the neighborhood of opposition needed because the magnitudes will be unusual- continuously for a period of 20 months. The ly faint because of the atmospheric extinction main objective of this program was to gather at low altitudes and the disadvantageous phase fairly homogeneous data on the magnitudes of relationship nearer to superior . Also, many of the newly discovered faint planets 1 Department of Astronomy, University of Cincinnati, Cincinnati, Ohio. will be in the region of perihelion during the first 87 88 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS opposition and, due to their eccentricity, will be tions are so well in hand that it is now possible at a greater heliocentric distance in the second to turn to new and more interesting problems, opposition. the solution of many of which will be aided by Recent work at the Naval Observatory and the experience gained in the work described by Kuiper has disclosed new problems related above. An accurately computed dynamical to the physical constitution and history of the orbit of Pallas from the time of its discovery minor planets. Their solution requires con- up to the present will permit an independent current photoelectric observations of individual and highly reliable determination of the value minor planets, including changes in color, of the constant of precession. G. W. Hill magnitude, and polarization, and extending prepared a list of 13 planets that are especially over a wide range of phase angles. Some suitable for determining the mass of Jupiter, evidence already exists that two classes of provided the observations and accurate per- minor planets can be recognized from polariza- turbations extend over a period of about a tion observations. Laboratory work of a century; work on this problem has begun. similar nature with known materials and con- Another program, already well underway, will ditions should offer some clues to these prob- use the first four minor planets for determining lems. Theoretical studies are also desirable. the position of the equinox instead of observa- The Minor Planet Center has experimented tions of the sun and the major planets. This with all available types of computing machines project also requires accurate computations in recent years, and it has now concluded that of their orbits. More penetrating researches the perturbations of all ordinary minor planets may now be made upon the families of minor may be most readily handled either by the planets originally recognized by Hirayama. method of variation of vectorial constants or Electronic machines have not yet been used by Hansen's method. The computations re- as effectively for the calculation of general quired are no longer a formidable problem. perturbations as for that of special perturbations, Similarly, the problems of preliminary orbits but the problem has been attacked and will and differential corrections have been reduced undoubtedly result in progress in the near to a matter of minutes. The same results future. The availability of electronic com- could easily be attained for all comet puting machines now makes the vistas in all computations. these fields more attractive than at any The routine aspects of minor planet computa- previous time in history. Planets, Satellites, and Comets By Gerard P. Kuiper *

The planets and the satellites must be In some respects the science of planetary as- studied from distances (—1 ) tronomy has greatly changed from its distin- that are intermediate between that of the stars guished predecessor. Eight- or ten-decimal ac- (>105 a. u.) and that of the bulk of the earth curacy and a formidable theoretical apparatus (< 10~6 a. u.). Likewise, some of the problems have often given way to one-decimal accuracy they present are intermediate—related both to and order-of-magnitude arguments. Observ- astrophysics and geophysics—but this situation ing programs, which occupied the major observ- is rather recent. atories for decades, have been replaced by The following paragraphs, part of a preface isolated observations by single astronomers who to a book now in preparation, may serve as a build new apparatus and use it once or twice backdrop on which to view some of the 20th under optimum conditions. "Astronomical ac- century problems of planetary astronomy. curacy" is a compliment in an age different from ours. Until the introduction of photography with large telescopes and the birth of astrophysics, both toward Nevertheless, there has been no breakdown the close of the 19th century, planetary astronomy— of standards. What has happened is that a with some exceptions—was astronomy. The system of new branch of science was born, an offshoot the "fixed" stars was of profound interest philosoph- of young and vigorous astrophysics, which was ically; but to the practicing astronomer it was largely a convenient frame to which the complex motions of the added to classical planetary astronomy but was planets could be referred. The great discoveries in not immediately integrated with it. And so astronomy were made in efforts to interpret the plane- it happened that 10-decimal accuracy and tary motions. Copernicus, , , , 1-decimal accuracy existed side by side, in a Lagrange, Laplace, Gauss, and Poincar6, to name but field now occupied by scientists of very different the greatest, created the concept of modern while studying the planets. interests and training. The study of planetary The phenomenal growth of astrophysics, and the motions and the great problem of the stability exciting of the galaxy and the observable of the solar system rose slowly and laboriously universe, led almost to an abandonment of planetary on foundations laid by the great masters. studies. The number of astronomers has always been New computational techniques proved increas- small and for better or for worse has not followed the ingly valuable in the solution of special prob- enormous increases of physicists and chemists. Celes- tial mechanics began to look like a nearly finished lems. Dr. Clemence reviews this basic part of discipline in which further progress was hopelessly planetary astronomy elsewhere in this series. difficult. Physical observations of planetary surfaces, The program of this addition particularly of Mars, led to controversies and specula- consists of a great variety of things that at tions. More and more this branch of planetary work, including the study of the moon, became the topic first sight may seem to lack coherence. A par excellence of amateurs—who did remarkably well geophysicist, considering the task of a plane- with it. The Memoirs of the British Astronomical tary astronomer, might well be skeptical. Association became the chief record for the develop- He could reason that, since the objects of study ment of markings. Astronomers with are some 10s times more distant than the earth, large telescopes were so occupied with the engaging 10 problems of stars, nebulae, binaries, clusters, the the obtainable information might be 10 galaxy and the universe, that astronomy almost en- times smaller, and furthermore, that the ab- tirely became the science of the stars. sence of material contact or near-contact with the planets deprives the astronomer of several

1 Yerkes Observatory, Williams Bay, Wis. sources of information: seismic studies, heat 80 90 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 flux measurements from the interior, gravity and for H2). Nitrogen in our atmos- surveys, magnetic surveys, study of magnetic phere is a good example of invisibility; even at variations, , chemical and crystal- sunset, when the sun's rays pass through 320 km lographic analyses of the crust, oceanographic atm., four-fifths of which is N2, it does not add studies, detailed atmospheric studies, etc. telluric lines to the solar spectrum, even at high One might wonder what there is left for the dispersion. A second handicap affects the inert astronomer to do. or noble gases, He, Ne, A, Kr, Xe. They are Classical or gravitational astronomy, how- of great interest for the study of the origin of ever, has put the planets on a coequal basis the planets because they will not have combined with the earth in some aspects. The mass of with other elements during geologic time and the earth in absolute units is known no better must therefore always have been part of the than the masses of Venus, Mars, Jupiter, planetary atmospheres. (This statement does Saturn, , and the Moon, with Mercury not apply, of course, to radiogenic He4, He3, and Neptune closely following. Nor do we and A40.) Because these gases range in atomic know the planetary orbits with an accuracy weight from 4 to 130, we can use them—in the greatly different from that of the earth's case of the earth, where they can be observed— orbit. Hence, the radiation received from the to determine empirically how the fraction of sun is known equally well for all. The atmospheric constituents lost by evaporation of the planets can be determined with con- during the process of planet formation depends siderable precision; hence the equilibrium on molecular weight. Since the original com- temperatures can be computed quite well, position of the , from which the even if with somewhat higher precision for some planets condensed, was very probably the cos- of the planets than for the earth, whose mic composition (essentially the composition of is difficult to determine accurately. The the solar atmosphere), and since the abundance periods of rotation of the planets can be found of heavy elements, like Fe, Mg, Si, etc., may be (Venus is still an exception), often with con- estimated, we may compare the terrestrial siderable precision, as well as the planetary abundance table with the cosmic abundance obliquities. With the radiation intensity and table to obtain the desired ratio, assuming that the obliquity known, theoretical climatology no appreciable fraction of the heavy elements may be applied to the planets provided some- was lost. Now it would be exciting if one thing can be said about planetary composition. could apply this approach to other planets. This, too, is within the range of possibilities. But the very property which makes the noble It is immaterial whether the source is 109 gases so interesting in tracing the early history miles away or just in front of the slit of the of the earth (i. e., their stability in the atomic spectrograph, provided enough light is avail- state due to the completeness of their outer- able to make a spectrogram of the desired most electronic ) makes them at present dispersion during, say, one night's exposure. undiscoverable on other planets. Such high There are, however, serious handicaps, not energies are needed to excite these atoms that resulting from distance but imposed by Nature their absorption spectra are far in the inacces- itself. In the first place, some of the most sible ultraviolet. abundant atmospheric constituents are homo- It is clear, then, that two important advances nuclear molecules (H2, N2, O2), and because of may have to await the development of an their symmetry the vibration-rotation spectra extraterrestrial observing station. If planetary are forbidden. Since the electronic absorption spectra could be taken from 200- or 300-km spectra H2 and N2 are in the ultraviolet part of elevation, entirely new and vital information the spectrum that is cut off by our atmosphere could be gotten on the composition of plane- (X<0.3M), we cannot discover these gases on tary atmospheres. One would then get an other planets unless such enormous quantities undistorted picture of atmospheric composition. are present that pressure effects destroy the The handicaps imposed by Nature would have symmetry of the molecules and make their been overcome. infrared spectrum faintly visible (as on Uranus Some of this much-desired information can NEW HORIZONS IN ASTRONOMY 91 probably be obtained with a large telescope at to the equator, of slightly different reflectivity a ground observatory. By introducing com- and, sometimes, color. In addition, separate pensation methods, Hiltner has succeeded in clouds are occasionally seen, perhaps one per measuring polarizations down to a few units in decade on the average; they must, of course, be the fourth place. It should be possible, in huge to be visible, with diameters about as large principle, to do the same in spectrophotometry. as the entire earth. They seem to be caused by This would allow one to observe the Raman major eruptions from the interior and, because spectra of H3 in the Jovian planets and to make of the very pronounced differential rotation a rough determination of the H2/CH4 ratio. within the planetary atmosphere, they are soon Conceivably even the He abundance might stretched out into belts parallel to the equator, thus be found. Similarly for the terrestrial merging with the belts already present. It is planets, N2 and possibly other gases might quite interesting how the belt system of Saturn become observable, at least if they are more has changed with time. Jupiter, likewise, shows abundant than the strongly scattering CO2. It major upheavals in its cover, every decade even seems possible, in principle, that the solar or so, not at all related to its position in its ec- lines Mg II at XX2795.5 and 2802.7 A, including centric orbit, and therefore not caused by solar their variable emission components, could be heating. These two planets apparently do not observed regularly from the ground from the have a sun-induced meteorology, as do terres- Raman lines in the spectrum of Jupiter, at trial planets, but one largely governed by re- XX3163 and 3172 A. The problem is probably leased internal heat. It has been pointed out largely a question of instrumental development. by meteorologists that is much Related to the problem of atmospheric com- more efficient in stirring an atmosphere than position are at least four others capable of solar heating from the outside. empirical pursuit. They are the problems of Fascinating and unsolved problems are posed atmospheric clouds, the planetary albedo and by these observations. Do these occasional its wavelength dependence, planetary polariza- bursts come from the deep interior? Then how tion as a function of the phase angle and as a do they get through the mantle of solid hydro- function of wavelength, and the spectrum of gen usually assumed to surround the denser the infrared emission of the planet to space, core ? Or is contraction of the mantle itself the which at least for the terrestrial planets must cause ? What is the explanation of the brightly balance the absorbed solar radiation. For the colored clouds—brick red, chocolate , Jovian planets a finite fraction of the radiation blue, amber, white, etc.—observed especially emitted probably still derives from gravitational during these upheavals? Is the Red Spot a contraction and from cooling off of the planetary cloud cap over a floatingisland , and if so, what interior. floats in what? Or is it merely a long-lived The atmospheric clouds on Jupiter and Saturn vortex in the Jupiter envelope? (It is not at- are almost certainly atmospheric condensation tached to the core because its rotational speed products, as is true for most of the clouds and is variable.) These questions, in turn, are re- haze on Mars. Occasionally Mars has lated to the unsolved problem of the depth of dust clouds also, while the nature of the cloud the gaseous envelope and the hydrodynamics of cover on Venus is still in doubt. Uranus and the differential rotation on both Jupiter and Neptune probably have no visible clouds, al- Saturn. Of the greatest interest and promise is though markings have been reported on Nep- the recent discovery of a whitish cloud on Ju- tune. The reality of these markings is not con- piter emitting radio bursts at irregular inter- finned by observation with large telescopes, or vals; the power of the bursts measured at 22 Me by the model required for the atmospheres of is 10* times that of a terrestrial lightning Uranus and Neptune to account for the enor- discharge, but the spectrum is quite different. mous quantities of CH4 and H3 observed. Sat- The cloud problem for Mars is simple by urn has a general haze cover which forms the comparison, and has been largely clarified by a bottom of the visible atmosphere. It usually is combination of the brightness at different wave- divided into a number of distinct belts, parallel lengths, polarization measures, and studies of 92 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. 1 general behavior. By contrast, the identity of by implication, the surface temperature at the Venus cover is not at all obvious. Different night. suggestions have been made: droplets, Enough has probably been said to illustrate dust, and salt particles, but none of these fit the wide variety of planetary problems, many all the observational data on polarization, in of them unsolved. One could mention others: visual and infrared light, on the brightness precise diameter determinations and resulting variations with phase, on the planet's color, planetary densities; their interpretation in and on the appearance of the cloud belts. In terms of models based on the physics of matter conversations with Professors W. E. Grath of under very high compression; determination Bonn and P. Harteck of Rensselaer Polytechnic of the moments of inertia and the resulting Institute, the possibility was discussed that the refinements of the planetary models; the related particles are polymerized C3O2, formed as a problem of planetary composition; and relation result of the impact of solar ultraviolet radia- of the interior to the envelope for the Jovian tions on an atmosphere largely composed of planets. In a separate class is the problem of CO2. This possibility will be checked by labo- the dark Martian features and their variations, ratory experiments. Here is an exciting prob- and the question as to whether they are organic lem that, once it is solved, will also deepen our or inorganic. understanding of the history of the earth, so Satellites present a host of problems of their similar to Venus in dimensions and mass. own. Needless to say, the moon is foremost A curious problem is that of the planetary among them, but they are all different and radiation to space. For a planet or a satellite interesting. , with its orange color and without an atmosphere, there is no difficulty. its atmosphere of ; , with its Black-body radiation will be a good approxima- slightly yellowish color and its probable at- tion for such an object, covered as it probably mosphere, of unidentified composition; , is with fine powder or debris after its long with its bright orange color, but lack of at- exposure to a vacuum, to solar radiation varying mosphere; the satellites of Saturn; the with the rotation, to cosmic rays, and to inter- greatly different mean densities and, hence, planetary debris. But how does a planet like compositions of the satellites; the great variety Uranus or Neptune establish its temperature of albedos, from near unity for the inner balance with the absorbed solar radiation? satellites of Saturn to 0.07 for the moon; the What atmospheric constituent radiates effec- differences in albedo sometimes found for tively between 20M and 100/*, where the maxi- different hemispheres of the same body, ex- mum of the emission should occur? With ceptionally large for ; the spots on the H a homonuclear molecule and CH a sym- Jupiter satellites and on the moon; etc. The 2 4 are unique and of unrivaled metrical molecule, and NH3 and H2O frozen out, a situation arises which is not unlike that beauty in a large telescope; their snowy com- in the upper layers of the earth's atmosphere. position seems reasonably well established. One can understand why Jupiter does not have There, too, appreciable quantities of (in this such a ring. The high density of Io indicates case ultraviolet) solar radiation are absorbed, that the surroundings of Jupiter were quite but the atoms cannot dispose of this energy hot at the time Io formed; and, even if some by radiating in extensive infrared bands. The snow later condensed within the result is that the atmosphere has a hot upper of Jupiter, it would almost certainly have fringe at nearly 10 times the mean equilibrium- evaporated subsequently by solar radiation. radiation temperature of the earth (^ 250° K). Saturn's ring appears to be just safe from The thermal radiations in the radio-frequency such destruction. But why don't Uranus and range will be of special interest since they will Neptune have rings? That is very puzzling presumably arise from the solid surfaces, and and further work must be done to lower the therefore give a measure of the depths of the thresholds set by present observations. atmospheres. Thermal radio emission of Mars Comets are much stranger objects than will determine the mean daily temperature and, planets or satellites. Their ephemeral nature NEW HORIZONS IN ASTRONOMY 93 shows that they "do not belong" near the appears, which has not been very often in earth; that they were not formed there. The recent years. A good bright comet, like that exposure to strong solar heat and corpuscular of 1882, would be a major scientific event. radiation causes the comets to disintegrate Problems of evolution and origin are as com- rapidly, and show heads and long tails that are pelling in solar-system studies as they are in among the weirdest structures seen in Nature. studies of stars and galaxies. It seems as if No wonder pamphlets were written in past ages some of these are being solved. On the other on whether the appearance of a comet might hand, a good look at the problems in geophysics mean the death of a king. Comets are unad- will temper any excess optimism. Much of justed to their surroundings, as we observe what we would like to know about planets, them near the earth; for this very reason they satellites, and comets is hidden from our view; demonstrate, by their violent reactions, certain and even if we could see it, explanations would properties of these surroundings, as the old not follow automatically, as geophysics shows. and adjusted planets and satellites no longer do. Clearly, one must be satisfied with limited Comets have thus become of the greatest objectives and with explanations and theories interest in the study of the properties of solar of a general kind. But because of the many radiations and interplanetary magnetic fields, repercussions these problems have on geophysics as well as for their own sake. The unpredicta- and related sciences, any advance will also have bility of the occurrence of a "new" comet and a general interest. The limits imposed by the irregularity of the stimulating solar beams telescopic equipment and other practical con- add to the excitement when a bright comet siderations have by no means been reached.

381014—56-

Airglow and Aurora By A. B. Meinelx

The fact that the background of the moonless describe a complex flux of radiation composed night sky is not entirely dark has been known of both terrestrial and extraterrestrial sources. for centuries, while references to the aurorae The extraterrestrial component arises from: extend back into the dim recesses of Nordic (a) interplanetary scattering of sunlight (zodi- antiquity. Yet only since the turn of the 20th acal light), (b) interstellar scattering and emis- century have these phenomena been the subject sion excited by stars, and (c) the integrated of systematic study. The presence of an radiation of individual unresolved stars. The emission spectrum for the aurora was recognized atmospheric component of the light of the night as early as 1867 when Angstrom observed and sky arises from: (a) scattering of starlight and measured the wave length of the green auroral extraterrestrial radiation in our atmosphere, line. The presence of an emission spectrum (b) atomic emissions in the form of mono- from the light of the night sky was noted as chromatic emission lines, (c) molecular emis- early as 1895 by W. W. Campbell, who used a sions in the form of emission bands and possibly new visual spectrograph at Lick Observatory continua, and (d) Cerenkov radiation and + many years before the spectrum was redis- N2 emission from cosmic rays. Assigning to covered photographically by V. M. at each source a percentage of the total light of the Observatory. The light of the night sky night sky can only be done with uncertainty since as a detached phenomenon of physical interest it is difficult to separate clearly some effects. was recognized by the pioneering investigations In the visible region the ratio of extraterrestrial of in 1910, V. M. Slipher in 1919, Lord to terrestrial must be approximately unity. Kayleigh in 1920, and in 1923. For a Upon certain occasions aurorae are super- historical account of early night sky and auroral imposed on the night sky. The auroral radia- research, the reader is referred to the articles by tions are caused by two basic mechanisms: Dufay (1928), Fabry, Dufay, and Cojan (1934), (a) heavy-particle excited atomic and molecular Dejardin (1936), (1942), Dufay (1938), emissions (in the homogeneous arc phase), and and Swings (1949). (b) electrical-discharge excited atomic and Many young scientists may wonder why molecular emissions (in the rayed-structure airglow and aurorae are included in a report on phase). However, most of the light may be new horizons in astronomy. Historically, both produced by secondary electrons in both cases. phenomena are noteworthy because of their To describe the appearance of the light of astronomical consequences, although both arise the night sky on the spectrum of a faint astro- solely within the gaseous envelope of the earth. nomical object, such as a distant nebula, it is To most astronomers the light of the night sky best to portray each spectral region separately. is a nuisance to be tolerated by necessity, since In the region sensitive to the 1030^-0 emulsion, little can be done about it. However, in spite the XX4800-3100 spectrum is characterized by of the unpleasant problems they make for the many bandlike features superimposed upon a observer, the light of the night sky and the continuum which is strong in the blue and aurorae are phenomena of wide astronomical rapidly disappears shorter than X3900. As a and astrophysical interest. consequence of this spectrum it is difficult to The term "light of the night sky" is used to detect weak astronomical features such as the broad, but shallow, H and K lines in the spec- 'National Astronomical Observatory, and Yerkes Observatory, Williams Bay, Wls. trum of faint galaxies. 05 96 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. 1 The region

Solar-Terrestrial Relationships: Weather and Communications By Walter Orr Roberts*

Relatively few of the many branches of temperature, depend very much on local ele- astronomy can be said to have applications of ments such as topography and latitude; a direct practical importance. The field of solar- worldwide circulation change that produces terrestrial relationships is an exception. Two rain at one spot may produce drought in decades of practice have shown the radio another. The ties of sun to local weather are operator that solar variations can make the therefore most complex. difference between success and failure in getting The mechanism responsible for responses of a message through. Today, we may look for weather to solar activity clearly does not in- improvements in practical weather forecasting volve substantial changes of the sun's total because of new discoveries made on the common energy; rather, it involves changes in selected border between meteorology and astronomy. spectral ranges where only a relatively small Just how far away these advances are will portion of the sun's total energy output is depend, at least in part, on the magnitude and concentrated. It is also clear that the most the quality of the effort made in this field. pronounced weather responses are likely to be The effects of the sun on radio communica- found at the upper levels of the earth's atmos- tions are well known, even if poorly understood. phere, where the sun's variable energy is The influence of solar activity on weather is principally absorbed. However, meteorolog- less apparent, and the reality and nature of the ical observations at such levels are still rather effects are still somewhat controversial. Enough sporadic. Improvements in all types of weather clues exist, however, to convince a substantial forecasting are likely to come from research on group of research workers that changes in the the nature and causes of persistence and sudden sun's shortwave emanations and in its corpus- changes in the dynamics of the general circula- cular radiation are reflected in worldwide tion of the atmosphere. The motions of the weather patterns, and that practical gains in upper levels and the possible nature of the both long- and short-term rainfall forecasting mechanisms that start disruptions in the pat- and temperature forecasting for large areas of terns of momentum and mass transport in the the earth can be expected when these connec- earth's atmosphere require particular attention. tions are better understood. The fact that the Dynamic meteorology is still in its infancy, but sun exhibits rather stable long-term trends the solar astrophysicist has much to contribute. improves the prospects for season-in-advance Ten years ago there seemed no real prospect weather forecasts based on variations in solar that major advances in long-term weather activity. forecasting would result from solar observa- The sun's effects on weather obviously must tions; today they promise advances of in- be on a large scale. Several factors in the so- calculably vast practical significance. called "general circulation" of the earth's Without much exaggeration it can be stated atmosphere give evidence of such solar in- that modern solar-terrestrial research began fluences. The results of these large-scale when R. C. of the Royal Observa- changes, in terms of surface rainfall and tory in England made his classic observation of • High Altitude Observatory, Boulder, Cote. a great solar flare visible in integrated light at 100 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. 1 11 a. m. on September 1, 1859. At the instant in a rather clear-cut way to changes in the of the flare, a fluctuation occurred in the character and level of solar activity as measured, strength of the earth's magnetic field. On for instance, by the average intensity of the the 2 days following, a violent magnetic storm emission of the sun's green coronal line. Al- was recorded by magnetometers at the Kcw though our understanding of the solar physics Observatory. Balfour Stewart, the director of the situation is very imperfect, we infer that at Kew, concluded that the terrestrial magnetic the observed large changes in the coronal disturbances were related to the flare; however, emission probably signify large changes in the the idea was so unbelievable that most research total ultraviolet or X-ray emission of the sun. workers dismissed the sequence of events as a Nonetheless, the total changes of energy of coincidence. the sun's emission must be very small, probably The connection between flares and magnetic measured in thousandths or millionths of the storms is now amply demonstrated, although . Not only do the world weather the problem remains of explaining just how a changes require vastly greater energy than the solar flare produces these effects and the even sun's variations can supply but terrestrial more spectacular phenomena in the aurora physics suggests that the level at which the and ionosphere. In the years since Carring- X-ray or ultraviolet energy is absorbed must ton's great flare, a whole new science has be so high that it is difficult to imagine that it developed to solve the problems common to could exert a significant influence all the way solar physics and to terrestrial atmospheric down into the weather sphere, far below. Yet physics. the problem stands, a challenge and a mystery. Both sides of this joint attack have greatly No one can doubt that real progress in fore- enriched our observational knowledge. The casting will come when we have a satisfactory spectroheliograph, the coronagraph, and the physical explanation of the link, whether it solar radio telescope have been the astronomical be a trigger mechanism for amplifying infinites- milestones. Geophysically, the night sky spec- imal solar changes or some yet unknown cause. trophotometers and ionospheric sounding de- The multitude of phenomena of the iono- vices have competed in importance with earth- sphere and the many variations of the effects based, airborne, and rocket-borne cosmic-ray with latitude and longitude are equally enig- counters, with photometers, and with magne- matic problems. Few astronomers or geophysi- tometers operated with ever-increasing regular- cists today doubt that the earth is subject to ity from an ever-increasing diversity of stations. highly irregular and unpredictable "showers" of Great advances, particularly in meteorology, solar corpuscles, or that these corpuscles control have come from the simple fact that we now important effects in aurorae, the ionosphere, and can prepare daily or twice daily hemisphere- earth magnetism. Yet, in spite of the brilliant wide charts of airflow and temperature dis- victories of the Chapman-Ferraro theory of such tribution at a wide range of altitudes from effects, we know that the theory is only a crude which large-scale dynamical studies can be beginning, and that it is grossly inadequate to made. account for major phenomena that are well ob- In every facet of terrestrial atmospheric served. We particularly need new research to physics, tantalizing connections between sun improve our predictions for communications in and earth have been found. But many pieces polar latitudes. of the jigsaw puzzle remain to be put in place. Many simple, but major, questions of solar- Advances of great importance in ionospheric terrestrial relationships are yet unanswered. radio communications and in meteorology will For example, do cosmic rays come from the sun ? undoubtedly appear when we have a satisfac- A few short years ago it was hard to find a work- tory theoretical explanation that fits together ing cosmic-ray physicist who would admit that the many known solar-terrestrial relationships. the sun as much as gently modulated the cosmic A specific example may help to illustrate rays that were believed to come in all directions the expected course of progress. We now know from outer spaqe. Astronomers generally as- that world weather patterns sometimes respond sumed, without any serious doubts, that per- NEW HORIZONS IN ASTRONOMY 101 haps well over half of the energy of the universe understanding of the sun's radiational and cor- was bound up in these mysterious high-speed par- puscular emissions, particularly those emissions ticle emissions. Now the view prevails that a stopped high in the earth's atmosphere. At major part of the low-energy cosmic rays come the moment, it is hard to say whether the solar from, or are at least strongly controlled by, the physicist or the rocket and satellite launchers sun; and how many of the more energetic cosmic will be first to specify the nature of the sun's rays also come from nearby, rather than from variable emissions as seen from an altitude, let reaches of space, is an open question. The observational-theoretical picture of solar us say, of 250 kms. In any event, the ad- activity needs integration. The vast complex vances in the decades immediately ahead will of sunspots, solar magnetic fields, flares, plages, be spectacular, and the practical stakes in long- prominences, and the corona needs to be welded and short-range weather forecasting may well into a consistent and understandable unit. surpass the very important gains that can From such a unified picture of solar activity, reliably be expected in the field of communica- and of the quiet sun as well, will come better tions forecasting.

Solar Physics

By Leo Goldberg1 and Donald H. Menzel3

At the present time, astronomy is in a re- such agencies supports a wide variety of studies markable period of rapid growth wherein the and is frequently responsible for so-called rate of accumulation of knowledge is higher "seed" research, in which the expenditure than at any time in the past. In such a of relatively few thousands of dollars may lead dynamic situation, existing facilities rapidly to the germination of a whole new field of become obsolete and the new instruments study. The next step in the nurturing of the required for continued development of the seedling may well require the construction of science are generally larger and more expensive equipment costing hundreds of thousands or than their predecessors. Unfortunately, the even millions of dollars. It is at this point traditional sources of financial support of large that planning on the national level must enter astronomical facilities have been seriously to evaluate the potentiality of the study with diminished. Few private donors are in a reference to astronomy as a whole. It would position to contribute gifts of the magnitude be unfortunate indeed if the question of cost necessary for modern research installations, should be a major factor in any decision against and the larger foundations appear to be con- embarking in some new, exciting, and otherwise centrating their primary efforts in other areas. worthwhile project. This country is capable Thus, apart from occasional gifts or contracts of extending far greater financial support than from industrial concerns, the burden of support it has in the past to provide an enormously must fall upon Government agencies, which expanded effort in basic research. The really can function wisely only if they are guided by serious bottleneck in astronomy is the current the collective, considered judgment of re- shortage of topflight astronomers. We must sponsible scientists. This means that as- therefore guard against wasting precious tronomers as a group must assume the re- scientific manpower in large but relatively sponsibility for deciding which large-scale unproductive projects, especially those fi- facilities are urgently needed for the continued nanced by public funds. The shortage of development of their science, and they must astronomers is a national problem which should also solve the organizational problems con- be a major concern of the National Science cerned with the cooperative use of these Foundation. facilities. No branch of astronomy has progressed more We do not imply that all future astronomical rapidly since 1940 than has solar physics. research should be directed by planning com- We do not except radio astronomy, which, as mittees. On the contrary, we strongly believe a special technique of tremendous power that the best new ideas and the of rather than a separate branch of astronomy, new fields of investigation will come from has supplemented optical methods in contribut- individuals or small groups of investigators ing importantly to virtually every field of who must continue to receive the strongest solar research. possible financial support from the National Following a period of phenomenal growth at Science Foundation, the Office of Naval Re- the beginning of the century, sparked by the search, and other Government agencies. The invention of the spectroheliograph and the solar regular program of grants and contracts by tower and by improvements in spectroscopic gratings, solar research progressed rather slowly ' The Observatory, University of Michigan, Ann Arbor, Mich. • Harvard College Observatory, Cambridge. Mali. through the 1920's and early 1930's. Largely 103 104 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS through the efforts of George Ellery Hale, and the chemical nature of the atmosphere. observational solar physics had advanced much The temperature gradient, in turn, must be more rapidly than the theoretical interpretation sufficient to drive the solar radiation through of the data. This gap, however, began to close the semiopaque gas. In an idealized atmos- in the late twenties, with the development of phere, where the atmospheric gases are arti- theories of ionization equilibrium and of atomic ficially stirred and where no heat is transferred structure. from or into the moving volume, the tempera- Although by 1930 a certain number of suc- ture gradient must follow the so-called adiabatic cesses had been achieved from the application law. Such an atmosphere is in neutral equi- of the new physical theories to the observa- librium. Exchanging equal masses of gas at tions, they were far outnumbered by many different atmospheric levels produces no effect conspicuous failures. In 1930, numerous road- on the temperature distribution. As the mass blocks filled almost every avenue of solar from the lower level rises and expands, it physics. It is useful to survey these barriers assumes the same volume, pressure, and from the vantage point of 1956, both to note temperature as that of the mass it replaces. those that have already been removed, or are in As long as the temperature decreases upward the process of being eliminated, and to assess at a rate slower than that specified by the how far we may now expect to move down the adiabatic condition, the atmosphere will be road before new obstacles appear. stable. Radiative transfer will predominate and the flux of energy will be essentially con- The stant over the entire solar disk. However, if One of the most fundamental problems in the natural gradient is steeper than the solar physics concerns the structure and adiabatic, a mass of gas displaced upward circulation of the gaseous layers comprising becomes warmer than its surroundings. It will the photosphere, the region in which the continue to rise, like a hot-air balloon. Analo- continuous spectrum is formed and in which gously, a mass displaced downward becomes most of the Fraunhofer absorption takes place. cooler than its surroundings and thus descends. The solution of this problem calls for the con- An atmosphere, so constituted, becomes un- struction of a model photosphere in which stable and convection sets in. the chemical composition and the distribution When convection is mild, the energy con- of temperatures, pressures, gas velocities, and veyed upwards will essentially equal that con- magnetic fields are completely specified every- veyed downwards. Transfer by radiation may where in the photosphere. In theory, we can still predominate. However, as convection derive all of this information from observations increases, nonlinear effects enter so that we no of the continuous spectrum and the Fraunhofer longer have a precise balance. We shall then lines. Practically, however, the problem is have a net outward transport of energy by frightfully complicated. For example, mea- moving airmasses. The convective and radia- surements of solar limb darkening and spectral tion fields may be said to "couple" together. energy distribution will yield the distribution The at the outer boundary will of temperature with depth if we know what vary from point to point over the solar disk in processes produce the continuous spectrum. such a way as to maintain a constant average. We can then calculate the pressure distribution The convective field will carry momentum as if the atmosphere is in hydrostatic equilibrium. well as energy and thus will produce either dimi- This assumption is certainly wrong if the at- nution or increase of the effective gravity, with mospheric gases are in violent, turbulent motion consequent distention or compression of the caused by convective instability or electro- atmospheric gases. magnetic forces. To construct an accurate Unsold has shown that convective instability model we must evaluate the relative contribu- tends to exist in two separate levels of the solar tions of radiation and convection in the atmosphere, one below the photosphere and the transport of energy through the atmosphere. other near the top of the reversing layer. The Stability depends on the temperature gradient lower one arises from ionization effects of hvdro- NEW HORIZONS IN ASTRONOMY 105 gen, the dominant chemical constituent of the the assumed atmospheric model. Third, we solar atmosphere. If a volume consisting must know or determine what processes cause largely of ionized gas is displaced upward, it the broadening of the spectral lines. The most expands and cools. The cooling effects lead to important broadening mechanisms include ran- electron capture. The released energy raises dom Doppler effects caused by thermal agita- the temperature above the value prescribed by tion or by macro- or micro-turbulence, colli- the adiabatic relation, and thus we find a hydro- sions, Stark effect, and radiation damping. gen convective zone. The second zone is Widening from hyperfine structure and Zeeman associated with the region where the atmosphere effect may not always be negligible. The is transparent except in the absorption lines. parameters necessary for calculating the Stark The "blanketing effect" of the absorbing atoms and collisional broadening are very difficult to again changes the temperature gradient in come by, theoretically or experimentally. such a way as to induce convective instability. Twenty-five years ago only a rough beginning When we turn to the Fraunhofer spectrum for had been made towards the quantitative inter- information on chemical composition and gas pretation of the solar spectrum. Virtually all motions, we encounter even more difficult analyses of the Fraunhofer spectrum were problems. The shape and intensity of a solar based on the idealized assumption that the absorption line depend in a very complicated continuous spectrum originated in a sharply manner on a variety of factors, of which the bounded radiating surface, the photosphere, and abundance of the responsible element is only that the lines were produced in a cooler super- one. The "f-value," "line strength," or ab- imposed layer, the reversing layer, which was sorptive power of an atom for a given radiation supposed to be at constant temperature and is also significant. The population of atoms in pressure. Only very crude guesses were avail- the two levels responsible for the line formation able for most f-values and line-broadening is also extremely important. The Boltzmann parameters, and accurate intensity measure- equation fixes these populations if the medium ments had been made for but a mere handful of is in thermodynamic equilibrium. The temper- solar lines. Nevertheless, H. N. (1929) ature is the determining parameter. Likewise, accomplished a remarkable feat in achieving the Saha equation, under similar conditions, what he has called the "zero-th approximation" fixes the amount of dissociation. However, the to the chemical composition of the sun, espe- very fact that energy is flowing through the cially in view of the fact that only recently have medium requires departures from this equilib- astrophysicists displayed any general agreement rium condition. An exact determination of the that the "first approximation" has finally been significant populations theoretically requires reached. advance knowledge of all the atomic constants, No one doubted the existence of large tem- such as areas for absorption of radiation or perature and pressure gradients in the solar collisional excitation, and the appropriate photosphere, but attempts to construct a model transition probabilities. The equations of sta- solar atmosphere were frustrated by a total lack tistical equilibrium are most cumbersome and of knowledge concerning the origin of the con- difficult to solve. Further progress would tinuous spectrum. Calculations by Menzel and seem to depend on the development of new Pekeris (1935) gave too low a figure for the techniques for the obtaining of approximate absorption by transitions involving the negative solutions near thermodynamic equilibrium. hydrogen ion. In 1938 Wildt (1939) noted the We need to know the mechanism that the existence of a bound state of H —, and suggested atoms employ for removing energy from the line. that photoelectric ionizations and free-free The process may involve pure absorption, transitions of this negative ion were the major coherent scattering, or noncoherent scattering, causes of the sun's continuous opacity. Wildt's and it is usually not obvious which mechanism suggestion was confirmed by the work of or combination of mechanisms operates for a Chandrasekhar and associates (Chandrasekhar given line at a given height in the atmosphere. and Breen, 1946) and Chalonge and Kourganoff The character of the line depends critically on (1946). 106 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. 1 Numerous calculations of model solar photo- oms and negative hydrogen ions. Errors evi- followed. Even now, we must admit dently exist either in the model or in the theory that the photospheric model is known only in of the continuous absorption, or perhaps in both. the first approximation. Although most Third, no spherically symmetrical model of the workers agree on the temperature distribution photosphere appears capable of representing the in the intermediate layers, the structure of the center-to-limb variation in the intensities of deepest and uppermost layers remains uncer- both low and high excitation lines of Fe I tain. Since most of the radiation from these (Bohm, 1954). Nor can the center-limb vari- regions is absorbed by the overlying gases, we ation in the CO line intensities be accounted for must use theory rather than direct observation on such a model (Newkirk, 1953). Similar re- for the analysis. In consequence, the location sults have been obtained by Elste (1955) at and extent of the convective zone are not Gottingen and by others. precisely known. Indeed, the magnitude of The evidence is accumulating that the line- what appear to be convective processes in the intensity observations can be reconciled with outer atmosphere suggests that even the inter- theory only when account is taken of temper- mediate layer may not be free from effects of ature and velocity fluctuations in the photo- the hydrogen zone. Radiation from the upper sphere and chromosphere. The recognition of layers can best be observed at the extreme limb temperature fluctuations is not new, since they of the sun, where terrestrial atmospheric seeing are reflected in the brightness fluctuations of hampers observation. The blanketing effect of the granules, but their influence on the Fraun- the absorption lines and possible departures hofer spectrum represents a second-order effect from thermodynamic equilibrium further com- which could not be detected before the advent plicate the problem. of modern techniques of observation and analy- Additional difficulties of even more basic sis. It is now generally agreed that the phe- character stand in the way of a definitive model nomenon of the granules is a consequence of of the photosphere. Just a few years ago it turbulent convection in the photosphere and seemed possible that a straightforward analysis that, on the average, the bright and presumably of measurements of solar limb darkening and hot granules are ascending and the cooler, inter- absolute spectral energy distribution would de- granular regions are descending in the photo- termine an accurate solar model. It was hoped sphere. If the convection is sufficiently violent, that the model could be tested and further re- however, coupling of the mechanical and radi- fined by calculation of absorption-line profiles ative fields may occur, with nonlinear effects. in terms of a specific chemical composition of The optical thickness of the atmosphere might the atmosphere. For several reasons, this hope vary significantly between the bright and dark now seems to have been optimistic. First, no areas, with an appreciable change of line in- existing model has succeeded in reproducing in tensity. detail the profiles of strong absorption lines. In Another consequence of the turbulence is the particular, the observed central intensity has excessive broadening of the centers of the always been greater than the calculated, but the Fraunhofer lines as compared with that to be discrepancy has usually been attributed to in- expected from thermal at a perfections in the theory of line formation. temperature of 6,000° K. The root-mean- Various devices such as fluorescence effects and square turbulent velocity derived from indi- noncoherent scattering have been employed to vidual line profiles and from the curve of growth fill in the line center, but none has proved quite is about 2 km/sec. Richardson and Schwarzs- satisfactory. Second, the empirical dependence child (1950) supposed that the radial velocities of the continuous opacity upon wavelength, de- of individual granules could be obtained from rived from Pierce's (in press) latest measures of the measurement of minute Doppler shifts from limb darkening, differs significantly from the point to point along the length of Fraunhofer values calculated in terms of currently accepted lines, under conditions of high spectroscopic electron pressure distributions and the theory of resolution and exceptional seeing. Careful continuous absorption by neutral hydrogen at- measurement of selected Mount Wilson plates NEW HORIZONS IN A8TRONOMY 107 did indeed reveal irregularities in the lines, may be expected to reduce exposures to a small which, upon measurement, yielded a root-mean- fraction of a second. Such a reduction would square velocity of about 0.4 km/sec. This effect greatly reduce the effects of atmospheric was interpreted as arising from the macroturbu- scintillation and improve our determinations lence of relatively large elements, the higher of the lifetimes of small-scale solar phenomena. velocity of 2 km/sec being attributed to the Finally, it is already apparent that spectroscopic of elements smaller than the resolution has far exceeded the image resolution thickness of the "reversing layer." and that limitations in external seeing quality It had been supposed that the observation will seriously hinder the discovery of the details of the random Doppler shifts was limited by of solar atmospheric structure. the resolution of the solar image; i. e., by the Poor external seeing is caused both by dis- external seeing. Recent work by McMath turbances inherent in the atmosphere and by (1956) and McMath, Mohler, and Pierce local disturbances of the air produced by solar (1955) at the McMath-Hulbert Observatory heating of the telescope and mirrors. The Nation- has revealed, however, that limitations in al Science Foundation panel for a national as- spectroscopic resolution are at least of equal tronomical observatory, under the chairmanship importance and that air turbulence within the of R. R. McMath, has recently considered this spectrograph may "wash out" the small problem and has concluded that two courses Doppler effects. On McMath-Hulbert photo- of action should be taken. First, the present graphs made with the vacuum spectrograph, site survey for the national observatory should all photospheric lines show a characteristic be extended to include observations of "zigzag" appearance even when the seeing as well as nighttime seeing. Second, an in- quality is less than average. Further, when vestigation of instrumental seeing should be the seeing is good, the lines fluctuate in in- undertaken at the earliest possible moment tensity from point to point along the solar with the goal of designing a in disk. Thus the temperature and velocity which heating effects would be minimized. fluctuations in the solar atmosphere may now It is probable that the solution of the problem be directly observed in the Fraunhofer spectrum, of telescopic seeing is now the most important at least for elements about 2 seconds of arc in instrumental challenge in the field of solar diameter or larger. physics. Several inferences may be drawn from these The vacuum spectrograph or other spectro- results, the most general of which is that the graphs that may be designed to eliminate interpretation of photospheric and chromo- instrumental turbulence will initiate a new spheric observations on the basis of spherical phase of research in solar spectroscopy that will symmetry should now be abandoned, and the take many years to run its course. The end observational emphasis directed towards de- result should be a much finer elucidation of termining the nature of the velocity and the structure of the solar atmosphere, including temperature fluctuations in the solar at- its composition and circulation, than is possible mosphere. The impact of this new develop- at present. As a valuable byproduct we may ment upon the National Science Foundation is also expect a better understanding of the likely to be very great. First, important funds mechanisms of absorption-line formation. As will be required both to modernize existing previously mentioned, the fact that the observed solar spectrographs and to construct new ones. central intensity of a very strong line is always Second, an enormous quantity of new and greater than that calculated from theory has valuable data will become available for analysis, been a major difficulty. However, it is already which will necessitate funds for assistance in obvious from the vacuum spectrograms that reduction and analysis. Third, new impetus some if not all of the discrepancy between will be provided for the development and theory and observation will vanish when we refinement of image-tube techniques. At pres- allow for fluctuations in observed central ent photographic exposure times are not less intensity, which have generally been blurred than 10-20 seconds, whereas the image tube by poor seeing inside the spectrograph. In 108 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. 1 addition, theorists will have to include effects optical spectrum during eclipses, from motion of mechanical as well as radiations! transfer pictures in Ha obtained with large corona- of energy. graphs, from observation of the radio emission Finally, it may not be out of place to point at high frequencies, from observation of the out that if temperature fluctuations are present central intensities of certain strong Fraunhofer in the solar photosphere they undoubtedly occur lines and of the XI0830 absorption line of He I. also in stellar atmospheres, especially in the from observation of emission lines in the rocket highly turbulent atmospheres of the supergiant ultraviolet, and from observations of the earth's stars. ionosphere. Interpretation of the observations from the conventional point of view of spheri- Chromosphere and corona cally symmetrical chromospheric models leads By definition, the beginning of the chromo- to serious discrepancies. In particular, one set sphere is the upper boundary of the photosphere. of observations has seemed to require a tempera- It is also the transition region, perhaps 20,000 ture for the low chromosphere in the neighbor- km thick, in which the electron temperature hood of 5,000°-6,000° K, whereas another set increases from its photospheric boundary value seems to demand much higher temperatures of of 4,000°-4,500° K to 1,000,000° K or higher, in about 20,000° K or greater. the inner corona. It is therefore of strategic A possible path out of the dilemma was importance, since within its confines probably pointed out in 1949 by Giovanelli (1949), who occur the processes that maintain the corona at suggested that the chromosphere is nonuniform its high temperature. More than 20 years ago and that it consists of alternate "hot" and Menzel (1931) called attention to the anomaly "cold" columns. Hagen's (1954) measurements posed by the coexistence in the flash spectrum of the center-limb variation of the radio emis- of emission lines of both neutral metals and sion also point to a model of this general type. ionized helium. Menzel pointed out that the Strong support for the nonuniform model has observed intensity of He+X4686 could not be recently been provided by the detailed analysis explained by a temperature of 6,000° K even if of flash spectra by the Harvard-High Altitude the chromosphere were composed of pure he- Observatory group (Athay, Evans, and Roberts, lium, and suggested that the required ionizing 1953; Athay, Billings, Evans, and Roberts, radiation might come from a hypothetical 1954; Athay and Roberts, 1955; Pecker and ultraviolet excess in the solar spectrum. Athay, 1955; Matsushima, 1955; Athay, Men- The problem posed by Menzel is still para- zel, Pecker, and Thomas, 1955; Athay and mount for theories of chromospheric structure, Thomas, 1955, 195Ga, 1956b; Athay and Men- even though the solution in terms of an ultra- zel, 1956; Billings, , Evans, and Lee, violet excess has been ruled out. Since the 1954) and by studies of the structures of 1930's, at least three important developments Fraunhofer lines at the McMath-Hulbcrt Ob- have paved the way for a renewed attack on the servatory (McMath, Mohler, Pierce, and Gold- chromospheric problem. First, the discovery berg, 1956; Mohler and Goldberg, 1956). of the high-temperature corona inevitably re- Although the exact geometry and physical quires the existence of large temperature gradi- properties of the hot and cold regions have ents somewhere in the chromosphere. Second, not yet been clearly established, interesting the observation from rockets of the far ultra- working hypotheses have been suggested whicli violet solar spectrum has shown that the hypo- clearly indicate the lines along which future thetical ultraviolet excess does not exist. researches should be planned. Observations of Third, the observations clearly show that the the flash spectrum at future eclipses will con- chromosphere is not a static phenomenon and tinue to play a leading role in the investigations that hydrodynamic or aerodynamic processes of the chromospheric structure, but they may play an important role in fixing its physical now be supplemented by additional new tech- state. niques which have recently been developed. Evidence bearing on the structure of the Of great importance are observations of micro- chromosphere comes from observation of the wave radio emission, which should be carried NEW HORIZONS IN ASTRONOMY 109 out with large antennae of the highest possible no longer follows Maxwell's law. Instead, resolving power. The accurate measurement the equation represents the velocities as lying of emission profiles of solar lines in the rocket between the two extremes of isotropic scattering ultraviolet will also be crucial for the testing of without change of velocity, or of reemission chromospheric models. Finally, the Fraun- under thermodynamic equilibrium at the tem- hofer spectrum itself constitutes a rich but perature described by the root-mean-square largely neglected source of chromospheric velocity. The equation bears a close resem- information. The interpretation of the limb blance to the Milne- equation of spectrum is complicated by the two-dimensional radiative transfer, with solution between the integration, both radial and tangential, of the extremes of monochromatic radiative equil- chromospheric emission. On the disk, the line ibrium and local thermodynamic equilibrium. of sight integration is only one dimensional. The solution of the transfer problem, espec- Indeed, the complex structures of the X10830 ially when electric currents and magnetic helium line and of the cores of the hydrogen and fields are present in the medium, can rarely ionized calcium lines, as revealed by the vacuum be expressed in terms of known functions. spectrograph, are undoubtedly a consequence Instead we shall have to determine families of of the temperature fluctuations in the chromo- solutions, with ranges of boundary values and sphere and should be analyzed from this point initial conditions. In general, these calcula- of view. tions will require the use of high-speed comput- The dynamic nature of the chromosphere is ing machines. The National Science Founda- graphically portrayed by the Ha motion pic- tion may well be called on to furnish funds for tures recently obtained by R. B. Dunn of the development of models of the dynamic Harvard and the Upper Air Research Observa- chromosphere. tory with the 15-inch Sacramento Peak corona- Closely associated with structural problems graph. These remarkable photographs record of the solar atmosphere is the question of origin the rapid changes that occur in the irregular of the solar radio emission. It is well known structure of the outer chromosphere. Since that an ionized gas, with specified electron the chromosphere is basically a dynamic phe- density, possesses the so-called "plasma" oscilla- nomenon, with kinetic transport of both tion frequency equal to the "critical frequency" energy and momentum, a true model of this at which the phase-velocity becomes equal to region of the solar atmosphere should depict zero. Many persons have supposed that nat- and account for the internal motions. One ural oscillations of such a medium will lead to may question whether any static solution of the emission of radio waves on the plasma the fundamental equations exists. Small per- frequency. A detailed analysis indicates, how- turbations about such a hypothetical static ever, that the plasma oscillations, which re- solution, for example in terms of a linear semble sound waves in their longitudinal combination of normal modes, seems out of character, cannot be converted into transverse keeping with the actual observations. Instead, waves by means of boundary conditions alone. we suspect that the mean energy transport, In short, plasma oscillations cannot produce determined by the generation, transverse radio waves. When the medium can never be fulfilled in a purely radiative contains a magnetic field, however, the rigidity atmosphere. of the medium associated with magnetism The solution will depend on the magneto- resolves any longitudinal wave into vibrations hjdrodynamic in their having one transverse component. A dis- nonlinearized form. The Bhatnagar-Gross- turbance propagated from regions of high into Krook (1954) equation appears to be a useful regions of low density tends to have its vibra- substitute for the much more complex Boltz- tions amplified. In this way, as Krook and mann transport equation. The simplification Menzel have shown, a magnetohydrodynamic represents a new approach to problems of gas wave at low density will eventually escape from dynamics in which the distribution of velocities the sun's outer boundary as a true electro- 110 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS V«L. magnetic wave, with polarization determined versity and complexity of solar phenomena. A by the orientation of the internal magnetic second reason is that only in rare cases have the field. most modern spectroscopic techniques been ap- Low-frequency radiation tends to be trapped plied to the study of solar details. The changes or attenuated in an atmosphere containing on the solar surface are so gross and easy to ions and electrons. Thus, only higher fre- observe that solar observers have been tempted quencies are able to penetrate from the chromo- to make large numbers of observations with in- sphere through the overlying coronal regions. struments of relatively low resolving power, Similarly, a given disturbance rising through the either visually or on uncalibrated plates. solar corona permits lower and lower radio The rapid acceleration of observing programs frequencies to escape. Study of the changing during the International Geophysical Year gives spectrum of solar radio noise, especially in to those participating in the program the re- combination with optical data, should give sponsibility of maintaining high standards for valuable information about the solar atmos- their records. It is to be hoped that the need phere. J. P. Wild's (1953) dynamic spectrum for large masses of data will not distract the analyzer presents the observational facts in attention of solar physicists from the really basic graphic form. The Cornell measures of solar problems of solar activity, which pertain to the radio noise, as analyzed by H. W. Dodson origin and physical nature of solar phenomena. (1953), have already yielded significant cor- In the end, a complete understanding of the relations with flare and other solar activity. effects of solar radiation on the earth is more The corona itself poses many difficult prob- likely to come from the solution of these basic lems. There is little doubt that the streamers, problems than from correlations developed be- domes, and arches owe their basic structure to tween one or two observed solar quantities and general and local magnetic fields. Magnetic various geophysical data. focusing must be important in certain of these Knowledge of the physical state of the various phenomena. The forces associated with electric kinds of solar phenomena must precede correct currents undoubtedly play a significant role in theories of their origin. In the light of recent determining the motions of prominences and technical developments—which include the per- the general condensations that occur from the fection of gratings and the magnetograph by solar corona. The detailed theory of fluid gas the Babcocks, the construction of the vacuum dynamics necessary to account for these phe- spectograph by McMath, and the development nomena will undoubtedly be very complicated. of a large coronagraph by the Harvard-Air Force group on Sacramento Peak—we can look Solar activity forward to a fruitful period of quantitative The term "solar activity" includes all types of analysis of the spectra of solar details during the phenomena of a transient nature except the forthcoming sunspot maximum. small-scale fluctuations in velocity and temper- An understanding of the detailed structure of ature associated with granulation. Typical of sunspots, including the three-dimensional dis- these phenomena are sunspots, faculae, tribution of temperatures, pressures, gas ve- regions, flares, filaments, prominences, and locities, and magnetic fields, should be a pri- coronal disturbances. Solar activity occurs mary target during this period. There also throughout the solar atmosphere; e. g., the sun- seems to be no reason why high-dispersion spots and faculae in the photosphere, the flares studies of flare spectra should not lead to know1- in the chromosphere and corona, and the prom- edge of the vertical temperature and pressure inences largely in the corona. No attempt will distributions in the flare-emitting layers, which be made in this brief survey to summarize the can provide insight into the mechanisms of problems related to solar activity in any sys- origin both of the flares themselves and of the tematic way. The entire subject is still in a ultraviolet radiation responsible for ionospheric highly chaotic state, hardly less so at present disturbances. In this connection we consider than in the 1930's. One reason for this state that the physical investigation of faculae and of affairs is the aforementioned inherent di- the associated "plages faculaire" has been NEW HORIZONS IN ASTRONOMY 111 grossly neglected. It is well established that stimulate the birth of new theories as well as flares break out only in regions occupied by the applications of existing ones. plages. The latter, which are also regions of For example, the solution of many problems abnormal magnetic activity, are generally lo- of solar structure or solar activity waits upon cated above the photospheric faculae. There further developments in the field of generalized must exist, then, certain physical properties gas dynamics concerning the motions of high- associated with faculae and plages that favor temperature gases in the presence of various the outbreak of flares; yet the properties of types of force fields. Indeed, the United States these regions have hardly been studied. appears to be lagging in this important field, Our understanding of the physics of prom- which has applications to physics, geophysics, inences is equally weak. The development of and industry as well as to astrophysics. the motion-picture technique by McMath and Several symposia on this general subject have by Lyot in the early 1930's has been followed already been held and have been useful both in by large numbers of reasonably accurate meas- terms of summarizing what little knowledge urements of prominence motions, chiefly in Ha exists in this area and in stimulating new radiation. Some of the observed motions have thinking. If a really top-notch group of in- been explained plausibly, but nevertheless only vestigators could be assembled to work together qualitatively, on the basis of the combined for a relatively long period—for months, or even laws of hydrodynamics and electrodynamics by years, instead of a few days—to study the Alfv6n and by Menzel, but a detailed quantita- problem in its broadest aspects, we could tive description of prominence behavior is still expect significant new advances. For a prob- somewhere in the future. Here again, the lem of this magnitude, support would be development of a successful theory is retarded necessary for a number of years, during which by uncertain knowledge of the temperatures, headquarters of the project would be main- pressures, turbulent motions, and magnetic fields tained at a particular university with some within the prominences. The first three of rotation of personnel. A center of this sort these quantities are readily within the reach of devoted to theoretical problems of fluid gas modern instruments and theory. But one must dynamics could be as important to astronomy avoid the temptation to tackle these important and to other branches of science as a cooperative problems with equipment yielding marginal observing facility. data that can be interpreted only on the basis of doubtful assumptions or even wishful References thinking. ATHAY, R. G., BILLING*, D. E., EVANS, J. W., AND ROBERTS, W. O. Finally, although we have devoted major 1954. Astrophy.s. Journ., vol. 120, p. 94. attention to observational trends in solar ATHAT, R. G., EVANS, J. W., AND ROBERTS, W. O. physics, we have not wished to minimize the 1953. Observatory, vol. 73, p. 244. importance of support by the National Science ATHAT, R. G., AND MENZEL, D. H. Foundation and other Government agencies for 1956. Astrophys. Journ., vol. 123, p. 2S5. ATHAT, R. G., MENZEL, D. H., PECKER. J.-C, AND theoretical research of varying shades of purity. THOMAS, R. N. The so-called theoretical projects that are most 1955. Astrophys. Journ. Suppl., vol. 1, p. 505. likely to receive support today are those that ATHAT, R. G., AND ROBERTS, W. O. involve detailed computations— usually by high- 1955. Astrophys. Journ., vol. 121, p. 231. speed calculators—based on theories already in ATHAT, R. G., AND THOMAS, R. N. 1955. Astrophys. Journ. Suppl., vol. 1, p. 491. existence. In short, practical results are guar- 1956a. Astrophys. Journ., vol. 123, p. 299. anteed, but such projects are no more theoretical 1956b. Astrophys. Journ., vol. 123, p. 309. than those that involve the measurement of BHATNAQAB, P., GROSS, J., AND KROOK, M. stellar magnitudes or of intensities. 1954. Phys. Rev., vol. 94, p. 511. We do not question the value of accurate com- BILLINGS. D. E., COOPER, R. H., EVANS, J. W., AND putations to astronomy and solar physics. LEE, R. H. 1954. Electronics, vol. 27, p. 174. However, the National Science Foundation B6HM, K.-H. should also consider how its support might 1954. Zeitschr. Astrophys., vol. 35, p. 179. 112 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS

CHALONGE, D., AND KOURGANOFF, V. MENZEL, D. H. 1946. Ann. d'Astrophys., vol. 9, p. 69. 1931. Publ. Lick Obs., vol. 17. CHANDRASEKHAR, S., AND BREEN, F. MENZEL, D. H., AND PEKERIS, C. L. 1946. Astrophys. Journ., vol. 104, p. 430. 1935. Monthly Notices Roy. Astron. Soc, vol. DODSON, H. W. 96, p. 77. 1953. In Kuiper, ed., The sun, p. 692. MOHLER, O., AND GOLDBERG, L. ELSTE, G. 1956. Astrophys. Journ., vol. 124, p. 13. 1955. Zeitschr. Astrophys., vol. 37, p. 184. NEWKIRK, G. GIOVANELLI, R. G. 1953. Thesis, Univ. Michigan. 1949. Monthly Notices Roy. Astron. Soc., vol. 109, p. 293. PECKER, J.-C, AND ATHAT, R. G. HAGEN, J. P. 1955. Astrophys. Journ., vol. 121, p. 391. 1954. Jouru. Geophys. Res., vol. 59, p. 158. PIERCE, A. K. MATSU8HIMA, S. . In press. 1955. Astrophys. Journ. Suppl., vol. 1, p. 479. RICHARDSON, R. S., AND SCHWARZSCHILD, M. MCMATH, R. R. 1950. Astrophys. Journ., vol. Ill, p. 351. 1956. Astrophys. Journ., vol. 123, p. 1. RUSSELL, H. N. MCMATH, R. R., MOHLEB, O., AND PIERCE, A. K. 1929. Astrophys. Journ., vol. 70, p. 11. 1955. Astrophys. Journ., vol. 122, p. 565. WILD, J. P. MCMATH, R. R., MOHLER, O., PIERCE, A. K., AND 1953. In Kuiper, ed., The sun, p. 676. GOLDBERG, L. WILDT, R. 1956. Astrophys. Journ., vol. 124, p. 1. 1939. Astrophys. Journ., vol. 89, p. 295. Stars and the Galaxy

Spectral Classification By W. W. Morgan *

In a classical review titled "Some Problems As early as 1897, Miss had of Sidereal Astronomy" prepared by H. N. reached the conclusion that— Russell in 1917, the main object of astronomy ... a single series was inadequate to represent the is stated as being peculiarities which presented themselves in certain . . . the development, on the basis of collected facts, cases, and that it would be more satisfactory to assume of satisfactory theories regarding the nature, mutual the existence of collateral series. relations, and probable history and evolution of the She devised a system of 22 "groups" which objects of study. formed a sequence; most of the "groups" were Russell suggests that before the stage of then divided into "divisions," which, in turn, formation of a general theory is reached, two introduced a second dimension. It was from courses may be followed: this pioneering work of Miss Maury that the (1) to collect masses of information, as accurate field of spectroscopic parallaxes derives, through and extensive as possible, by well tested routine the later work of Hertzsprung, Russell, Adams, methods, and leave it to the insight of some fortunate and Kohlschutter. and future investigator to derive from the accumulated The situation in 1920 was, therefore, approx- facts the information which they contain regarding the general problems of the science; (2) to keep these imately as follows: spectral classification was greater problems continually in mind, and to plan the considered to be primarily a one-dimensional program of observations in such a way as to secure as problem; in addition, differences in luminosity soon as practicable data which bear directly upon were observable by means of certain sensitive definite phases of these problems. line-ratios. The early work of Miss Maury The present survey gives an approximate had not been adopted as standard, nor was it picture of the research situation in spectral continued on any appreciable scale. classification in 1956, a comparison with If a single development could be considered Russell's survey, and some suggestions of the of primary importance in spectral classification probable directions for future work. since 1920, it would be the realization of the The , which was fundamental two-dimensional nature of the published between 1918-1924, remains the most problem. The critical work here was that of important fundamental body of data in Rus- Bertil Lindblad, who developed criteria for sell's first category. Russell's conclusions are, the separation of stars of differing luminosity to a large extent, based on the Harvard work. by use of plates of low dispersion; this last property permitted the extension of the spectro- The interval 1917-1955 scopic parallax method to great numbers of Two fundamental conclusions were drawn by stars too faint to be observed with higher Russell from the study of the lines of stellar dispersion. spectra: (a) that almost all of the observed Another fundamental development was the lines are identifiable with known elements; formalizing of the earlier work (in particular, and (b) that the vast majority of stellar spectra that of Lindblad) in the two-dimensional fall into a single, continuous linear sequence. Yerkes system of Morgan and Eeenan. As far Later work has confirmed fully Russell's first as the actual criteria of classification are con- conclusion; the second requires some qualifica- cerned, there was little new in the latters' tion. system; the principal development lay in the

1 University of Chicago, Chicago, III. recognition that spectral classification per se 115 116 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS is a two-dimensional process and that the made on this most promising field for future designation of the "luminosity class" should galactic studies. be by means of an empirical notation similar A most important development has been in in nature to the ordinary "temperature type." the field of classification of the faint inhabitants The process of determining spectroscopic of dust clouds—the stars—by Joy absolute magnitudes was thus divided into two (Mount Wilson), Herbig (Lick), and Haro stages: (1) the determination of a spectral (Tonantzintla). The work on these strange type and luminosity class; and (2) the calibra- systems of stars is also in its infancy. tion of the empirical luminosity classes in terms of absolute magnitude. The first of these Present and future problems operations is purely a description of the spectra. This topic will be considered under three This description is a fundamental property of general heads: (1) astrophysical interpretation; the latter, and there are no compelling reasons (2) spectral classification as applied to problems why it would have to be revised with the passing of galactic structure and stellar evolution; and of time. In this respect the second operation (3) instrumental needs. can be improved with the passing of time, due (1) Astrophysical interpretation.—Very prom- to more and more accurate calibration data ising beginnings have been made in the past few becoming available. years in devising semiquantitative methods of Another development in the intervening classification for stars in certain restricted areas years has been the devising of methods for of the HR diagram. Some of these have made segregating certain kinds of stars by means of use of measures of line depths and equivalent spectra of exceedingly low dispersion. The widths to furnish more objective data for classi- most important applications of these methods fication. Of especial importance is the work have been to the use of objective-prism of Greaves, Baker, and Wilson at Edinburgh, of spectrograms. Chalonge and associates in Paris, and of Hos- In 1950 attention was called to the possibility sack at Toronto. Accidental accuracy of a high of classifying members of certain "natural order has been achieved. The principal difficul- groups" on objective prism spectra. For ty here is probably in devising a frame of refer- example, a single criterion is sufficient for the ence sufficiently precise to make full use of the segregation of a well-defined group of early A extremely accurate individual results obtained main-sequence stars; this is carried out by by these investigators. means of the series of hydrogen. If a It is from investigations such as these that we group of stars is selected having the character- will be able to decide how uniquely the stars can istic of maximum intensity for the Balmer be classified in a two-dimensional array, and lines, it is found that the stars in this group what the relationship is between the "normal" possess the characteristics of closely similar and the "peculiar" stars. The practical limit absolute magnitude and spectral type. Among attainable in the accuracy of spectroscopic par- the most interesting of the groups identifiable allaxes will also be determined from high-dis- on objective-prism plates is that of the blue persion investigations. giants. Nassau and collaborators at Cleveland (2) Spectral classification as applied to prob- have prepared such a catalog of northern blue lems of galactic structure and stellar evolution.— giants, and this work has been extended to The application of the southern and fainter northern method to large numbers of fainter stars is now stars by Haro and associates at the National quite feasible. The onty thing needed for a Astrophysical Observatory in Mexico. great increase in our knowledge of the arrange- Of comparable importance has been the ment of the stars in space in the neighborhood development of a method of classification of of the sun is more investigators; the method and late M giants in the infrared spectral region by equipment are now quite adequate for the work. Nassau and collaborators. A catalog of the New programs will probably be of two kinds: brightest of these objects has been published (1) work on the O-associations, galactic clusters, recently, but a beginning has scarcely been and other objects associated with spiral struc- NEW HORIZONS ENT ASTRONOMY 117 ture; and (2) investigations of the amorphous The requirements for such "maximum effi- disk and halo population. ciency" spectrographs have been outlined An interesting group for further investigation clearly by Bo wen (1947, 1952), who has also in the second category is that of the late M designed and supervised the construction of giants, which are most important for the de- instruments of this type which are now in use termination of the kinematical properties of the with the Mount Wilson-Palomar reflectors. disk population. However, for their observa- Special attention should be called to the new tion at great distances it will be necessary to 60-inch spectrograph designed by devise infrared techniques for measurement of which has recently been put into opera- radial velocity and absolute magnitude. tion. The faint limiting magnitude obtainable Some of the most dramatic developments of with this instrument illustrates how great a gain the 30-odd years since Russell's survey have in efficiency is possible with reflectors in the been in the field of stellar evolution. The second category of size. Special mention should theoretical work in this field, up to 1951, has also be made of the exceedingly efficient nebular been well summarized by Stromgren (1952). spectrograph designed and constructed under This work, together with later developments, the direction of Mayall for the 36-inch Crossley now makes possible the application of Russell's reflector of the Lick Observatory. second criterion quoted at the beginning of this There are now several large and most useful summary. It is now possible to formulate objective prism Schmidt cameras in operation. precise observational experiments concerned Especially important new instruments of this with various aspects of the great general type have been put into operation recently at problem of stellar evolution. Hamburg-Bergedorf, the University of Mich- One of the most inviting practical problems igan, and Vanderbilt University. Exceedingly of the future is the invention of a system of efficient slitless spectrographs have been de- spectral classification according to stellar age, signed and constructed recently by Herbig or relative evolutionary state. (Lick) and Meinel (University of Chicago). (3) Instrumental needs.—The principal in- References strumental need for present and projected observational problems in this field is probably BOWEN, I. S. 1947. Publ. Astron. Soc. Pacific, vol. 59, p. 253. not for new or larger telescopes but for more 1952. Astrophya. Journ., vol. 116, p. 1. efficient spectrographic equipment for use with STROMGREN, B. existing instruments. 1952. Astron. Journ., vol. 57, p. 65.

381014—56

Stellar Atmospheres By Marshal H. Wrubel'

The outer region of a star in which the optical phere follows from the temperature distribution spectrum is formed is referred to as the atmos- by using the hydrostatic equation, the absorp- phere. This definition delimits a shell lying tion coefficient, the ionization equation, etc. between the deeper regions, which cannot be The resulting "model atmosphere" combined directly observed, and the higher regions, which with the theory of the formation of absorption contribute very little energy to the optical lines is used to predict the observed spectrum. spectrum (but which may be all-important in The only adjustable parameters of such a solar the radio spectrum). Although it is convenient model are the abundances of the elements which to treat this shell separately, it is certainly not are determined by a comparison with observa- independent of the layers lying below and tions. above it. Although the reasoning is straightforward, Any discussion of stellar atmospheres logically many problems arise even in the static model. begins with the sun. Here is a star on our To list only a few: doorstep; and the variety and accuracy of the (1) The temperature distribution of the solar data we can accumulate far surpasses that upper layers is uncertain because of the for any other star. We may examine the difficulty of observation close to the limb of the radiation received from every point on the sun. sun's disk and for every wavelength that (2) The abundance of the second most penetrates the earth's atmosphere. abundant element, helium, cannot be deter- In spite of centuries of solar observation, mined directly since it is effectively "dead new techniques continue to introduce new weight" in a low-temperature atmosphere. phenomena. To mention only two of the most (3) The fundamental data of transition interesting: the Babcocks' magnetograph, probabilities for many lines are still missing. which gives a tracing of the atmosphere's The great emphasis on nuclear and molecular magnetic field point by point; and the McMath- physics in recent years has attracted physicists Hulbert vacuum spectrograph, which has re- away from research in atomic spectra. A vealed a remarkable complexity of line struc- possible solution is for some young astronomers ture. Both of these innovations are triumphs to learn the laboratory techniques and to do the of observational technique and will create work themselves. challenges to the theoretician. (4) The theories of the shape of the line- The theoretical models that have been ex- absorption coefficient, i. e., broadening, absorp- plored until recently may be characterized as tion versus scattering and noncoherent proc- being in "hydrostatic and radiative equilib- esses, are not in satisfactory form. Of partic- rium"; that is, the atmosphere is supported by ular interest in this connection is the recent use hydrostatic pressure and radiation is the sole of a shock tube at the University of Michigan means of energy transport. The logical chain to obtain laboratory data on broadening. is this: the temperature distribution is found (5) The problem of predicting the profile either theoretically or by limb darkening of an absorption line formed when radiation measurements, and it depends on the broad passes through many layers of different physical continuous absorption characteristics of the gas. characteristics has been solved only for very A complete physical description of the atmos- special cases. 1 Indiana University, Bloomlngton, Ind. (6) The cumulative effect of many absorp- 119 120 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS tion lines (blanketing effect) is difficult to direction is being planned for the forthcoming evaluate. International Geophysical Year. Many prob- However, overshadowing all these difficulties lems in this particular field are treated in- today is the problem of describing a dynamic dividually in other articles of this volume. atmosphere. The theoretician no longer can The techniques developed to describe the avoid including large-scale motions in his solar solar atmosphere have been applied with models but the kinematical description is far some success to "normal" atmospheres of other from clear. One important source of informa- spectral types. In some cases it is possible tion is the "granulation"; but the details are to obtain information that the sun will not obscured by "seeing" in the earth's atmos- provide directly: for example, the abundance phere. It is to be hoped that Schwarzschild's of hydrogen relative to helium can be found unorthodox plan to photograph the sun from in principle from the spectra of hot stars. above the "seeing" will be successfully carried Unfortunately, no conclusive value can as yet out in the near future. be given. Although simple models (such as those used Problems immediately arise because of ob- by de Jager) involving temperature inhomo- servational limitations. There is just not geneities in adjacent columns of the atmosphere enough light to work with. Photoelectric may prove serviceable for a while, eventually techniques for the absolute spectrophotometry we must have a model obeying the hydro- of stars are currently being developed by several dynamical laws (better still the hydromagnetic observers and should yield important astro- laws) and including the transport of energy by physical data. Nevertheless we have only a combination of convection and radiation. indirect measurements of limb darkening, and Increasing use will undoubtedly be made of these for only a few eclipsing binaries. high-speed electronic computers to obtain As the available light goes down, so does the numerical solutions to these difficult problems. possible dispersion, and observations become Computers can certainly not create ideas, but less reliable. Some objects are out of reach of they can explore the consequences of various all except the largest reflectors. assumptions to an extent previously impossible. Surveys can be carried through by using It is a challenge to find efficient machine pro- rough curve-of-growth methods based on ideal- cedures for calculating model atmospheres and ized models, but eventually we need detailed line profiles under a variety of physical con- model atmospheres for a wide variety of spectral ditions. Computer techniques will probably types. These are constructed by using theoret- be second nature to astrophysicists a few ical temperature distributions and are tested academic generations hence; at the moment, by requiring the flux to be constant at every however, theoreticians have been very con- depth. However, it is difficult to apply this servative in adopting these new methods. criterion in early-type stars where the flux As if the existing observations do not provide is jagged due to the absorption edges of hydro- sufficient problems, astronomers have watched gen and helium. with great interest as rocket spectrographs and As we go toward the cool stars, the im- radio telescopes have broadened the limits portance of the blanketing effect grows and a of the observable spectrum. Lyman a has at premise upon which other models are founded last been photographed and even soft X-rays breaks down—it is not possible to separate have been detected. Observations such as the lines from the continuum. Entirely new these will have a profound effect on future theoretical techniques will have to be combined models of the sun. with laboratory data on the absorption of mole- There is very little generally accepted theory cules before we will have as detailed an inter- concerning the spectacular transient solar pretation of the atmospheres of stars of late phenomena. Many solar observatories are in- spectral type as we have for the sun. volved in the accumulation and classification Nature has provided a unique tool for study- of data from which theories will eventually ing the outer regions of late-type giants. arise, and a highly coordinated effort in this Binary systems like 31 Cygni, in which a small NEW HORIZONS IN ASTRONOMY 121 hot component is gradually eclipsed by the References extended atmosphere of its giant companion, AIXER, L. H. provide a wealth of detailed information. The 1953. Astrophysics. Ronald Press, New York. observations indicate a "network of prom- , A., BD. inences" far more extensive than that of the 1955. Vistas in astronomy. Pergamon Press, sun; and if an understanding of a dynamic London and New York. ClIANDRASEKHAR, S. atmosphere is important in the sun, it is vital 1950. Radiative transfer. Oxford University here. Press, London. As the spectroscopic peculiarities become HYNEK, J. A., ED. more unusual, the extrapolation of normal tech- 1951. Astrophysics. McGraw-Hill, New York. niques becomes less successful. Thus, our KoURGANOFF, V., AND BUSBRIDGE, I. W. present understanding of the atmospheres of 1952. Basic methods in transfer problems. Ox- pulsating variables, magnetic stars, Wolf- ford University Press, London. Rayet stars, etc., is rudimentary at best. KUIPER, G. P., ED. Nevertheless, the tendency is to make quantita- 1953. The sun. University of Chicago Press. tive measurements whenever possible and not UNSOLD, A. 1955. Physik der Sternatmospharen. Revised to be satisfied with visual estimates. ed. Springer, Berlin. It is obviously impossible to cover all aspects WOOLLEY, R. V. D. R., AND STIBBS, D. W. N. of such a varied subject in a short survey. 1953. The outer layers of a star. Clarendon Fortunately several detailed discussions of solar Press, Oxford. and stellar atmospheres have recently appeared, WRUBEL, M. H., ED. and references to individual papers will be 1954. Proceedings of the National Science Foun- found there. dation conference on stellar atmospheres. Indiana University. Bloomington, Ind.

Aerodynamic Problems in Stellar Atmospheres By Richard N. Thomas*

The term "cross-field research" occurs with methods of description of a gas under conditions increasing frequency in astronomical studies. departing drastically from local thermodynamic This new emphasis does not imply that as- equilibrium. tronomy, in the past, failed to utilize advances These problems are all of vital interest to in other sciences; rather, it indicates our grow- astrophysics because our observations show ing awareness that the physical problems of that such configurations occur in astronomical matter studied by the observational astron- objects. Several of these problems, those con- omer are not far removed from those studied by nected with the ablation from a rapidly moving the laboratory scientist. Classically, astro- solid object, form the fundamental basis of physics has dealt with matter under condi- meteor astronomy and have led to the creation tions that were strange and abnormal, com- of astroballistics, a new field comprising both pared to those in the terrestrial laboratory. astronomy and aerodynamics, which is de- Astronomy has been an observational science. scribed elsewhere in this series of articles. The objects of its inquiry have been obscured Others of these problems seem to occur with by the earth's atmosphere and situated at great increasing frequency as we acquire more and distances; and the phenomena studied were better data on stellar atmospheres and are incapable of reproduction in the laboratory. thus faced with the need to construct a theory Of all the "cross-field" liaisons that exemplify sufficiently precise to explain the data. In our changing views, none is more interesting this review, we want to elaborate on certain than that between modern aerodynamics and aspects of stellar atmospheres and to emphasize astrophysics. the continued interplay of astrophysical theory Modern astrophysics not only concerns itself and observation with aerodynamic theory and with the common problems in the two sciences experimentation, in a kind of checkerboard but also recognizes that the data required to pattern. The exciting aspect concerns the support a theory or an anticipated future common physical conditions occurring in the development in one field may already exist in and the aerodynamic experi- the other. Also, the analytic methods that ment. The frustrating aspect is the lack of are gradually being developed for a specific facilities for experimental astrophysics, as it problem in one field may already be common becomes more and more evident that we can knowledge in the other. For example, the actually duplicate in the laboratory physical advent of high-altitude rockets, satellites, and situations of direct astronomical interest. the possibility of extraterrestrial manned ob- Ten years ago, theorists concerned with jects will break the barriers imposed by the stellar atmospheres were completely preoc- terrestrial atmosphere and the astronomical cupiod with the notion of a gaseous atmosphere distance scale. Yet a satisfactory development whose properties were fixed by the three of ultraspeed aerodynamics requires careful assumptions: radiative transfer of energy, local attention to the problems of ablation from a thermodynamic equilibrium, and the absence solid moving through a gas, of radiative energy of any velocity fields large enough to affect loss from a gas compressed rapidly to a high the structure of the atmosphere. The discovery degree, of the attendant ionization and chemical by Schwarzschild, 50 years ago, of the over- reactions in the gaseous atmosphere, and of the whelming efficiency of radiative energy trans- 1 Harvard College Observatory, Cambridge, Mass. port, relative to convective, completely dom- 123 124 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 inated all thinking in subsequent years. The thermal energy—emission lines, excesses of later recognition of atmospheric convective ultraviolet radiation, etc. The situation, there- zones in stars where hydrogen was not mainly fore, was far from clear cut. ionized was of minor consequence; even in (If the reader expects the above history to be these zones radiation dominates the energy climaxed by a happy solution in this review, he transfer. There were, however, three "types" will be disappointed. If these problems had of atmosphere where "contrary" evidence was been solved there would be little point in record- becoming difficult to rationalize away: the ing their apparently illogical history. We dwell atmospheres of the variable stars, where the on the above pattern of thinking because, un- notion of a velocity field had long existed but happily, much of it is today's pattern of think- where the exploration of the physical con- ing also. At present we can report only on sequence of its existence had been rudimentary; efforts to resolve these difficulties, not on their the atmospheres of certain supergiants, where success.) the evidence for large velocity fields had existed On the other hand, when the thermal dissi- for 10 years but nothing had been done in the pation of the mechanical energy of the velocity way of incorporating them into an atmospheric field was included, the effect on the atmospheric model; and the solar chromosphere, where model was, and is, likewise far from clear. This considerable evidence of velocity fields and of energy transfer represents the most serious and nonequilibrium phenomena caused the break- the most fascinating problem of all. First, any down of virtually all our classical notions of a such energy transfer establishes a cyclic process, stellar atmosphere. a mechanical energy source and a radiative sink. There are two ways in which a velocity field Such a situation is ideal for setting up a de- can perturb an atmosphere: through a transfer parture from the conditions of local thennody- of momentum or a transfer of energy. What namic equilibrium. Second, such energy transfer little thinking on the influence of velocity fields introduces the possibility of a perturbation on that did take place in these earlier days was the radiation field produced by the star. Con- concerned largely with the momentum effect. sequently, the obvious subject for investigation Velocity fields were simply imposed as an added is the problem of the energy transfer associated pressure term, and the thermal and radiative with velocity fields in stellar atmospheres. properties of the atmosphere were supposed to The solar chromosphere offers the most nat- remain invariant. Unfortunately, the velocity ural starting point for an investigation of this fields discussed are often of such size that, if the sort. First, from a general point of view, there temperature of those atmospheric regions where is one strong boundary condition on this investi- the velocity fields exist are not higher than we gation of aerodynamic effects in stellar atmos- would infer from our customary theories, the pheres: the "classical" stellar atmosphere mod- motions must be supersonic. Such a situation els have been remarkably successful in repro- made it difficult to introduce the experience of ducing the features of the continuous spectrum the aerodynamicist in describing velocity fields in the visual region. How can we introduce in a compressible gas because such supersonic such things as Shockwaves without affecting the motion, particularly of the quasirandom sort re- continuous spectrum? In the sun, the large quired by the astronomical data, is in aerody- velocity fields seem confined to the outer at- namics always accompanied by Shockwaves and mosphere, where the total mass is too small to by the rapid conversion of mass motion into make appreciable contribution to the continu- thermal. Thus we had the unhappy situation of um. We may, however, determine how far an empirical fact (large velocities) forced into an down toward the photosphere the effect of the artificial framework (no dissipation of me- chromospheric velocity fields extends by look- chanical energy) because astronomers were re- ing at weaker and weaker lines and the wings of luctant to abandon the atmospheric models the strong lines. With this information, we may which had hitherto worked so well. Moreover, be able to answer the question of what observ- there were absent many of the features one able consequence would result from an increase would expect to occur in a rapid production of by some factor in the energy of the velocity field. NEW HORIZONS IN ASTRONOMY 125 Or, alternatively, what would we have to do to spicules arise from granulation. In this case, turn the solar atmosphere into that of a Wolf- all stars with atmospheric convection zones Rayet star, where the whole atmosphere seems should have , that is, all main- a composite of aerodynamic motions? sequence stars of spectral class later than~^40. Second, with a view to finding the origin of But what of the earlier spectral classes? these velocity fields, the closeness of the sun Should we expect no extended atmospheres and our ability to resolve its disk are of enor- here, no velocity fields? Yet the existence of mous value. Many astronomers believe that differential velocities between emission and the chromospheric velocity fields simply reflect absorption lines in stars of classes O and B the motion of the solar spicules, which are is well known. In a more extreme case, we supposed to arise from the granulation, whose refer again to the Wolf-Rayet stars, where the origin lies in the hydrogen convection zone. whole atmosphere seems composed of a system Some of these astronomers regard magneto- of jet-ejections, or streams of outgoing and in- hydrodynamic effects as essential in the produc- going material. Whence arises this motion? tion of spicules, from granules; others believe Aside from the possibilities of the hydrogen the transition can be effected from the aero- convection zone (and possibly a helium con- dynamic behavior of a compressible gas in a vection zone in the hotter stars ?) and induction gravitational field alone; and still others believe effects in time-dependent magnetic fields, we that the connection between spicules and have two other sources of mechanical energy: granules has yet to be established. Yet in all stellar rotation and stellar pulsation. these points of view several experimental and It is interesting to note that the first sugges- theoretical aerodynamic problems stand out— tion that aerodynamic turbulence might arise problems which we can investigate jointly in in stellar atmospheres arose from the observa- astronomy and in aerodynamics: (1) What is tion of the high Reynolds numbers associated the behavior of a supersonic jet when radiation with stellar rotations. Furthermore, we note processes form a major dissipative mechanism ? that "shell" stars are most often those having a Recent solar work on the stability of an rapidly rotating nucleus. What is the aero- ionized gas against radiative cooling suggests dynamic state of such a rapidly rotating gaseous that these stability considerations may fix, to atmosphere ? Does the star set up a system of a major degree, the permitted kinetic tempera- aerodynamic turbulence, anisotropic for the ture of the gas in the jet. (2) How does a largest scale motion, with the greatest velocities supersonic jet composed of an ionized gas be- along the direction of rotation? This would have in a magnetic field, with the field (a) conform to laboratory experience. Or, is the parallel to the jet axis, and (b) perpendicular? rotation velocity so high (greater than our (3) Suppose we have a velocity field, with time- laboratory experience) that rings of matter are dependent terms, in a gaseous atmosphere in a ejected? This problem has hardly been gravitational field. How rapidly do the acous- touched, astronomically or otherwise. tic waves, produced by this field and propa- The concept of stellar pulsation is relatively gated upward, die out? What is the inter- old; the implications of stellar pulsation, ference pattern ? (4) If we superpose a fluctu- relatively new. Given a gaseous atmosphere ating magnetic field on this velocity field with of arbitrary temperature and density distribu- time-dependent terms in an ionized medium, tion, and an impressed oscillation of some ther- can induction effects accelerate a nontrivial modynamic variable at some level in the amount of gas? (5) Last but hardly least in atmosphere, what happens? Remember, we our incomplete list of problems, How do we are not dealing with the terrestrial atmosphere, specify the spectroscopic state of the gas under with temperature ~ zero so far as activation the environment of these four problems, and of internal degrees of freedom are con- others in the same physical sequence ? cerned. We need a terrestrial disturbance Let us return for a moment to the suggestion of ~Mach 10 to cause any but the most trivial that spicules are the source of the mechanical luminosity. In the star, the ambient temper- energy which causes the chromosphere and that ature ranges upward from~4,000°, and even 381014—66 10 126 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS for Mach 2 one expects to find an observable of the h3Tdrogen convection zone and the extent change in ionization and excitation. A literal of the atmosphere. The convection zone interpretation of velocity-curves would suggest appears to determine the phase relation between Af>2 in a wide class of variable stars. In- the luminous flux and pressure at the "origin" deed, in some stars emission lines appear over of the disturbance; the extent of the atmosphere some portions of the cycle, and we might determines the character of the disturbance, interpret their occurrence as implying the and, in a sense, has a self-amplifying aspect. existence of localized regions of high kinetic The more the dissipation of energy in the at- temperature. On the other hand, some stars, mosphere, the more extended the atmosphere, whose velocity variations imply compression hence the greater the chance for a shock to waves of sufficient amplitude to be shock- develop and dissipate energy. waves, do not show these emission lines. Should we interpret these differing results as All these problems are tantalizing. Few suggesting that the compression wave just have been investigated, none solved. Our does not have a sufficiently long path to develop greatest progress has been in our finally recog- into a shock before it is out of the visible nizing what the problems actually are. Our atmosphere? greatest challenge is the need to develop some Actually, two physical parameters seem to physical insight to the solution of these prob- play a major role in aerodynamic phenomena lems which lie on the most advanced frontiers in the atmospheres of variable stars: the extent of both astrophysics and aerodynamics. Emission Nebulae By Lawrence H. Aller 1

Galactic nebulae that show an emission line (b) Photographs with the aid of Fabry- spectrum are of two types. Those associated Perot etalons and suitable plate-filter com- with the type I population are the so-called binations are extremely valuable for studying diffuse nebulae; those associated with the the internal motions in extended diffuse neb- type II population are the planetaries. The ulae. This technique, introduced by Fabry diffuse nebulae are normally irregular in form, and more than 40 years ago and often of low density and surface brightness, further developed by Baade, Goos, Koch, and and sometimes of considerable size (e. g., Minkowski, has been applied with great success , 30 Doradus in the Large Magellanic by Georges Courtes at Haute-Provence. Stud- Cloud, and NGC 604 in the ies may be made with relatively small wide- Spiral). The planetaries are often symmetrical angle telescopes that will yield considerable in form and smaller than the diffuse nebulae. information on the mass motions and internal Many have a higher surface brightness and turbulence. It is important that such in- higher density than the better known diffuse vestigations be carried out not only in Ha but nebulae. also in the forbidden radiations of [O II] and [O III] whenever possible. Types of observations (c) Spectrograpbic studies fall essentially Because of their different characteristics, the into three categories. First are slit spectro- kind of equipment and technique best suited graphic studies of the line and continuous for the one type of emission nebulae is not spectra to obtain line identifications and in- necessarily best for the others. In studies of tensities from which the physical state in the diffuse nebulae, wide-angle cameras of high nebula or nebular filament may be inferred. speed and detectors sensitive to low surface For this type of problem a so-called nebular are desirable. For studies of the spectrograph with high speed and low dis- planetaries a large scale (and therefore a large persion is necessary for all but the brightest telescope) is essential, since the angular di- objects. Planetary nebulae have been studied ameters of all but the largest of these objects rather intensively, particularly by Bowen and amount to only a few seconds of arc. The his coworkers, but the spectra of diffuse nebulae principal types of observation that can be (except Orion and a few special objects) have made are the following: not been adequately investigated. Spectro- (a) Direct photographs with the aid of nar- graphs of high dispersion may be used for row band-pass filters to isolate monochromatic studies of the internal motions in small, bright nebular radiations. In particular, interference planetaries. Olin Wilson's multislit has proved filters have proved very useful in picking up particularly valuable in such studies. Finally, faint Ha emission regions throughout the slitless spectrograms are essential for the study . When most of the nebular radi- of monochromatic images of small nebulae ation is concentrated in a few strong, well- when radiations that fall close to one another separated lines, the filter technique is extremely are to be separated; e. g., the [N II] X6548, valuable for studying problems of stratification 6584 and Ha 6563 Lines. and structure. (d) Polarization measurements are important 1 University of Michigan Observatory, Ann Arbor, Micb. in these objects where scattering by small solid 127 128 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL. 1 particles occurs and in radio sources. Electron Interpretation of the observational data in scattering is rarely a factor in the continuous terms of physical parameters spectra of diffuse nebulae. The densities and temperatures of the radiating (e) Radio-frequency observations of the free- gases may be established from the surface free emission from diffuse nebulae have been brightness and from spectroscopic data. If the made by Haddock and his coworkers. Even- nebula is smooth and unobscured and an tually, with the development of larger antennae, estimate of the electron temperature is avail- it should be possible to measure the emission able, a measurement of the surface brightness from the brighter planetaries as well. These in Ha or preferably in the Balmer continuum objects have high surface brightnesses but small will provide an estimate of the electron density. areas so that their total flux is small. The Both diffuse nebulae and planetaries are fre- 21-cm data supply information on those regions quently highly obscured. If a measurement of of space where hydrogen is neutral. the Balmer decrement is available, an estimate of the space absorption can be found readily. Observable parameters Boggess III has shown how one may The following data can be obtained for typical combine optical measurements of surface bright- emission nebulae: ness in Ha with radio-frequency measurements (a) The apparent size, form, and structure to get the space absorption and densities for Can be measured in each of the stronger mono- certain diffuse nebulae for which Balmer decre- chromatic radiations. ments have not been measured. (b) The surface brightness and isophotic con- The forbidden lines provide estimates of the tours can be measured for each of the stronger electron density N« and give us our best guesses monochromatic emissions. concerning the electron temperatures T(. If all (c) The wavelengths, identifications, and in- the radiations originate in the same strata one tensities of the lines and the energy distribution may obtain N( and T« from a comparison of the in the continuous spectra can be measured for forbidden line intensities of [O III] and [N II], typical nebulae. for example. Often, however, the [N II] and (d) The distances of diffuse nebulae may be [O III] lines are predominantly produced in found from the apparent magnitudes, color different regions of the nebula. Hence, some excesses, and luminosity classes of their illumi- other method is needed. A number of years nating stars. The distances of the planetaries ago, Ufford, Van Vleck, and I called attention eannot be established so easily, and for the to the fact that the intensity ratio of the 3726 moment we shall have to be content with and 3729 [O II] lines varied from one nebula to statistical distances based on the association of another, apparently depending on the density. these objects with the central bulge of our Seaton calculated improved target areas for the galaxy and external galaxies. collisional excitation of the [O II] lines so that (e) The relationship between the structure quantitative estimates of the density could be and internal motions can be established from a made from the observed line intensities. Oster- comparison of isophotes of slitless spectrograms brock has recently derived densities in the fila- or monochromatic photographs and radial veloc- ments of the Network Nebula, in the Orion ity measurements. It is possible to establish Nebula, and in other objects from the X3726/ whether the motions are orderly expansions or X3729 data. turbulent streaming, and to correlate the motion with the appearance of the object. O. C. Some outstanding problems Wilson has given a rather complete picture for At the moment, the physical nature of the the planetaries. Similar studies for the diffuse excitation of the bright line spectra of the gase- nebulae should prove worthwhile. ous nebulae appears to be well understood. The NEW HORIZONS IN ASTRONOMY 129 continuous spectra of normal, purely gaseous Perhaps the most exciting problems are those nebulae are less well understood, but seem to be posed by the gaseous nebulae that are strong interpretable in terms of recombination of hy- radio sources. Oort and Walraven recently in- drogen and helium ions with electrons and the terpreted the continuous spectrum of the Crab double-photon emission. Further quantitative Nebula as "synchrotron" radiation from fast measurements of the nebular continua are electrons moving in a magnetic field. Possibly needed to test various theoretical suggestions. such radiation is always or nearly always associ- Further calculations of transition probabilities ated with strong radio-frequency emission, for forbidden lines, especially those of iron in although powerful radio sources do not neces- various stages of ionization, are needed. Con- sarily emit a strong optical continuum. The tinuous absorption coefficients should be calcu- origin of the magnetic fields and the source of lated for abundant ions, and further target the energy of the particles is not explained. It areas for the collisional excitation of metastable is not sufficient to blame it on the supernova levels should be computed. The accuracy of process; the detailed mechanism must be worked the available theoretical target areas ought to out. be improved. New equipment and observational techniques Numerous theoretical problems connected have expanded and are continuing to expand with the structure and especially the kinematics our empirical knowledge of emission nebulae. of gaseous nebulae remain to be formulated and Theoretical interpretations of the forms and solved. For example, what holds together the kinematics are lagging behind. The problems thin filaments in such an object as the Network posed are among the most difficult in theoretical Nebula? It is generally assumed that magnetic astrophysics, a fact which makes the subject of fields are responsible, but the character and par- bright-line nebulae all the more challenging. ticularly the origins of these fields remain ob- Some suggested references scure. Any real understanding of the forms of diffuse nebulae must await some fundamental The following books and review articles con- progress in the aerodynamics of compressible tain further references to the literature: ionized gases moving at supersonic velocities in ALLER, L. H. the presence of magnetic fields. Classical hydro- 1956. Gaseous nebulae. Chapman and Hall, dynamics and turbulence theories will require London. VAN DB HULST, H. C, BD. much amplification and extension before they 1955. Second symposium on cosmical aerodynam- can be applied to the diffuse nebulae and ics held at Cambridge, 1953. Published planetaries. for Int. Astron. Union by North Holland The relation between the diffuse gaseous Publishing Co. and by Interscience Press, New York. nebulae and the solid particles of the inter- WURM, KARL stellar medium requires further theoretical and 1951. Die planetarische Nebelu. Akademische observational study. Verlag, Berlin.

Double and Multiple Stars By Otto Struvex

The emphasis at the present time is upon bottleneck near the inner Lagrangian point, Llt problems connected with the origin and evolu- of the critical zero-velocity surface (instability tion of double stars. G. P. Kuiper's discussion of type A), or through an opening at the first in three articles in the Publications of the Astro- outer Lagrangian point, Zj (instability of type nomical Society of the Pacific2 in 1935 and 1955 B). Kuiper thought that the instability of fi forms a sound basis for future work. He attrib- Lyrae is a combination of types A and B, but utes the origin of double stars to the formation more recent work seems to indicate that both of separate condensations in a primordial inter- components are unstable and lose matter in the stellar cloud which possessed too much angular form of two streams moving in opposite direc- momentum, or too much mass, for the forma- tions through the bottleneck near L\. There tion of a single star. In this respect the origin can be little doubt that this phenomenon pro- of a multiple system does not differ significantly duces the secular change in the period of 0 from that of a . But the later evo- Lyrae. The loss of matter in /3 Lyrae, if the lution of a binary differs from the evolution of a streams are ultimately expelled into space, must loose cluster. A visual double star is a very amount to about 5X10" gr/sec in order to ex- stable formation and is not likely to change plain the increase of the period. Since the mass very greatly during many hundreds of millions of the star is of the order of 103* grams, the of years, while a galactic star cluster may be expectancy of /3 Lyrae should be of the order of dispersed in the Milky Way in an interval of ten 100,000 years. or a hundred million years (though at least one No other star is known that closely resembles cluster, NGC 67, is believed by H. Johnson and /3 Lyrae (in itself a good indicator of the short A. Sandage to have an age of 5X10* years). duration of its particular stage of evolution). Since we have good reasons for believing that However, many other systems have gaseous some galactic clusters are very young (the asso- rings or streams and many of them, according ciations of Orion and f Persei, the to F. B. Wood, have changing periods, especially in , NGC 6231, etc.) while others are when at least one component is a large much older (, , Praesepe, NGC whose volume fills its loop of the critical zero- 725 and NGC 67), it would be of great interest velocity surface. to know whether any binaries in these clusters One of the most pressing needs in this field is show differences in their properties that might a systematic, prolonged study of the periods of be related to the differences in their ages. As all suitable eclipsing and spectroscopic binaries. yet, our information concerning all types of The photometric observations would be rela- binaries in galactic clusters is too fragmentary tively simple, but should be made with photo- for us to draw any conclusions. electric means. Results of great value would Of special interest is the question of the evo- become available in a few years; but even more lution of close binary systems. From Kuiper's important would be a list of accurate periods work on /3 Lyrae, we know that two types of during the next 10 years which could later be instability may occur, with matter escaping compared with similar lists at intervals of about from one or both components either through the 25 years. We cannot start too soon with such a program! 1 Department of Astronomy, University of California at Berkeley, The spectroscopic observations could, for Calif. » Vol. 47, pp. 15, 121, 1935; vol. 67, p. 887. 1955 many systems, be carried out with telescopes 131 132 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS of moderate size. We need periodic deter- and many more would enable us to locate the minations of the velocity curves of all the regions in which these emission lines are brighter spectroscopic binaries. Here again produced. Joy's original interpretation of RW results may be expected immediately, not only Tauri in terms of a ring around the brighter with regard to the periods but also with regard component of the binary, or my later interpre- to changes in the orbital elements. , tation in terms of a ring of gas surrounding the , 5 Orionis, and many more should be entire system, may have to be modified in the reobserved at least once every 10 years. light of Huang's idea. One of the most puzzling questions in the During the interval 1940-1950 several field of double-lined spectroscopic binaries is McDonald astronomers made a superficial sur- the interpretation of the tendency of many vey of the velocity curves and spectroscopic systems to show a strengthening of the violet- features of about 100 eclipsing variables. Most displaced absorption lines relative to the red- of the systems presented no peculiarities and displaced components, irrespective of which the results were sufficient for statistical studies, of the two stars happens to be approaching. but a considerable number yielded results of This phenomenon was discovered by Miss exceptional interest. A few of these stars Maury at Harvard in the spectra of have since been investigated more exhaustively and V . But later observations at (U Cephei, U Sagittae, SX Cassiopeiae, RX McDonald Observatory failed to show this Cassiopeiae, RZ Scuti), but quite a number variation. Yet, independently, we found that have not yet been scrutinized. I can mention it was conspicuous in Spica and in several other here only a few of them. WX Cephei (J. short-period spectroscopic binaries of early Sahade and C. U. Cesco) shows amazing class B or O. A recent study of v Scorpii by changes in the intensities of the two Ca II S. J. Inglis shows that it is present in the absorption components. They are unlike the helium lines but that it may be reversed in changes discussed above. DN Orionis has an sign in the hydrogen lines. I had previously exceptionally small range of velocity and, assumed that there is more gas in front of therefore, a very small mass function. It is the advancing hemispheres of each component one of the most peculiar binaries in this respect, than in front of the receding hemispheres. resembling XZ Sagittarii (J. Sahade). Both However Inglis' result suggests that it may must possess subgiant components of very small be caused by a difference in ionization and mass (about 0.1 or 0.2©) but of large di- excitation. As far as I know, there are no ameter. These components violate the con- photoelectric measurements that would help ventional mass-luminosity relation. AW Pegasi us in explaining this phenomenon. Since it and V 377 Centauri (Sahade) are even more occurs only in close systems (it is also con- peculiar, and we have been unable to interpret spicuous in many systems) their spectra. Details about these and other I suspect that it is in some way related to the systems may be found in the Astroph3rsical existence of gaseous streams. Journal or in the Contributions from the S. S. Huang has recently asked whether W. J. McDonald Observatory. there might not be accumulations of gaseous The famous visual double star 7 Virginis material at those two Lagrangian points which has been suspected by Lick astronomers to correspond to the equilateral triangle solutions vary in light. Since both stars can be observed of the restricted three-body problem. Unless spectroscopically with high dispersion, a test the mass ratio of the binary is very different of each component (or of the integrated light from one, these solutions are not stable, but of both) for variability would be of interest. there could still be regions of increased density Any whose orbit is known should in the gaseous ring at L% and L4. It is possible be investigated at least in integrated light for that a more careful study of the eclipses of the variability. The discovery of an RR Lyrae emission features in the hydrogen lines of RW or another pulsating component in a double star, Tauri, RW Persei, SX Cassiopeiae, U Sagittae, whose masses are known, would be a major NEW HORIZONS IN ASTRONOMY 133 triumph. It might even be possible from the prepared to find that the giants lose mass by colors (or spectra) and magnitudes to pick out expansion. A. J. Deutsch's work on the ex- promising pairs, but the chances for success panding shell of the system of a Herculis sup- can hardly be estimated. I recommend this ports this hypothesis. important search only to those who have What is the nature of the so-called Trumpler other problems that are certain to give sig- stars? Some of them occur in binary systems nificant results. but we have as yet no firm knowledge of their A few eclipsing binaries must be observed as masses. As Trumpler has shown, their lines nearly continuously as is possible, both photo- have large red shifts which, if interpreted by electrically and spectroscopically. Among the theory of , would give these I should list e Aurigae, , very large masses (or very small radii). There /S Lyrae, UX Monocerotis (R. Lynds suspects are some reasons for believing that their masses that the small, bright, and hot component of are not more than about 75o> But how do this system is a short-period pulsating variable, we then explain the red shifts? while the other component is a large and cool Are all post-novae and all SS Cygni stars subgiant). No spectrographic equipment is close binaries? And why are there so many now available for such continuing observations. Wolf-Rayet stars which are components of I therefore endorse Th. Dunham's project for close binary systems? Are all of these objects the construction of a new spectrographic the products of the evolution of binaries (as telescope. Huang suggests in a recent article in the Astro- I hope that someone will have observed the nomical Journal8) and are the RR Lyrae stars of « Aurigae over a wide range of the products of nuclear evolution of single wavelengths. It is important to know whether stars (or of wide visual binaries) uncomplicated the depth of the eclipse (a totality began Dec. by the loss of mass resulting from the instability 15, 1955, ending 330 days later) is really the of a close binary? It is not yet clear why loss same in all wavelengths. of mass into space (as distinguished from the What is the dispersion of the mass-luminosity formation of streams within a binary, or the relation, expecially at the upper end? If exchange of mass between the components of nuclear evolution proceeds without appreciable a binary) should be significantly different. change in mass, we should expect some scatter We must assume that once a stream of gas is among the O- and B-type stars. Most of our generated in a binary it is somehow expelled present information comes from the velocity from the entire system, and that the absence curves of double-lined binaries, and this prob- of such streams in single, slowly rotating stars lem is looked after by the astronomers at Vic- impedes their loss of mass by "corpuscular toria. Special attention should, however, be radiation." devoted to such stars as J. S. Plaskett's massive It would be unreasonable to conclude this binary, in which the more massive component report without stressing the importance of has the weaker absorption lines; the profiles continuing routine observations of all classes of the lines of this component show irregular of double and multiple stars. In preparing changes. his program, an investigator should not be It is not yet clear whether the yellow and red concerned solely with the solution of immediate giants in Population I obey the conventional problems but should devote a part of his time mass-luminosity relation. K. O. Wright's work to the accumulation of accurate observations on seems to indicate that it does, and which will bear fruit long after his death. my own earlier measurements suggested only The most important astronomical problems— a relatively small departure. D. M. Popper's those of evolution of stars, of perturbations in work on f Aurigae and other similar systems double star orbits, of slow periodic or irregular may soon solve this problem. It may help variations, and many more—require tens or us in constructing the evolutionary tracks of hundreds of years of persistent effort. It is massive stars, but here again we must be •VoLftl, p. 40, 1956. 134 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS especially difficult to sustain an enthusiastic every astronomer has profited from the observa- attitude on the part of a scientist if he does tions of his predecessors; he should therefore not expect to gather the harvest resulting be willing to pass on to his successors a store from his labor. Nevertheless, we must all of good observations to make up for what he follow the dictum of Ejnar Hertzsprung: has inherited from former generations. Stellar Dynamics ByU. Van Wijk1

Although stellar dynamics is largely a theo- attempt to interpret the present kinetic prop- retical field, anyone working in it has a strong erties of the stars in terms of the conditions interest in observational results concerning the under which they were formed. Examples of motions of the stars—their radial velocities and this approach are the papers by Belzer, Gamow, proper motions. and Keller (1951) and van Wijk (1956). Here One of the more active areas of stellar dy- we approach the border line between stellar namics has been the study of the interaction dynamics and the study of galactic structure. between stars and interstellar clouds insofar as But two subjects which are just beyond this it affects the motions of the stars. In papers border line may still be mentioned here because by Spitzer and Schwarzschild (1953) and by of their intimate connection with our present Osterbrock (1952), the effect of randomly dis- subject: the work on pseudointegrals of motion tributed interstellar clouds has been considered. by Lindblad and the study by Blaauw (1952) One may hope that the 21-centimeter band of the dissolution of very loose galactic clusters. observations will ultimately lead to a more specific idea about the distribution of inter- References stellar clouds and that the effect of nonrandom BELZER, J., GAMOW, G., AND KELLER, G. distributions, such as a spiral arm, may then 1951. Astrophys. Journ., vol. 113, p. 166. BLAAUW, A. be considered. 1952. Bull. Astron. Inst. Netherlands, vol. 11, A field which is basically quite fruitful but p. 414. which has been lying fallow recently is that of CHANDRASEKHAR, S. spherical clusters and, particularly, the escape 1943. Astrophys. Journ., vol. 98, p. 54. OSTERBROCK, D. of cluster stars. The papers of Spitzer (1940) 1952. Astrophys. Journ., vol. 116, p. 164. and Chandrasekhar (1943) are still the basic SPITZER, L., JR. references, and it may well be that further 1940. Monthly Notices Roy. Astron. Soc. London, progress will strongly depend on the ready vol. 100, p. 396. availability of electronic calculators. SPITZER, L., JR., AND SCHWARZSCHILD, M. 1953. Astrophys. Journ., vol. 118, p. 106. An interesting recent development is the VAN WIJK, U. 1 Princeton University Observatory, Princeton, N. J 1956. Astron. Journ., vol. 61, p. 277. 135

Variable Stars By C. Payne-Gaposchkin*

Variable stars play a dual role in astronomy. be borne in mind, however, that the Magellanic First, they furnish standards of luminosity and Cepheids may differ in detail from galactic distance. Second, their own physical proper- Cepheids of like period. The study of color ties throw light on energy sources, the processes and spectral variation of galactic Cepheids of pulsation and explosion, and the circum- should therefore be pressed, with especial stances of stellar formation. attention to the effects of selective absorption. Here, again, material from the Southern Standards of luminosity Hemisphere is of particular urgency; the The RR Lyrae stars, Cepheids, novae, and rich region seems comparatively free supernovae have been the most important of absorption, and radial velocities of Cepheids indicators of distance. Their average lumi- in this area will fill out our information on the nosities (mean for the pulsating stars, maximum relation of the Cepheids to galactic rotation, for the explosive ones) are known with an un- and thus on the dimensions of the galactic certainty of a few lOths of a magnitude. We system. are now concerned with determining the disper- The absolute magnitudes of novae have sion of luminosity for stars of these groups. recently been placed on a firm basis by the The absolute magnitudes of RR Lyrae stars light curves of the novae in Messier 31. Prob- will emerge from critical comparisons of the ably the novae in other nearby galaxies will color-magnitude arrays of globular clusters. be the most valuable means of extending this Probably they will be found to be related to the information. metal richness of the cluster stars. The The absolute magnitudes of supernovae, striking differences between the noncluster RR the distinction between the two types, and their Lyrae stars of our neighborhood and those near relationship to ordinary novae must again the , when interpreted in the rest on studies made with the largest telescopes. light of the data from globular clusters, will furnish a new approach to the dimensions of Physical problems our galaxy. These studies, which require Detailed study of the brighter variable spectroscopic analysis and photoelectric ac- stars is continually revealing new types of curacy, can be made only with the largest variation. Future progress in telescopes. The extension of this work to the astronomy calls for understanding the relation- nearby, bright globular clusters of the Southern ships of these types to one another and of the Hemisphere would be most important. physical processes involved in them. While The dispersion in the absolute magnitudes the discovery of variable stars should certainly of Cepheid variables and its relation to color continue, it seems probable that systematic and spectrum is likewise an investigation to study of chosen typical specimens is more be carried out in the Southern Hemisphere. important than indiscriminate search for faint The Cepheids in the Magellanic Clouds furnish variables. However, it should not be assumed perhaps the most urgent and profitable problem that all undiscovered variable stars are faint; in applied variable star astronomy; the study even among the naked-eye stars there may still should be made concurrently with photo- be a number of important examples hitherto electric and spectroscopic equipment. It should undetected. Means for tracing the variations

> Harvard College Observatory, Cambridge, Mass. of brightness of stars that are spectroscopically 137 138 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS interesting, such as shell stars and Of stars, photometry will probably reveal differences of should not fall into abeyance; in this regard composition. Equally important is the devel- the maintenance of adequate sky patrols opment of the physical theory that bears on the (not necessarily with powerful instruments, propagation of waves through the envelopes of as the profitable stars are likely to be brighter these groups of stars. than the tenth magnitude) is a project worthy The long-period and other red variables in- of continued support in both hemispheres. volve similar theoretical problems. Their It is important that the variable stars chosen spectra also invite quantitative observational for systematic investigation be studied with work: the identification and intensities of the greatest available precision. Brightness bright lines are of obvious importance, and and color should be measured photoelectrically. quantitative study of the absorption spectra Spectra should be photometrically analyzed, (admittedly beset with grave observational and all spectra should be suitably standardized difficulties) is also a problem of the first for this purpose. Estimates of the intensities importance. of absorption and emission lines are not good Explosive variables.—The most difficult prob- enough for modern purposes. It should be lem, which can be attacked only with the emphasized that the red and near infrared largest telescopes, involves the physical proc- are now as accessible as the photographic esses of the supernovae, and especially the region, and often provide data that are equally unsolved problem of their spectroscopic changes. or more significant. The study of the novae proper has always Lastly, a star that is varying should be been hampered by the unexpectedness of the studied concurrently by photometric and spec- outbursts. Too often, alas, a has been troscopic methods; this is of especial importance studied from a sense of duty, and each contribu- when the variation is not strictly periodic. tion thus tends to be isolated from related Eclipsing stars.—Of unique importance be- studies. Every suitably equipped observatory cause of their contribution to our knowledge of should have a nova program prepared in ad- stellar masses and dimensions, eclipsing stars vance, and the supplies necessary to put it into have entered a new era with the coming of action. The program should from the first photoelectric photometry. Any eclipsing star include the red and near infrared, and especial that gives a two-spectrum velocity curve and a attention should be paid to photometric good light curve adds to our fundamental standardization. knowledge of stars. The stars are so faint that the Several groups of eclipsing stars give unusual necessary spectroscopic studies require power- promise. The combinations of a blue star and ful equipment. The likelihood that all are a with an extensive atmosphere are double stars and the possibility that many are important in the study of stellar envelopes; in spectroscopic binaries of short period make addition to the well-known specimens now being them of exceptional interest. Photometric actively studied, others are undoubtedly to be studies might reveal that some of them are be found in lists of "composite spectra." eclipsing stars; in addition, it would be impor- The remarkable between the short- tant to learn whether the recently observed period UX Ursae Majoris, the system of Nova rapid changes of SS Cygni are a common feature Herculis, and the pair that constitutes AE in the rest of the class. Aquarii invite a search for such stars, which Irregular variables.—The faint irregular vari- will throw light on the processes of stellar explosions. able stars in nebulosity are destined to play an Periodic variables.—The classical Cepheids and important role in ideas of the origin and develop- the parallel sequence of Population II variables ment of stars. The discovery of these "T are rich in possibilities. Concurrent studies of associations" and the concurrent spectroscopic brightness, color, spectrum, and radial velocity and photometric study of their behavior are will be needed before the interrelationships of extremely important provinces of variable star these groups are clarified. Careful spectro- astronomy. Interstellar Matter By Lyman Spitzer, Jr.1

The study of interstellar matter may be Stromgren at the Yerkes Observatory, is the divided into two main parts. Firstly, there is accurate color measurement for stars of all the measurement of the interaction between spectral types. Accurate measures of absorp- interstellar material and electromagnetic radia- tion lines yield very precise determinations of tion. The objective of this research is to spectral type, and hence of the intrinsic stellar determine, from the observed radiation reaching color. The color excess, or deviation between the earth, the properties of the material the actual color and the intrinsic color, may between the stars. While this study is pri- then be found for any star. With this develop- marily observational, the interpretation of the ment we should be able to determine the observations in terms of the number of absorb- density of small solid particles at points in the ing atoms or of the properties of the absorbing sky very close together, and thus obtain a more solid particles may involve much theoretical detailed picture than is now possible of the analysis. distribution of interstellar material. Secondly, there is the physical analysis of Another suitable program for photoelectric those processes that occur in interstellar space. spectrophotometry is the precise measure of In this study the density, composition, and interstellar line profiles. With this technique spatial distribution of interstellar material are it is possible to determine both the number of obtained from the observational research de- absorbing sodium and calcium atoms in the scribed above, and the various physical proces- line of sight to each star and the precise velocity ses that occur are deduced from the laws of distribution of the atoms. Still another pos- nature, as determined in terrestrial laboratories. sibility is the measurement of much fainter Thus, for example, the internal temperature interstellar lines than are customarily observed. of the small solid particles or grains is com- Thus, detailed information can be obtained on puted theoretically, the interactions between the spatial distribution of such interstellar atoms and grains are analyzed, and the methods atoms as neutral Ca and ionized Fe and Ti. by which interstellar material can condense At present the interstellar lines produced by into a new star are deduced. these atoms can only barely be detected in a At present our information on material in very few stars. space is seriously limited both by observational Radio astronomy is another very important inaccuracies and by a lack of information on observational development in this field. As the the relevant laws of nature. During the next size of radio telescopes increases more and more, few decades significant advances in both these the information about the distribution of neutral respects may be anticipated. hydrogen will become more and more detailed. New experimental techniques promise a Information about other types of atoms may great increase in observational accuracy. Of also be obtained, perhaps, with radio tech- particular importance is the imminent develop- niques. Evidently our exploration of inter- ment of photoelectric spectrophotometry. The stellar clouds will become much more precise combination of a photoelectric cell and a and much more extensive in the years to come. spectrograph will have many important ad- This increasing detail of information, and vantages in interstellar research. One pos- this very great increase in the accuracy of inter- sibility, which is already being explored by stellar maps that can be drawn, will naturally 1 Princeton University Observatory, Princeton, N. J. lead to an expanded effort on the analytical and 139 140 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 theoretical front. For example, data are grad- precise, elaborate, and ingenious laboratory ually accumulating on what might be called approach, together with a considerable amount interstellar meteorology—the great problem of of theory, is probably required to shed light on the motions and transformations of the inter- the situation. The interactions of grains with stellar clouds. As in terrestrial meteorology, atoms, including the rate at which the grains this subject must include not only ordinary grow and are destroyed and the various selec- dynamics but also the effects of radiation, both tive ways in which they interact with different emitted and absorbed. Changes of state, in- types of elements, are of particular interest volving the combination and dissociation of astronomically. Another important problem molecules and atoms, must also be considered. concerns the magnetic properties of such grains, These phenomena are probably influenced pro- the solution of which also probably demands a foundly by the presence of both large-scale combination of observation and theory. magnetic fields and cosmic rays, which have no A central theoretical problem is, of course, counterpart in terrestrial meteorology. Evi- the one of evolution. How do clouds evolve dently the comprehensive analysis of the many over a period of a billion years or more, and, in simultaneous processes will be a vast and chal- particular, how are new stars formed from the lenging adventure. interstellar medium? The detailed answer to Another theoretical problem concerns the these questions is perhaps the most difficult physical chemistry of grains. The boundary part of the entire field, and from a logical stand- conditions in this problem are simple. We point should perhaps be postponed until the know approximately the conditions under which rest of the subject is fully developed. In view atoms in interstellar space gather together to of the very great interest in this particular form solid particles. However, the nature of problem, however, it appears certain that dur- these particles is completely unknown. In par- ing the next few decades much new work will ticular, the detailed structure of the grains, the be done. While no final results may be ex- forces holding them together, and, in fact, all pected in this area for some time, it seems very except the grossest properties of these small likely that ingenuity and hard work will produce solid particles are completely uncertain. A many glittering and intriguing hypotheses. 21-Centimeter Research By Bart J. Bok '

The discovery of the 21-centimeter line of paraboloid at the Naval Research Laboratory neutral hydrogen by Ewen and Purcell (1951) in Washington. Also, most of the work until and the subsequent confirmation of this dis- now has been done with electronic equipment covery by Dutch and Australian groups (Muller with bandwidths of 40 and 15 kc/sec (equivalent and Oort, 1951; Pawsey, 1951) marked the to resolutions in radial velocity of 8 and 3 birth of a new era in the study of our Milky km/sec) and very few observations have been Way system and of the universe of galaxies. made with bandwidths between 5 and 10 That radio radiation with a wavelength of 21 kc/sec (equivalent to radial velocity resolu- cm emanating from neutral atomic interstellar tions of 1 to 2 km/sec). Within a year or so hydrogen might be sufficiently strong to be there should be in operation at least a half recorded was first suggested by van de Hulst dozen radio telescopes with apertures of 60 to (1945) and was predicted independently by 100 feet, all of them steerable and with surface Shklovsky (1949). But not until the actual tolerances sufficiently strict to permit full use discovery of the line did astronomers begin to at the 21-cm wavelength. Research at 21 cen- realize how powerful a tool for future research timeters has just been started with two large had become available. radio telescopes—the 83-foot instrument at Research in the 21-cm field may at present Dwingelo in Holland and the 60-foot George R. be divided roughly into three broad areas. radio telescope of Harvard Observatory The first concerns the spiral structure of our (see the recent article by Heeschen, in press). galaxy, and has yielded spectacular first results. Several of these instruments apparently will Really detailed analysis, however, has only have electronic equipment capable of record- begun, and will undoubtedly be fruitful in ing 21-cm profiles with a bandwidth of 5 kc/sec future studies of spiral structure. The second (equivalent to a resolution in radial velocity of category is that of related optical and radio 1 km/sec). The group at the California Insti- studies of the fine structure of the interstellar tute of Technology is planning to use two 90- gaseous medium in our galaxy. We shall list foot antennas as an interferometer with variable below some of the topics of research most separation of the two component antennas. important here, but it is already clear that the Moreover, three very much larger paraboloids "21-cm astrophysicist" has waiting for him a intended in part for use in 21-cm research are multitude of specific research problems. The now either under construction or on the draw- third category is that of extragalactic 21-cm ing board: the 250-foot paraboloid for Jodrell research, one in which the work has barely Bank, the 140-foot paraboloid for the Inter- started but which obviously will attain increas- University Radio Observatory now under con- ing importance as the reach of our equipment sideration by the National Science Foundation, increases. and the large instrument planned for Australia. One should bear in mind that up to the present all but one of the five principal radio Spiral structure of our galaxy telescopes that are active in the 21-cm field are The group has contributed most paraboloids with apertures of the order of 25 effectively to the study of the spiral structure to 30 feet, the one exception being the 50-foot of our galaxy. The early work by the Leiden group culminated in the famous joint paper by 1 Harvard College Observatory, Cambridge, Mass., and Common- wealth Observatory, Canberra, Australia. van de Hulst, Muller, and Oort (1954) which 141 142 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS contains the picture of the spiral structure of begun a study rather similar in scope to that of our galaxy in its broad features as observable Schmidt. It seems established that one Inner from a Northern Hemisphere radio observatory. Arm is present which presumably is identical For the Southern Hemisphere a preliminary with the optically discovered Sagittarius Arm, survey has been completed by Christiansen but further evidence exists for the presence of and Hindman (1952), and the most recent some structure in the internal neutral-hydrogen results were presented by (1955) conglomerate, close to the center of our galaxy. at the Troy meeting of the American Astro- Two major problems arise in any analysis of nomical Society. Since the original Dutch spiral features from 21-cm profiles. First, a survey was completed, there has been con- large group of problems relates to the probable siderable activity in the study of the spiral shape of the curve that gives circular velocity structure for the part of our galaxy between of rotation as a function of distance from the galactic longitudes 40° and 200°; that is, from galactic center. In 21-cm research we have no Cygnus to well beyond . The most direct way of obtaining distances, and we can extensive research has been done by Westerhout only assign to a given observed feature a dis- (in press), who has derived and analyzed 21-cm tance based on its observed radial velocity of profiles for the entire section between galactic approach or recession. If our assumed form latitudes —10° and +10°. To supplement the for the curve relating distance from the galactic work of Westerhout, we have the published center to circular velocity of rotation is in and unpublished results of Matthews (1955) error, then all of our derived distances obvi- and of Davis at Harvard and of Heifer and ously will be in error also. A second and related Tatel (1955) at the Carnegie Institution. All cause for difficulties lies in the fact that, on the these researches indicate that the simplified basis of 21-cm observations alone, we have no descriptions of three primary spiral arms in this way of disentangling effects of peculiar motions section, the , the , and from those caused by purely circular galactic the Distant Arm, are only very rough first ap- rotation. proximations. The available evidence sug- We should be able to outline on a fairly gests major branching of the principal spiral homogeneous basis the general features of the features; it is evident, furthermore, that all of spiral structure of our galaxy once the southern the spiral features are definitely not precisely Milky Way observations are completed and in the same plane. Comparable detailed studies analyzed as have been the profiles for the north- for the Southern Hemisphere are still lacking, ern Milky Way by the Leiden observers. but the first results by the Australian groups Great interest attaches to the future clarifica- suggest that here, also, complexity rather than tion of the details of spiral structure, but here simplicity are to be found. progress will be slow at best, and, initially at The study of the spiral structure in the part least, somewhat uncertain. We have already of our galaxy interior to our sun offers a variety indicated that the dynamical study of the law of complex problems. The principal source of of circular galactic rotation is imperative if difficulty is that, even on the simple assumption we wish to derive meaningful distances for the of purely circular galactic rotation, a feature features found through our 21-cm observations. with a given radial velocity may be produced To deal conclusively with problems of the by hydrogen clouds at either one or another branching of spiral arms and with fine structure distance; or, more generally, that a particular phenomena, our only hopeful approach appears feature may well be the blurred combined to be through coordinated optical and radio impression produced by features at two different studies of sections of the Milky Way, one by distances from the sun. The early results one. For proper analysis of 21-cm features, obtained by the Leiden group have been we shall need extensive and precise optical summarized in a paper by Kwee, Muller, and data—for example, for OB stars and associations Westerhout (1954); and, more recently, Schmidt in the area of the sky covered by a particular (in press) has obtained useful new evidence. At 21-cm study. For proper interpretation of the Harvard, Heeschen and Howard (1956) have 21-cm features, we should at all times attempt •at NEW HORIZONS IN ASTRONOMY 143 to obtain radio profiles for a fairly wide range Lilley found the ratio between the average of galactic latitudes and for points in the area densities of gas and of dust to be close to of investigation spaced as close together as 100:1 inside the Complex. Lawrence, appears to be reasonable for a particular radio Menon, and I (Bok et al., 1955) used the 24- telescope. Preliminary work in this field by foot paraboloid to study some very densely Matthews (1955) has already demonstrated obscured spots inside the and that the results of closely spaced surveys will Taurus Complexes. We found no marked be well worth the effort, and that correlated differences in 21-cm signal strength when we radio and optical studies will provide insight compared intensities for the highly obscured into the fine details of spiral structure. Fur- fields with those found for neighboring fields thermore, a great deal of new information should still inside the complexes but definitely less be obtainable regarding the fine structure of obscured. We deduced from our observations the interstellar medium and the association that the large dust complexes are regions of between interstellar hydrogen and certain markedly high concentration of interstellar classes of stars. A related problem of great atomic neutral hydrogen, but that the dis- interest is the relative distribution of neutral tribution of the atomic hydrogen inside the hydrogen and inside a spiral arm, complex does not mirror precisely that of the or branch of an arm. cosmic dust. Existing 21-cm observations suggest that the neutral hydrogen is more Astrophysical studies evenly distributed than the cosmic dust, but The 21-cm astrophysicist is essentially a no really significant conclusions can be drawn spectroscopist who has for study only one until the really large paraboloids become avail- single spectrum line. Fortunately, the micro- able for studies of this nature. A careful wave spectrographs presently in use are really search for neutral hydrogen concentrations by high-dispersion instruments and thus much 21-cm interferometer techniques should prove can be learned from a single 21-cm profile. of great interest in connection with this prob- The 21-cm astrophysicist realizes that he is lem; a negative result would be fully as im- the fortunate fellow who has first achieved portant as a positive one. When the necessary access to the all-pervading medium of neutral observations have been made, their interpre- hydrogen of our galactic system, which means tation will still present a good many difficulties. probably as much as 95 percent of all the Our lack of information regarding the presence interstellar hydrogen. In this section, we shall of molecular hydrogen complicates analysis describe first some astrophysical problems that considerably, and as yet the possible effects have proved practicable for paraboloid anten- of variations in temperature inside the large nas with apertures of 20 to 30 feet, then turn dust complexes are by no means understood. to problems whose solution is aided by antennas The assumption that the interstellar medium with apertures of 50 to 80 feet, and conclude of neutral atomic hydrogen possesses one single with some astrophysical problems that we may characteristic temperature is at best a very attempt to study with antennas of 100-foot crude one, and in future research temperature and larger apertures. variations undoubtedly will have to be con- The work of Lilley (1955) and of Heeschen sidered along with density variations. (1955) at Agassiz Station illustrates the first A second area in which 21-cm studies have group—astrophysical problems that may be already made important contributions is that studied with antennas of relatively small of the study—by combined optical and radio aperture. For the Taurus Complex of dark techniques—of regions of special interest, such nebulosity, Lilley found generally far stronger as the Orion region. The recent work by signals for fields well inside the dark complex Menon (in press) illustrates how much we can than for relatively unobscured fields bordering learn through detailed 21-cm studies. Op- upon the complex. It thus became evident tically, the Orion region is known by its that the large dust complexes have excess separate features: the Great Nebula in Orion, neutral hydrogen gas associated with them. with a total estimated mass not in excess of 144 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 1,000 solar masses; the and becomes impracticable. Once we have per- various OB associations and aggregates; fected our equipment, success in breaking down 's Luminous Arc of very great dimen- the interstellar medium into its basic compo- sions; and, finally, the large dust complex that nents should be within our grasp. is evidently obscuring much of the region. The The greatest contribution so far to 21-cm 21-cm observations by Menon show that all of research by the 50-foot paraboloid of the Naval these features are imbedded in a large "bowl of Research Laboratory has been in the study of neutral hydrogen jelly," with a total mass absorption features observed in distant discrete probably of the order of 60 to 100 thousand radio sources. Hagen, Haddock, McClain and solar masses. Menon's observations suggest Lilley (Hagen, et al., 1955; Lilley, 1955) have that this entire mass is in slow rotation and studied 21-cm profiles and related phenomena that it is, very likely, expanding at the rate of in the continuum at wavelengths of 21 cm and 8 to 10 km/sec. Here the 21-cm approach has less for a number of discrete sources, some of already added a dimension to our earlier un- thermal origin, others not. The first relevant related optical studies. The astrophysical and 21-cm profiles have exhibited certain amazing cosmogonical importance of the new approach absorption features which tend to confirm the is obvious. cloud or turbulent nature of the interstellar The third group of problems that are proving gaseous medium. For the future study of amenable to study by 21-cm techniques deals these profiles it will prove profitable to insist on with the fine structure of the interstellar the highest attainable angular resolution ac- hydrogen medium. The studies by Lawrence companied by very narrow bandwidth of the (in press) and by Heeschen and Drake (1956) recording equipment. Almost every one of serve to illustrate two different approaches. these sources studied has presented us with a Lawrence, working in close collaboration with different type of problem. The neatest ab- Munch, has studied at the Agassiz Station sorption features are probably shown by the 21-cm profiles for fields centered upon stars for Cassiopeia source, and it seems difficult to which Munch, at Mount Wilson and Palomar, escape the conclusion from the 21-cm observa- had found multiple optical interstellar lines. tions by the Naval Research group and by Two established coincidences between radial (in press) and his associates at Jodrell velocities of 21-cm features and strong inter- Bank that the source is in the Perseus Arm rather stellar absorption lines indicate that the same than within 1,000 parsecs of the sun. The interstellar gas clouds may be responsible. study of the source at or near the galactic Further study of these phenomena should prove center—or at least in the direction between our exceedingly rewarding, but we can hope to sun and the galactic center—offers a host of arrive at a clear understanding through 21-cm possibilities for future research and here very studies only with radio telescopes of large much greater angular resolution than is attain- apertures with associated electronic equipment able to date is urgently required. We shall of very narrow bandwidth. The research of return in the next section to the Cygnus source, Heeschen and Drake refers to the general which exhibits absorption features that origi- direction of the Pleiades Cluster. They find a nate far beyond the bounds of our own galaxy. double-peaked 21-cm profile—the radial veloc- But the brief references that we have given ity of one peak coinciding almost precisely with above indicate that 21-cm techniques offer that of the Pleiades stars, and that of the great opportunities for future research. The other peak being very nearly equal to the mean contemplated interferometer for 21-cm re- radial velocity for the strongest components of search, with 2-component 90-foot steerable the interstellar lines. In all these investigations paraboloids, should make possible an entirely of the fine structure of the interstellar medium, new approach to the study of discrete sources one must insist on high angular resolution ac- and their absorption features. Large parabo- companied by narrow bandwidth; without loid antennas with apertures in excess of 100 narrow bandwidth, the whole study of the feet also promise very significant contributions. distribution of individual gas-cloud velocities In cases where the 24-foot antenna is useless, the NEW HORIZONS IN ASTRONOMY 145 50-foot antenna begins to suggest; but the 100- portant in establishing the presence of a 21-cm foot antenna should be able to show the way and neutral hydrogen medium surrounding the the larger antennas should really be able to solve Cygnus discrete source, but the agreement some of the critical galactic astrophysical between the optically predicted red-shift veloc- problems of 21-cm research that are ahead of us. ity and the radial velocity determined by radio techniques has important consequences for Extragalactic research cosmogonical theory. To the Australian group at Sydney (Kerr, The 21-cm radiation from the Andromeda Hindman, and Robinson, 1954) and their Nebula and other nearby galaxies remains as associates (Kerr and de Vaucouleurs, 1956) yet undiscovered. Special equipment to aid goes the credit for having detected and studied in its detection and measurement is now under the first 21-cm radiation from beyond our own construction in Holland and elsewhere. The Milky Way system. Their studies of the importance of 21-cm studies for individual Magellanic Clouds, which have been blended galaxies of the can hardly be over- skillfully with the optical researches carried rated. If we can attain angular resolution on at Canberra (Buscombe, Gascoigne, and de sufficient to separate areas inside spiral arms Vaucouleurs, 1954), have demonstrated con- from those between the arms, very significant clusively that the 21-cm technique offers contributions can be made to the solution of entirely new possibilities for the study of the whole group of problems related to the galaxies beyond our own Milky Way system. dynamics of galactic rotation and of spiral These studies have already made important structure. We shall then have also wholly contributions both to the dynamics of the new evidence on the distribution of neutral Magellanic Clouds and to our knowledge of atomic hydrogen in spiral and other galaxies population distributions. Until it was found and new light will be thrown on the problems that the 21-cm signals from the Small and from of stellar birth and evolution. For the future, the Large Magellanic Clouds are comparable 21-cm research of nearby galaxies, notably in strength, many had doubted that the Small those with spiral structure, has a high priority. Cloud had anywhere nearly as much neutral Quite recently Heeschen (in press) has suc- hydrogen as the Large Cloud. Through the ceeded in detecting 21-cm radiation associated 21-cm study we are learning a great deal about with a cluster of galaxies. At a frequency near the dynamics of the two Clouds. Their rota- 1387 mc/sec, corresponding to a recessional radial tion seems well established, and the masses velocity of close to 7000 km/sec, he found evi- derived from rotational velocities are in good dence for a weak but consistently present signal agreement with those estimated in other ways. (&T=2°K±0.°2) from the direction of the Studies of the distribution of random velocities cluster of galaxies. The specifically promise to add materially to our knowledge of determined radial velocity of recession for the the dynamics of the gaseous interstellar medium same cluster is 6680 km/sec, which appears to in these two galaxies. indicate satisfactory agreement. The disper- The second big contribution to date to the sion in radial velocity derived from the 21-cm study of radiation from outside our galaxy observations appears to be of the order of 500 was made by Lilley and McClain (1955) in km/sec, somewhat smaller than the optical value their work on the principal absorption feature of about 800 km/sec. The total estimated mass in the discrete source in Cygnus. They have of neutral atomic hydrogen is a little less than established that an absorption feature in the 10" solar masses. Heeschen's result was ob- spectrum of the Cygnus source is present at tained with the 24-foot radio telescope at precisely the frequency predicted from the Agassiz Station. Information regarding the optically known velocity of recession, produced distribution of neutral hydrogen inside the by the red-shift. The absorption feature cluster can be obtained only with instruments obviously originates in the galaxy inside which of greater aperture, which afford higher angular the Cygnus source itself is imbedded. Not resolution. only is the work of Lilley and McClain im- A potentially very profitable approach to the 146 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS problems of intergalactic hydrogen is through ourselves to the single 21-cm line, we could have the study of absorption features. Field and available additional lines, preferably those that Heeschen have recently attempted to observe might give clues to the presence of atoms other in the spectrum of the Cygnus source absorp- than hydrogen, and components of bands that tion features produced by the general medium would permit the study of molecules, such as of intergalactic hydrogen between the Cygnus OH and H2. The field of galactic and extra- source and our sun. The results with present galactic microwave spectroscopy is indeed a equipment are inconclusive, but the analysis fruitful one for future exploration. of the available data shows that very rarefied absorption features might be detectable, even References for the low densities that presumably prevail in BOK, B. J., LAWRENCE, R. S., AND MENON, T. K. the intergalactic medium. 1955. Publ. Astron. Soc. Pacific, vol. 67, p. 108. We should at this point refer again to the BUSCOMBB, W., GASCOIGNE, S. C. B., AND DE VAU- COULEURS, G. interferometer array now under construction 1954. Australian Journ. Sci., vol. 17, p. 3. by Bolton and his associates at the California CARPENTER, M. S. Institute of Technology. A variable-spacing 1955. Paper presented at Troy meeting of the interferometer, with two basic components Amer. Astron. Soc. of 90-foot aperture each, should make acces- CHRISTIANSEN, W. N., AND HINDMAN, J. V. 1952. Australian Journ. Sci. Res., ser. A, vol. 5, sible for study a number of discrete sources p. 437. running literally into the hundreds. From DA VIES, R. D. the absorption features, much information . Paper presented at Jodrell Bank Sympo- should be obtainable—information relating not sium 1955, in press. only to the clouds between our sun and these EWEN, H. I., AND PTTRCELL, E. M. 1951. Nature, vol. 168, p. 356. discrete sources but also concerning the dis- HAGEN, J. P., LILLET, A. E., AND MCCLAIN, E. F. tances of individual discrete sources. 1955. Astrophys. Journ., vol. 122, p. 361. HEESCHEN, D. S. Concluding remarks 1955. Astrophys. Journ., vol. 121, p. 569. It is now just about 5 years since Purcell and . Astrophys. Journ., in press. HEESCHEN, D. S., AND DRAKE, F. D. Ewen first succeeded in detecting the 21-cm 1956. Astron. Journ., vol. 61, p. 5. line of neutral hydrogen. In the intervening HEESCHEN, D. S., AND HOWARD, W. E., III. years much progress has been made, and I 1956. Astron. Journ., vol. 61, p. 6. doubt if many would have dared to predict in HELFER, H. L., AND TATEL, H. E. 1951 that by 1956 equipment equaling in excel- 1955. Astrophys. Journ., vol. 121, p. 585. lence present-day radio telescopes and electronic KERR, F. J., AND DE VAUCOULEURS, G. recording equipment would have become avail- 1956. Australian Journ. Phys., vol. 9, p. 90. able at several institutions. But although we KERR, F. J., HINDMAN, J. V., AND ROBINSON, B. J. 1954. Australian Journ. Phys., vol. 7, p. 297. may look with some satisfaction upon our KWEE, K. K., MULLER, C. A., AND WESTERHOUT, G. accomplishments to date, in reality 21-cm 1954. Bull. Astron. Iiist. Netherlands, vol. 12, research has only just begun. No. 458, p. 211. I have purposely confined my remarks to LAWRENCE, R. S. 21-cm research in a most limited sense. I have . Paper presented at Jodrell Bank Sympo- studiously avoided referring to the importance sium 1955, in press. LILLET, A. E. of studies in the continuum at wavelengths of 1955. Astrophys. Journ., vol. 121, p. 559. 50 cm and less, which must become an integral LILLET, A. E., AND MCCLAIN, E. F. part of all 21-cm studies. I have also omitted 1955. U. S. Naval Res. Lab. Rep. No. 4689. all reference to other spectrum lines in the MATTHEWS, T. A. microwave range, notably to the line of deu- 1955. Publ. Astron. Soc. Pacific, vol. 67, p. 112. terium, the discovery of which would open up MCCLAIN, E. F. an entirely new field for research of the inter- 1955. Astrophys. Journ., vol. 122, p. 376. MENON, T. K. stellar medium by radio techniques. It would . Paper presented at Jodrell Bank Sympo- be very helpful if, instead of having to confine sium 1955, in press. NO.I NEW HORIZONS IN ASTRONOMY 147

MlJLLER, C. A., AND OoRT, J. H. VAN DB HULST, H. C. 1951. Nature, vol. 168, p. 357. 1945. Nederlands Tijdschr. Nat., vol. 11, p. 210.

PAWSBY, J. L. VAN DE HULST, H. C, MUUJSR, C. A., AND OOHT, J. H. 1951. Nature, vol. 168, p. 358. 1Q54 Bull Astron In8t Netherlands, vol. 12, SCHMIDT, M^ NO y . Paper presented at Jodrell Bank Sympo- sium 1955, in press. WESTEBHOUT, G. SHKLOVSKT, I. S. • Paper presented at Jodrell Bank Sympo- 1949. Astron. Zhurn., vol. 26, p. 10. sium 1955, in press.

Some Observational Problems in Galactic Structure By Harold Weaver1

The past decade has brought many new in- physical specification; color is frequently sub- sights into the problems of galactic structure. stituted for spectral type. If we fix the space New observational techniques—optical and variables and represent the distribution graph- radio—have produced results that were thought ically we obtain the Hess (1924) diagram. far beyond reach only a few years ago. We The space distribution of nonstellar material now realize the importance of the various may be represented by Dt(R, 0, z), which refers structural units of the galaxy (clusters, associ- to particles of kind i. ations, arms, etc.) as groups of evolving stars For R small or z large in D(M, S; R, 6, z) one of different ages. Data on the structure and finds pure or essentially pure Population II. the kinematic properties of the galaxy provide If to M, S are assigned ranges encompassing basic information for studies of stellar evolu- intrinsically bright O and early B stars, Baade tion; clusters and associations provide tests of found that in extragalactic systems the locus of stellar-model evolutionary calculations. It is maximum of D(M, S; R, 0, z) traces out spiral clear that we must relate our galactic studies arms where Population I stars and much inter- to those of other nearby resolved galaxies. stellar dust is to be found. Population I, Knowledge will advance most rapidly if we occurring in regions where z is small and R not understand how the structural details of our small, appears as an addition to, and inter- galaxy relate to those of the sequence of galaxies. mingled with, the general substratum of Popu- lation II stars forming the disc of the system. New approaches 2. Morgan, Sharpless, and Osterbrock's Several recent observations, to a considerable (1952a, 1952b) use of Baade's spiral tracer degree, have reoriented the whole trend of technique to produce the first picture of spiral thinking about galactic structure and have structure in the galaxy. forcefully indicated the need for new approaches 3. Oort, van de Hulst, and Muller's (1952) to the problems. Among these observations observation that the 21-cm radiation of hy- are the following: drogen indicates hydrogen concentration in 1. Baade's (1944) decisive demonstration spiral arms whose form and position confirm that the space and physical-specification vari- and extend the work of Morgan and his ables in the space-density function D(M, S; associates. R, 9, z) are strongly correlated. 4. Ambarzumian's (see Kourganoff, 1951) The stellar space-density function D(M, S; recognition of stellar associations and of their R, 0, z)dMdSdV will play an important part importance in galactic and stellar evolution in the discussion that follows; it indicates problems. symbolically the space density of stars in ab- In the following discussion we shall con- solute magnitude range M to M-\-dM, spectral- centrate attention on the problems that fall type interval S to S-\-dS, and volume element principally under 1 and 2; those that fall dV, centered at galactocentric position R, 0, z. principally under 3 and 4 will form the subject Other or additional variables can be used for matter of other papers. Effort will be made

1 , University of California, Berkeley, Calif. to avoid duplicating material included in 149 381014—56 -11 150 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Blaauw's (1955) excellent summary of the Stellar distance scale.—The present scale of International Astronomical Union symposium, absolute magnitudes (Keenan and Morgan, "Coordination in Galactic Research," which 1951) is not entirely independent of galactic should be read by all investigators in this field. rotation constants (first order theory only). In this discussion we shall critically examine These, in turn, are dependent upon an older the results of some recent researches in galactic absolute magnitude scale and on corrections structure; and, rather than stress the agree- for average interstellar extinction. Work of a ments, we shall stress the discrepancies between fundamental nature is here necessary to elim- observation and theory and between the inate the possibility of a closed circuit of observations themselves. While this approach systematic errors that do not easily reveal will accomplish our purpose—to expose areas themselves. Fundamental absolute-magnitude where problems need to be worked on—it scales are required—scales derived from stellar will show the field of galactic research in an associations, clusters, moving clusters, and untrue light. The reader must guard against stars of large parallax (with the latter two gaining the false impression that little has types of objects serving for establishment of been accomplished in the field. The fact is, zero point). The degree of agreement between rather, that much remains to be done. magnitudes (and also luminosity class IV, III, . . ., magnitudes) derived from Form of the arms different associations, clusters, and so forth, The form of the spiral arms first indicated by will be of extreme interest; comparisons with Morgan and his associates, using emission field stars in the solar neighborhood will like- nebulosities, and by Oort, using hydrogen wise be of importance. Several investigators concentrations, becomes increasingly complex are engaged in observations of this type and and less easily understood as additional observ- are producing very interesting results. ational data accumulate. Further work, both Observations of the ratio of total interstellar observational and theoretical, in detailed de- extinction to color excess are contradictory. lineation of the spiral arms by stars and by On the basis of both photographic and photo- gas is required. The Morgan-Meinel-Johnson electric studies (Baade and Minkowski, 1937; (1954) technique of very short spectra pro- Stebbins and Whitford, 1945; Sharpless, 1952), vides a powerful method of discovering distant the has been pointed out as spiral tracers. Southern Hemisphere observa- possessing an extinction law appreciably differ- tions of the 21-cm radiation are especially ent from that of other interstellar material. needed; 21-cm observations of high resolving A photographic spectrophotometric study of power in angle and in frequency are needed in O and B stars by Divan (1954) indicates both hemispheres. Such observations will, precisely the same extinction law in the Orion among other things, provide data for determin- region as in all other regions investigated. ing the forms of the gas condensations, and the On the other hand, very accurate photoelectric degree of association between OB stars and gas. observations by Johnson and Morgan (1955) indicate an "abnormal" ratio of extinction to The distance scale within the galaxy color excess for the Cygnus region in the sense The distance scale presently employed in tracing that relatively more ultraviolet extinction is spiral structure by means of OB stars is correct present in that region than in the - only to the extent that (1) absolute magnitudes region. Space variations of the and colors of the MK classification system are extinction law complicate photometric dis- correct, and (2) the ratio of total interstellar tance calculations and pose difficult problems extinction to color excess is known. for the theoretician attempting to explain the The distances of the hydrogen condensations distribution of particle size in the interstellar are correct only to the extent that (1) an material. The observational determination accurate differential-rotation function Aw(R) is and subsequent theoretical discussion of the available, and (2) the interstellar gas is free wavelength dependence of interstellar extinc- from large-scale peculiar motions. tion requires further attention. NEW HORIZONS IN ASTRONOMY 151 Hydrogen distance scale.—The differential parameter, 8.16 kpc, derived from the distribu- rotation function Au(R) for R^>R0 must be tion of RK Lyrae variables in the region of the derived from stellar motions or from the mass galactic nucleus. The excellent agreement of distribution in a model galaxy; the two values does not necessarily confirm is assumed in either case. If derived from their correctness. Baade's estimate could be stars, Aco(i?) is dependent upon the accuracy appreciably changed by a small revision of the of the stellar distance scale. For R

The product J?o«o is smaller than the Vo value gen. For the stars, accurate photometric dis- derived by independent fundamental methods. tances and radial velocities can be derived for Examination of Ro and w0 is required. Of the objects nearer than r0, an observationally set two quantities, Ro is, percentagewise, the more limit. Over the region of space thus defined uncertain. (shown in fig. l,a) the motion pattern can, at The u(R) curve derived by Kwee, Muller, and least in principle, be derived and regional de- Westerhout (1954) from 21-cm observations partures from circular motion deduced. Hy- shows waves which, they suggest, represent drogen observations provide no distance scale departures from circular motion which are independent of velocity; they permit study of correlated with position with respect to spiral departures from circularity only over the vol- structure. The general question of systematic ume of the galaxy in which tWx occurs (shown deviations from circular motion over sizable in fig. 1,6) in the range Ri

GALACTIC CENTER \ CENTER /

FIGURE 1.—The plane of the paper represents the plane of the galaxy, a, Crosshatched area indicates the region over which stars permit the study of departures from circular motion, b, Crosshatched area indicates the region over which 21-cm radiation of hydrogen permits study of departures from circular motion. volumes of the galaxy merits investigation on a galactocentric distance Rx is fixed at the point broad basis. Observations of both hydrogen at which strong turbulence in the interstellar and stars are desirable. Of particular interest hydrogen sets in. In the case of hydrogen, a will be tests for similarity of kinetic properties large peculiar motion occurring in the line of of stars and gas in the same volume of space and sight, but not within the indicated volume, for a correlation between gas and stellar motions could obscure a smaller peculiar velocity within and spiral structure as suggested by Kwee, the indicated volume and thus confuse the ob- Muller, and Westerhout. servation. To a lesser degree, variations in the Study of departures from circular motion is distribution of random velocities in the hydro- likely to prove easier with stars than with hydro- gen also obscure the situation. Such difficulties NEW HORIZONS IN ASTRONOMY 153 do not exist for the stellar observations. Hydro- even of a rudimentary kind, for details in a few gen observations, on the other hand, extend to properly chosen extragalactic systems would be a much greater distance than do stellar ones. of great interest. A start on this problem Comparison of flmax values derived at corre- could be made by measuring magnitude and sponding angles on either side of the galactic color differences between arm and interarm center will be of very great interest. regions of some of the most resolved extra- galactic systems. Association of spiral tracers and hydrogen Spiral tracers are highly concentrated in the No adequate test has yet been made of the com- arms; data on the space distribution of later- monly accepted hypothesis that arms or struc- type Population I stars, B 3, B 5, . . . , in and tures delineated by very early type stars are near the arms is lacking. We have no knowl- coincident in space with hydrogen gas density edge of how prominent the arms would appear maxima. There are indications that these may if the galaxy were viewed through a series of not always be coincident (Weaver, 1953; van de hypothetical filters, each of which would permit Hulst, Muller, and Oort, 1954). OB stars may us to see stars in only a small spectral-type appear where no hydrogen density maximum range, B 3, B 5,. . . . The ordinary method of occurs. analysis of star counts does not offer much Several tests of this hypothesis can be pro- promise of producing such a picture of the posed. The simplest involves investigation of galaxy. Rather, we should obtain the picture the simultaneous occurrence of hydrogen max- with the aid of accurate individual photometric ima and concentrations of early type stars over distances derived for each member of a large the surface of the sky. This test was inde- sample population of stars. An empirical model pendently proposed by W. W. Morgan during of the space distribution of any one particular a colloquium given at Berkeley, Calif., in Novem- type of star could then be constructed. Data ber 1955. An alternative test involves the such as density gradient perpendicular to the agreement of velocity maxima of the radial galactic plane, in the galactic plane, and per- velocity distributions of the OB stars (treated pendicular to the arm, and so forth, could be as particles in a star gas) and of the hydrogen inferred from the model for each stellar type. gas, properly weighted to match the distribu- An approach through such an empirical tion of the stars in longitude. A third test model avoids the concepts of a continuous involves the determination of the closeness of space-density law and of a space-invariant association in space of maxima in the space luminosity function. It does not involve the density functions of OB stars and of interstellar solution of an integral equation for the space- hydrogen as found from the 21-cm observa- density law or the correction of the apparent tions. Distances of the stars are to be derived space-density law with the aid of an average photometrically; distances of the hydrogen interstellar extinction relation in a specified maxima are to be determined from a rotation direction. Each star is treated individually; function based on the same photometric dis- maximum resolution is achieved. A limit of tance scale used for the stars. Different completeness in absolute magnitude M(r) can assumptions are involved in the three cases; be specified at each distance; the space density all three tests should eventually be made. of intrinsically faint stars is not inferred from a few intrinsically bright ones as when the Composition of the arms integral equation is solved. Understanding of the physics of the spiral Extensive observations are required for such arms will require, among other data, knowledge a program; a cooperative venture involving of their composition. Nearby extragalactic sys- several observers and telescopes might prove tems will undoubtedly play an important role advantageous. Stromgren's (1951, 1954) ac- in composition studies. Photographs of M 51 curate photoelectric method of spectral classifi- published by Zwicky (1955) are instructive; cation should play an important part in any analogous photoelectric measurements would be such plans. Stromgren's results indicate that invaluable. Isophotes and isocolor contours, the photometric probable error of one observa- 154 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS tion of a predicted absolute visual magnitude information on the space variation of D(M, is less than 0.2 mag; in distance estimate a S) R, 6, z) over the M-range and r-range probable error of approximately 15 percent accessible to observation. However, such a might be expected. With the 82-inch program now represents only a possible future McDonald telescope, observations could be ex- development; information on stellar popula- tended to apparent magnitude 14. Further tions which will provide data for stellar evolu- exploratory work in the use of this important tion studies must be provided more rapidly. observing technique is called for. The results Among observable quantities that provide obtained should provide data for the establish- such information with less completeness than ment of an empirical space distribution as a D(M, S; R, 6, z) but that are much more test. easily obtained are the following: The need of an accurate and fundamental 1. The luminosity function, L(M; R, 6, z), calibration of the absolute magnitudes and which is the integral of D with respect to S. colors of the MK spectral types and luminosity 2. Integrated magnitudes, colors, or spectra, classes as well as an accurate value of the ratio which represent, essentially, integrals of the of total extinction to color excess is reemphasized luminosity function over specific wavelength by a program of this type. ranges and which refer to a defined volume element of the galaxy or of some extragalactic Variations in the arms system. Within the spiral arms are found variations in 3. The ratio of number of stars brighter than the space density of stars and in the stellar some fixed M and within a defined volume mixture. Observational data leading to a element to the total luminosity of all stars better understanding of such variations would within the same volume element (Baum and be of considerable value. Schwarzschild, 1955). The wavelength ranges A pronounced example of the first kind of in which M and the total luminosity are meas- variation is found in the longitude range 110° ured must be specified. to 145° adjoining the rich Perseus star cloud. 4. The ratio of mass to luminosity (measured In this range of longitude there exists a deep in a specified wavelength range) of the material minimum in the number of OB stars (Weaver, within a defined volume element of the galaxy 1953); there is essentially a discontinuity in or of an extragalactic system. the Perseus arm as delineated by early-type 5. The ratio of luminosities of stars con- stars. tained within some defined volume clement of Variations in the content of the arms fre- the galaxy or an extragalactic system measured quently appear. The Perseus cloud contains O in several different specified wavelength ranges stars and late-type supergiants (Bidelman, 1943, to the luminosity of Uio same stars in some 1947); classical Cepheids are found within it. standard wavelength range. The Orion association contains principally main Basically, the integrated quantities represent sequence stars of type O 6 and later; it contains certain average properties of the stellar popula- no supergiants of type later than B (Sharpless, tion in the defined volume element. The ratios 1952, 1954). The association, which compare one portion of the population with has M 11 as its nucleus, contains main sequence another portion of the same population. The stars of type B 9 and later, and ordinary giants particular ratio taken fixes the portions of the (Russell, 1953). Gurzadian (1949) has pointed populations compared. All of these observable out many cases of this sort. quantities need further investigation and appli- A preliminary study of variations in the cation within our own galaxy and extragalactic arms might be made on the basis of galactic systems. clusters and Trumpler's cluster types. The luminosity function of stars characteriz- ing the solar neighborhood has been determined Luminosity functions and related quantities by van Rhijn (1936), by Luyten (1939, 1941), An observing program of the type outlined in and others. Van Rhijn's function shows a the preceding sections would provide detailed minimum at absolute magnitude -f 8; Luyten's NEW HORIZONS IN ASTRONOMY 155 does not. The reality of this feature requires or changes through evolution may, of course, checking. The test might be made with the make the establishment of such a mean function aid of late-type dwarfs discovered spectro- impossible on the basis of present knowledge. scopically (Vyssotsky, 1943, and coworkers, Luminosity functions of typical Population 1946, 1952). With the present functioning II would also be of great interest. Important of Schmidt telescopes, intrinsically faint stars results for luminosity functions of globular discovered spectroscopically should play an clusters have recently been given by Sandage increasingly important role in the determina- (1954) and by Tayler (1955); for nebulae by tion of L(M; Ro, 0, 0). Baum and Schwarzschild (1955), by Baum The van Rhijn luminosity function pre- (1955), and by Roberts (1955, and in press). sumably characterizes Population I; the ex- For these latter cases indirect methods using actness of the characterization may bear integrated magnitudes and multicolor observa- further analysis. The method of deriving tions were employed. the luminosity function entails the use of Variations have been found in luminosity volume elements of different sizes for different functions of typical Population II systems just luminosity ranges. The very bright stars are as they have in Population I systems. An observed over a volume element of great size, elliptical nebula appears to contain a much a cylinder of possibly a kiloparsec radius; the larger percentage of dwarfs than a globular very faint stars are observed within a spherical cluster. The luminosity function of the dwarf volume element of a few parsecs radius centered branch of a globular cluster appears remarkably on the sun. The highly luminous stars are similar to that of the stars in the solar neigh- known to be nonrandomly distributed in the borhood; a typical galactic cluster contains a galactic plane; indications are (McCuskey, much smaller percentage of dwarfs than does 1951) that the fainter stars also vary in mixture the solar neighborhood. Comparisons of this over moderate distances from the sun. The sort based upon many more cases than are now van Rhijn luminosity function was derived when available will be of very great interest. there was still belief in the universality of the They may well provide the key to an understand- function. There is no certainty that it refers ing of the mixture of stars we find in the neigh- to a homogeneous population of stars. The borhood of the sun, particularly if information function is, for the faint stars, essentially of the space density of stars of different spectral L(M; R, 0, 0); for the bright stars it is some types and luminosity classes in and between average L(M), the average being taken over the arms can be derived. a large and somewhat uncertainly specified volume element. The local system An important step towards a clearer under- Interpretation of space density results in the standing of the Population I luminosity func- region of the sun in terms of a simple picture tion could be taken by determining the lumi- of spiral arms may prove difficult because of nosity functions of a number of representative the interference of the "local system" (see Bok, Population I objects; namely, galactic clusters, 1937, for earlier references). The observations star clouds, associations. Such luminosity by Nassau and Morgan (1951) demonstrate functions could be determined with adequate that the distribution of bright OB stars follows accuracy by statistical means, provided an the outline of the local system. Heeschen and accurate magnitude sequence was available for Lilley (1954) have found the local system in each case. Of particular interest would be a study of variations in the form of the luminosity the 21-cm radiation of hydrogen. The struc- functions for stellar groups of different Trump- ture likewise appears among both the bright and ler type or different age, different size, and so dark nebulae. A thorough modern study of the forth. The possibility of forming from such local system would be of interest; it would also a sample a mean function representative of be of value as an aid in disentangling the Population I in the solar vicinity should be local system and what one ordinarily thinks of investigated. Variations in different groups as spiral structure. 156 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Other aspects of the galaxy GURZADIAN, G. A. 1949. Astron. Journ. Acad. Sci. U.S.S.R., vol. 26, For an understanding of the galaxy a knowl- p. 329 (for abstract see Astron. News edge of the interstellar material is no less Letter No. 54, p. 1, 1951). important than that of the stars. Problems HEESCHEN, D. S., AND LILLET, A. E. of some aspects of interstellar matter have been 1954. Proc. Nat. Acad. Sci., vol. 40, p. 1095. described where necessary: the degree of as- HEES, R. 1924. Seeliger Festschrift, p. 265. Springer, sociation of interstellar hydrogen and spiral Berlin. tracers, motions of the hydrogen, and the value HUMASON, M. L., AND WAHLQUIST, H. D. of the ratio of interstellar extinction to red- 1955. Astron. Journ., vol. 60, p. 254. dening. However, general problems of the JOHNSON, H. L., AND MORGAN, W. W. interstellar medium and of the 21-cm radiation 1955. Astrophys. Journ., vol. 122, p. 142. KEENAN, P. C, AND MORGAN, W. W. of hydrogen fall outside the scope of this 1951. In Hynek, ed., Astrophysics, p. 12. Mc- review. They will be found in the reviews of Graw-Hill, New York. those subjects. Likewise, problems of motions KOTJRGANOFF, V. within the galaxy have been introduced only 1951. Quelques documents sur la structure de la where necessary since they form the subject galaxy. (Inst. d'Astrophys., Paris, June, matter of a separate paper. In actual practice, 1951. Trans, by C. Payne-Gaposchkin in Astron. News Lett., No. 64, 1952.) the study of the distribution of stellar or KWEE, K. K., MULLER, C. A., AND WESTEBHOUT, G. interstellar material and the study of kinematics 1954. Bull. Astron. Inst. Netherlands, vol. 12, cannot be separated if a complete view of the p. 211. problem is to be gained. The distribution of LUTTBN, W. J. material and the motions of the material are 1939. Publ. Astron. Obs. Univ. Minnesota, vol. 2, No. 7. component parts of one problem that must be 1941. Ann. New York Acad. Sci., vol. 42, p. 201. considered as a whole. MATALL, N. U. 1946. Astrophys. Journ., vol. 104, p. 290. References 1951. Publ. Obs. Univ. Michigan, vol. 10, p. 19. MCCUSKBT, S. W. BAADE, W. 1944. Astrophys. Journ., vol. 100, p. 137. 1951. Publ. Obs. Univ. Michigan, vol. 10, p. 67. 1954. Quoted in Bull. Astron. Inst. Netherlands, MORGAN, W. W., MEINEL, A. B., AND JOHNSON, H. M. vol. 12, p. 117; an earlier value will be 1954. Astrophys. Journ., vol. 120, p. 506. found in Pub. Obs. Univ. Michigan, vol. MORGAN, W. W., SHARPLESS, S. L., AND OSTERBROCK, 10, p. 7, 1951. D. E. 1952a. Astron. Journ., vol. 57, p. 3. BAADE, W., AND MINKOWBKI, R. 1937. Astron. Journ., vol. 86, pp. 119, 123. 1952b. Sky and Telescope, vol. 11, p. 134. NASSAU, J. J., AND MORGAN, W. W. BAUM, W. A. 1955. Astron. Journ., vol. 60, p. 153. 1951. Pub. Obs. Univ. Michigan, vol. 10, p. 43. OORT, J. H., VAN DE HULST, H. C, AND MuLLEB, C. A. BAUM, W. A., AND SCHWABZSCHILD, M. 1955. Astrophys. Journ., vol. 60, p. 247. 1952. Trans. Int. Astron. Union, vol. 8, p. 512. ROBERTS, M. S. BlDELMAN, W. P. 1943. Astrophys. Journ., vol. 98, p. 61. 1955. Publ. Astron. Soc. Pacific, vol. 67, p. 327. 1947. Astrophys. Journ., vol. 105, p. 492. . Astron. Journ., in press. BLAAUW, A. RUSSELL, J. A. 1955. Int. Astron. Union Symposium, No. 1. 1953. Astron. Journ., vol. 58, p. 89. BOK, B. J. SANDAGE, A. R. 1937. The distribution of the stars in space. As- 1954. Astron. Journ., vol. 59, p. 162. trophysical Monographs, p. 97. Uni- SHARPLESS, S. L. versity of Chicago Press. 1952. Astrophys. Journ., vol. 116, p. 251. DIVAN, L. 1954. Astrophys. Journ., vol. 119, p. 200. 1954. Thesis, Univ. Paris, 1954, and Li6ge, Colloq. STEBBINS, J., AND WHITFORD, A. E. Inter. d'Astrophysique, 1954, p. 538. 1945. Astrophys. Journ., vol. 102, p. 318. FHICKE, W. STROMGREN, B. 1949a. Astron. Nachr., vol. 277, p. 241. 1951. Astron. Journ., vol. 56, p. 142. 1949b. Astron. Nachr., vol. 278, p. 49. 1954. Astron. Journ., vol. 59, p. 193. NEW HORIZONS IN ASTRONOMY 157

TATLEB, R. J. VTSSOTSKT, A. N., AND MATEER, B. A. 1955. Astron. Journ., vol. 59, p. 413. 1952. Astrophys. Journ., vol. 116, p. 117. VAN DB HULST, H. C, MuLLEH, C. A., AND OOBT, J. H. WEAVER, H. F. 1954. Bull. Astron. Inst. Netherlands, vol. 12, 1953. Astron. Journ., vol. 58, p. 177. p. 117. 1955a. Astron. Journ., vol. 60, p. 202. VAN RHIJN, P. J. 1955b. Astron. Journ., vol. 60, p. 208. 1936. Publ. Astron. Lab. Groningen, No. 47. 1955c. Astron. Journ., vol. 60, p. 210. VTSSOTSKT, A. N. WEAVER, H. F., AND MORGAN, H. R. 1943. Astrophys. Journ., vol. 97, p. 381. 1956. Astron. Journ., vol. 61, p. 268. VTSSOTSKT, A. N., , E. M., BT AL. ZWICKT, F. 1946. Astrophys. Journ., vol. 104, p. 234. 1955. Publ. Astron. Soc. Pacific, vol. 67, p. 232.

Galactic Clusters and Associations By A. Blaauw l

Research on galactic clusters and associations common circumstances. A vast amount of can be divided into two categories: the proper- observational data has been collected in the ties of the individual objects, and the properties past for the galactic clusters, most of which will of the system of the objects as a whole. The serve as a basis of more refined work made first aim of these investigations is, of course, to possible by the improved present and future arrive at a knowledge of the present condition observational techniques. Much less has been of the individual clusters and associations and done on the associations, whose importance of their present motions and distribution in has only recently become clear. However, space. But the ultimate purpose is to under- what has been done observationally so far, stand these conditions as a result of the origin especially in the field of photometry, meets the and evolution of the clusters and associations. standards required for our present demands. This problem is immediately connected with Having in mind that research on galactic that of the history of the galaxy and with that clusters and associations can contribute to our of stellar evolution. This double aspect makes understanding of the evolution of the galaxy the study of clusters and associations partic- and of the stars, we shall consider three major ularly significant. trends in this research, in the three sections There is no sharp division between galactic following. clusters and associations. This is best illus- trated by the fact that certain objects, such as Individual properties of the stars NGC 2244, may be classified under either of Individual properties of the stars in clusters these categories. Most associations must have and associations may be: the absolute magni- short lifetimes and will have dispersed among tude, the color, the spectral type, the angular the general population of the galaxy within one momentum, the duplicity or multiplicity, and or two revolutions of the galaxy (a few hundred the motion of the star with respect to the center million years). Galactic clusters are more of the cluster or association. Such studies are stable objects but this need not mean that their important because the stars in the cluster or origin is essentially different from that of the association can be assumed to be of approxi- associations. They may simply be remnants mately equal ages, to have formed at the same of the rich central parts of some associations with location in the interstellar medium, and, very small internal motions, that had a negative perhaps, to have uniform chemical composition. energy from the outset, or have disposed of a This last assumption is uncertain since the small positive energy by the ejection of part abundance of interstellar grains may have of their stellar content. Stellar associations varied within the cloud complex in which the offer an opportunity to study the properties of stars were formed. In principle, knowledge of the very youngest stars, and they seem to be these properties of the stars in a large number one of the best places to look for information of clusters and associations would provide an on the process of itself. Galactic ideal basis for studying (1) the properties of clusters offer a variety of star samples, each of stars of equal age but of different mass and which informs us about the properties of stars angular momentum; (2) the changes, with age, formed simultaneously and at the same loca- not only of the stars' position in the HR - tion in the interstellar medium, under certain gram but also of their rotational and duplicity properties. »Yerk«s Observatory, Williams Bay, Wls. 159 381014—56 -13 160 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 Trumpler's extensive investigations of galac- the Hyades, , and a Persei clusters. tic clusters have revealed the large variety of The use of differential galactic rotation as an patterns in the HR diagrams and this, combined indicator of distance has not produced valuable with modern concepts of stellar evolution, distances in the past, but it may do so in the shows that these objects represent a wide range future for the young clusters if these follow the of evolutionary stages. The modern equivalent spiral structure of interstellar hydrogen closely of these diagrams, based on accurate photo- and if radial velocities determine uniquely to electrically measured colors and magnitudes, which arm the clusters belong. H. L. Johnson's has already allowed more detailed discrimina- approach to the problem of the absolute magni- tion of the properties of different stars in the tude determination, based on the assumption of same cluster, and more refined comparison of identical "zero-age main sequences," may prove the color magnitude arrays of different clusters. to be very useful provided sufficient evidence It has, moreover, the great advantage of reach- can be produced to show that the chemical ing considerably fainter apparent—and, hence, composition does not affect the position of the also absolute—magnitudes than the spectro- main sequence. One must bear in mind, how- scopic method. ever, that this approach eliminates the detec- While color and absolute magnitude are the tion of composition differences unless these are most obvious quantities to choose in defining of a more pronounced character in the rapidly the physical properties of a star, they do not evolving part of the color absolute-magnitude describe them completely. Initial chemical com- diagram than in the slowly evolving part; the position and angular momentum cause vari- determination of the absolute magnitude scale ations of the array in the color absolute-magni- should be based, preferably, on the latter. tude diagram for clusters of equal age. We The extension of the color-magnitude arrays should consider also the duplicity or multiplicity to faint stars may require the identification of of the stars and their motions in the cluster. fainter cluster members than are known at These will probably have no relation to the present. This holds even for the nearest color-magnitude arrays, but they must have an clusters, the Hyades being a good example of important bearing on the process of the forma- a cluster whose faint members are still unknown. tion of the cluster from the interstellar medium It will in general be the case for the poorer and will have to be taken into account in subse- clusters observed against the rich background quent studies of the evolution of the clusters. of the Milky Way. So far, observations of There may be other properties of the stars in ad- the faint stars in clusters have usually been dition to those listed above, which, though not confined to a sample considered representative yet easily accessible, may provide important as far as the magnitudes and colors are con- clues; the magnetic fields might be among these. cerned, with little attention being given to the There can be little doubt that the first desider- estimation of the total numbers of faint mem- atum for future work on galactic clusters is the bers. Yet, this is an important item if we determination of accurate color absolute-mag- want to consider the density of the population nitude arrays for a large number of clusters. along the main sequence at zero age for differ- These should provide the basis for subsequent ent clusters. That differences in this respect studies of the secondary effects mentioned exist follows from the comparison of NGC above, which may be feasible only for a limited 752 and Praesepe. Such differences are prob- number of stars and a limited number of clus- ably related to the mass density and the state ters. These latter could then be selected so as of motion in the interstellar medium from which to represent the most useful variety of evolu- the cluster originated. Their observation may tionary stages. provide interesting data on the formation A critical step in the determination of the process of the clusters. color absolute-magnitude arrays is the conver- Existing data on the identification of faint sion of apparent magnitudes into absolute mag- cluster members can in many cases be con- nitudes. There are only a few clusters for which siderably increased by the measurement of new we have geometrically determined distances: proper motions in the fields of the clusters. NEW HORIZONS IN ASTRONOMY 161 The large collections of photographs of galactic very little, while the internal velocities are clusters, taken at several observatories 30 or generally large and very likely much the same more years ago, are becoming more and more as at the time of formation of the stars; there- useful for this purpose. In many cases, work fore, they can indicate the initial properties of this kind done in the years preceding World of the stars and may yield clues to our under- War II cannot be considered to be up to date standing of their process of formation. The for the future projects mentioned above. observations made so far have been concerned Comparison of the old epoch plates with modern mainly with the photometry (absolute mag- ones may give proper motions of several to nitudes and colors), the proper motions, and 10 times the weight of those available at present. radial velocities, while a beginning has been It is gratifying to know that several observa- made with the systematic investigation of the tories that possess fine collections of old plates duplicity properties. But all these referred have already started comparing these with only to the brightest part of the population and modern ones, whereas others have expressed the stars of spectral types about B5 and earlier. their interest in undertaking programs of this Very little is known as yet of the occurrence kind. International Astronomical Union Com- of fainter, later-type stars. Stars down to mission 37 on star clusters and associations at spectral type A0 along the main sequence the Dublin Assembly has adopted a resolution have been surveyed in the Orion association, in to the effect that observatories which possess III Cep, and in the most concentrated parts collections of early epoch cluster plates present of a few other objects. But in two of the lists of these to the president of Commission nearest associations, II (f) Perseus and I 37 and indicate to what extent they are in a , we are still ignorant about the mem- position to collaborate on work. bers of types B5 and later. Yet these two Observations of the stellar rotation velocities associations are favorably located at relatively and of the frequency of spectroscopic and high latitudes where the confusion by the visual binaries have, in the past, been given general galactic background is less serious than relatively little attention. Yet they would seem in most other cases. Surveys of the late B and to offer most interesting prospects for future later-type stars in associations will suffer from research. Why are there few or no spectro- a few drawbacks: the area over which the stars scopic binaries among the Pleiades although have to be classified is much larger than for visual doubles are plentiful, whereas spectro- the clusters—for the nearest associations of scopic binaries are frequent in other clusters, the order of 100 square degrees—and, since notably in those containing many early-type the internal velocities are large, members and stars? Are we facing here an evolutionary field stars cannot be distinguished uniquely. effect or is this difference directly related to The membership may have to be based on different initial conditions in the interstellar statistical evidence, like that, for instance, medium at the time the clusters were born? showing that the concentration in the sky for Whatever the answer may be, it is obvious a particular type of star is greater in the region that a systematic study of the frequency of of the association than outside it. The size rotational velocities and of the duplicity and direction of the proper motions may help characteristics in clusters of a variety of ages when the expanding character is pronounced, and appearances will ultimately give valuable but then it can be used only in the outer regions information on the formation problem. where the density of the population is small. The foregoing remarks referred in the first In spite of these difficulties, knowledge of the place to the work on galactic clusters. Although fainter membership of the associations must the basic problems are very much the same for be considered one of the most urgent goals of the associations, the emphasis in future work research on these objects. The first approach will probably be somewhat different. The should, of course, be by Schmidt telescopes or associations on the whole are much younger large-field astrographs equipped with objective than the clusters so that the stars have evolved prisms. 162 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Dynamical studies the associations in order to attain high accuracy Dynamical studies of clusters have, in the past, for the individual stars. In addition to the been mainly concerned with systems in a proper motion work we must consider radial state of approximate equilibrium, and have velocity observations as one of the principal dealt with the relation between internal velocity desiderata for future work. This has been distribution and mass distribution, the times mentioned also hi connection with the detection of relaxation, and the rate of escape of stars of spectroscopic binaries. from the cluster. Investigators usually as- The system in the galaxy sumed that the clusters are approximately the , several billion years. Investigations of the space distribution and the Recently acquired knowledge of the ages of kinematics! properties of clusters have shown clusters derived from increasing information that they occupy a flat subsystem in the galaxy with peculiar motions of the order of 10 km/sec. on stellar evolution offers the possibility for Only in recent years have the more accurate a fresh approach to dynamical studies of photometric distance determinations revealed clusters. Various questions come to our mind. the connection between open clusters and the We notice that the distribution of stars in the spiral structure. Hyades is quite irregular. We estimate its For future work in this field, also, the ac- age at about 109 years. Can we reconcile cumulation of a large number of accurate color- these two observations, estimate how many magnitude diagrams with the inferred knowl- stars must have escaped from the Hyades, and edge of the ages of many clusters forms the predict its future constitution? Suppose we basic requirement. Once this information has indeed collect data on the ages of a large num- been obtained one can study the distributional ber of clusters as suggested in the preceding and kinematical properties for different age section, can we explain the numbers of clusters groups. For the youngest clusters the evidence of various ages as a result of the predicted rate already obtained for the distribution along the of disintegration ? spiral arms will be strengthened, and it will be The dynamics of associations presents prob- interesting to see where, along the spiral struc- lems of a somewhat different kind. Especially ture of interstellar gas, the formation of clusters in the youngest objects, the forces acting on the used to take place. Data on proper motions, stars are probably mainly due to the interstellar already referred to, and mean radial velocities matter present in the region of the association. will provide information on the velocities of If we make the reasonable assumption that these objects at the time of formation. only a small fraction of the interstellar matter The most intriguing of these studies may, has condensed into stars, its mass may very however, well be those dealing with the some- likely exceed the total mass of the stars many what older objects. Shall we be able to recon- times. Probably the most promising aspect struct from their present distribution and of the study of the motions in associations is motions their distribution at the time of forma- that it will give us direct information on the tion ? This may be possible for the slow moving state of motion of the generating interstellar clusters whose ages are now of the order of one cloud complex, and it is intimately connected revolution of the galaxy (200 million years). with the complicated problems of the dynamics If we assume their space velocities to be of the of cosmic clouds. It is clear, however, that we order of 5 km/sec, they have traveled during must obtain the most precise data on the their lifetime a distance of about 1,000 parsecs motions of individual stars. For the brightest from their origin. This is probably not too far ones (m

Globular Clusters Observed through a Crystal Ball By W. A. Baum1

At the focus of a large telescope, a globular volume decreases as the inverse third or fourth cluster is one of the most magnificent sights in power of the distance from the nucleus, and the sky. One can see thousands upon thou- there is no well-defined outer boundary to their sands of stars all crammed together within a domain. It would be very valuable to have a few minutes of arc. Embedded like jewels here more complete and accurate analysis of the and there in the sparkling swarm, the brighter distances and velocities of all known globular stars run the full gamut of color; the brightest clusters in the galaxy, but the task is a big one. are brick red while others, slightly less bright, The stars in globular clusters are an entirely are distinctly blue. Toward the center of the different breed from those with which we are cluster, the population becomes fantastically familiar in the solar neighborhood. They have dense. It is amusing to consider what our sky different chemical abundances and they con- would look like if the earth were associated sequently follow different patterns of evolution with a star in the nucleus of a globular cluster. as they age. These differences can be deduced Perhaps the best way to judge what the from spectra and from color-magnitude dia- future has in store will be to examine the grams. Stars of the globular-cluster kind are present state of our knowledge and some sometimes referred to as Population II. They present lines of speculation to see the gaps are found not only in the clusters but also as which need to be filled. unattached vagabonds wandering through the Globular clusters constitute a very well de- same regions of the galaxy (nucleus and halo) fined class of objects, all pretty much alike in frequented by the clusters. They are found their general features. You could hardly similarly in the halo of M 31 and in nearby mistake one for any other type of system. Out dwarf galaxies. In general they seem to favor of more than a hundred observable globular regions which are free of dust. Since it does clusters in our own galaxy, only two or three are not necessarily follow, however, that stars still debatable cases. inhabiting a dust-free region are always the The distribution of globular clusters in the same kind (i. e., same chemical abundances and galaxy is almost spherically symmetric about associated patterns of evolution), one cannot the galactic nucleus, whereas stars like those say with certainty whether stars of the globular- in the solar neighborhood, along with gas and cluster kind do or do not constitute the pre- dust, are restricted to a highly flattened disk- dominant population in elliptical galaxies; if shaped region. This difference in distribution they do, then they are probably the most is associated, as one might expect, with a numerous kind of stars in the universe. This difference in angular momentum; globular is a fundamentally important question on clusters probably do not as a group share to any which more work is needed. large degree in the rotation of the galactic disk, but they have large individual radial velocities. Color-magnitude diagrams Except in the immediate region of the galactic One of the best ways to characterize the stars nucleus, the number of globular clusters per unit that compose a stellar system is to plot their 1 Mount Wilson and Palomar Observatories, Pasadena, Calif. color indices (or their spectral types) against 165 166 SMITHSONIAN CONTHIBUTIONS TO ASTROPHYSICS

i i i 1 1 i i i i i i i i

-6 —

My —

— -4 \ \ \ \ -2 ^^^^ — \ \ MI3 \ 0 \^ A M3 A MI3/IM3 f \ +2 I \ J/ 11 — \ f / J IMI3 \ J

_ \— •4 \ \w \ X M3 \ N +6 — X POP. i _ \ +8 — MI3 ^N 1 1 1 i i i i 1 I I N I I -.4 +.4 +JB -t-1.2 +1.6

FIGURE 1.—Schematic color-magnitude diagrams for M 3 and M 13. Sequences for the solar neighborhood are sketched with dashed lines for comparison. The diagrams have been adjusted vertically so that the mean of the cluster-type variables (circle) in M 3 falls at M,=0 and so that the evolutionary breaks from the main sequence fall at the same bolometric magnitude (dotted line). NEW HORIZONS IN ASTRONOMY 167 their magnitudes. In contrast to the diversity diagram to reach its present configuration is of color-magnitude diagrams found among uncertain, but it apparently is of the same open clusters which inhabit the galactic disk, general order as the age of the universe judged the color-magnitude diagrams of globular from the recession of galaxies, say, 5 to 10 clusters all have basically similar features. billion years. Each star probably spends more This does not mean that they are identical. than half of its life on or near the main sequence, There are indeed important differences among where it uses up only 10 to 15 percent of its them, but the differences mainly pertain to the hydrogen. As it proceeds up the vertical positions of the sequences rather than to the branch, its rate of rise is roughly proportional existence or nonexistence of various sequences. to the luminosity, and its radius expands as Color-magnitude diagrams for two well- the square root of the luminosity. By the known globular clusters, M 3 and M 13, are time it reaches M,~0, the star is evolving quite sketched in figure 1. The color indices span fast. It requires only about 2X108 years to an extremely broad range; the bluest stars have go the rest of the way to the top of the diagram indices around —0.5 on the PQ— V system while and about an equal length of time to traverse the reddest are around +1.5. The brightest the . The nova stage, if stars in globular clusters have absolute photo- indeed that is the star's actual route to the visual magnitudes of about —3, while the graveyard, must be less than 109 years' dura- faintest are presumably much below the present tion. Beyond that, its evolutionary schedule limit of observation at the bottoms of the is unknown. diagrams shown here. Near the top of the red-giant branch, the In the lower part of each diagram, from star starts to fluctuate small fractions of a M,~+4 downward, is a segment of the main magnitude at a relatively slow rate, its ampli- sequence. Stars which formerly lay a little tude increasing as the tip of the branch is above Mf~+4 on the main sequence have approached. In the same region of the color- recently evolved away from it and they magnitude diagram one occasionally finds a presently form the upper branches of the bonafide long-period variable, which suggests diagram. Stars which started from somewhat that the fluctuation may build up briefly to brighter levels of the main sequence, say large amplitudes as the hydrogen-spent star above M,~+3, have long since completed starts to collapse and slide back down the red- their evolutionary sequences and passed on giant branch. In any case, it apparently to the graveyard. When a star evolves away quiets down again as it travels rapidly toward from the main sequence it evidently moves the left along the horizontal branch. Then into the subgiant region at the right of the suddenly it breaks into another fit of oscilla- main sequence and climbs upward into the tion, this time as a cluster-type variable. Its red-giant branch. Since the total ascent is period and light curve are relatively steady, about seven magnitudes, the star is consuming but they are believed to undergo gradual fuel several hundred times faster during its secular changes as the star pursues its evolu- red-giant phase than during its childhood on tionary course through this region of the the main sequence. As a result it doesn't live diagram. Present data suggest that the period, very long as a red giant. When it starts to starting initially at about one day, becomes run short of hydrogen, the star is believed to progressively shorter, while the amplitude move toward the left along the horizontal rises to a maximum and then falls again. This branch and then dribble downward at the fall in amplitude is accompanied by a change extreme left of the diagram to the graveyard. in the shape of the light curve which may be Its death throes are suspected of being occa- associated with a change in the mode of pulsa- sional nova outbursts, after which it collapses tion. At the blue edge of the variable-star into a degenerate and cools itself region the star quiets down again until it slides into oblivion. off the blue end of the horizontal branch. If The time required for the color-magnitude the star then passes through a nova stage, one 168 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS VOL.1 might count it as the third fit of instability age alone would hardly seem sufficient to along the evolutionary track, but it evidently account for it. differs in kind from the other two. There are some fascinating differences be- The foregoing paragraphs outline the general tween the color-magnitude diagrams of various features of globular-cluster diagrams and their globular clusters. Features which differ notice- evolutionary interpretation as we understand ably from cluster to cluster include: (1) the them today. While certain basic facts now position of the main sequence, (2) the position seem to be fairly definite, it should be empha- (colorwise) of the top of the red-giant branch, sized that much of this information is only ten- (3) the shape and tilt of the red-giant branch, tative and some of it is even speculative. (4) the number of cluster-type variables, (5) Moreover, there are a few items which were the populousness of the horizontal branch at omitted altogether, because we haven't the the right of the cluster-type variables, and (6) foggiest notion how to fit them into the over-all the shape and tilt of the horizontal branch. evolutionary picture. These include the long- The last four of these features seem to be period Cepheids (between M,^ — 1 and mutually related; if one intercompares all of the M,~*—5) found in some of the clusters, the color-magnitude diagrams which have been in M 15, and the occasional photoelectrically calibrated on the same color blue stars found above the horizontal branch in system, he finds that they can be arranged into a few clusters. a series (call it series A) in which these last Altogether, there are so many points on four features tend to change mutually and which more photometric material is needed that progressively from one diagram to the next. it is difficult to suggest what should be under- The number of cluster-type variables thereby taken first. In some cases, the question is becomes a handy index for categorizing globular largely one of observing more stars or sampling clusters and their diagrams. It is rather puz- more clusters. This is true, for example, if one zling to note, however, that the second feature wishes to understand the nature of red-giant in the list above suggests arranging the dia- fluctuation, the properties of the red long-period grams into a different series (call it series B), variables, or the role of the long-period Cepheids; and that it is this latter arrangement which it is also true of investigating the differences spectroscopic observations tend to support. between cluster diagrams, which we shall dis- Data bearing on the first feature in the list cuss shortly. In other cases, the problem is above are still too scanty to tell us which rather to improve the accuracy of existing data. series, if either, the main sequences will favor, It has sometimes been tempting to overrate the but a prediction can be ventured: if series A finality of the present material and to neglect truly represents the effects of chemical abun- the need for further investigation. dances upon evolution, the arrangement of the Among the various branches of the cluster main sequences ought to be in accord with it. diagrams, the main sequence is observationally Actually, series B has somewhat the appear- the most difficult. Recent developments in ance of being series A folded in the middle, but photoelectric photometry, however, have made the number of clusters investigated is too it possible to measure extremely faint stars with small to provide more than a hint of any such complete freedom from the scale errors that relationship. If one intercompares a group of have hampered photographic efforts. As a cluster diagrams arranged in order according result of such developments, it has been found to series A, the tops of the red-giant branches that the stars comprising the main sequence of tend to be progressively bluer until one comes a globular cluster do not fall on the ordinary to M 92, after which the red-giant branches main sequence with which we are familiar. are progressively redder. Both extremes of They are subdwarfs. This result is of cos- series A might thus be associated with the red mogonic significance because it leaves little extreme of series B, and we could then forget doubt that globular clusters are made from a about series B having any independent sig- different recipe from that of the majority of nificance. This happy unification seems to stars in the solar neighborhood; a disparity in run into grief, however, if we attempt to NEW HORIZONS IN ASTRONOMY 169 interpret abundances from the spectra and to about the same limit that can be reached simultaneously to invoke abundances to account photographically, most of these clusters can for series A. If abundance differences are be reached with telescopes of modest aperture if reserved for explaining series A, and if B cannot they are adequately equipped. The approxi- be associated with it, then perhaps we should mate number of known variables, the integrated seek a phenomenon which does not affect the spectral class (or alternatively the integrated general evolutionary pattern but which does color index), and the approximate distance from affect the swollen atmospheres associated with the galactic nucleus are all parameters by the stars during their red-giant phases. The which clusters can be grouped and compared. fact that all the giants of a cluster are affected We have already cited the number of variables to somewhat the same degree suggests the as an index to the form of the color-magnitude possibility of an environmental phenomenon, diagram according to series A. We shall return such as surface accretion during passage of the in a moment to a discussion of the relationship cluster through the galactic plane or nucleus, of these integrated properties to the distance but such environmental effects have not been of the cluster from the galactic nucleus. regarded with much favor. A third possibility Among the clusters in table 1, color-magni- for explaining the existence of the two series is tude diagrams suitable for intercomparing based upon the fact that abundance differences the upper branches have thus far been pub- themselves are not necessarily one-dimensional; lished for NGC 4147, M 3, M 5, M 13, M 92, in principle there are as many parameters as M 2, M 10, and M 15. Work is now in progress there are elements in the periodic table. We on one or two others, but the majority haven't can imagine, for example, that some of the been touched. It should also be mentioned features of the color-magnitude diagram (maybe that there are several clusters which are too those abiding by series A) might be sensitive interesting to neglect in spite of the fact that to the hydrogen/helium ratio, while others they lie at less favorable galactic latitudes (maybe those abiding by series B) might be than those listed in table 1. Outstanding sensitive to the hydrogen/metals ratio. In any examples with low moduli are w Centauri, M 4, case, the problem of interpreting these series is M 22, and NGC 6397. Clusters lying close to a very challenging one. the galactic nucleus include NGC 6304, NGC Photometric observations of high precision 6316, NGC 6355, NGC 6453, and NGC 6522. are badly needed in more globular clusters if There are also two or three massive galactic the differences between them are to be fully clusters of great age, similar perhaps to M 67, understood. Although many color-magnitude whose color-magnitude diagrams may help to diagrams of various sorts exist, only a few have clarify evolutionary differences between halo been calibrated photoelectrically and adjusted and disk populations. accurately to a standard color system. There It is much more difficult to reach significant are plenty of favorably situated clusters whose samples of the main sequences than similarly diagrams would be valuable. Table 1 is a list to reach the upper branches. It requires a of clusters which lie 20° or more from the large telescope and a great deal of photoelectric galactic plane. Twenty-four of them are in observing time under first-rate conditions. the Southern Hemisphere while only 13 are in Thus far, M 13 is the only cluster for which the the Northern Hemisphere; however, several main sequence has been firmly located. Some of the southern clusters can be reached fairly tentative photoelectric results were used for well by northern telescopes. The modulus including M 3 in figure 1, and similar material represents a rough estimate of the magnitude also now exists for M 15. Earlier appearances of the horizontal branch; for the upper branches of normal main sequences in the photographic- to be well defined, the color-magnitude diagram ally extrapolated diagrams for M 3 and M 92 should extend about two magnitudes fainter have turned out to be spurious. It may be than the modulus without serious deterioration several years before a definitive amount of of accuracy. Since integrating techniques now data on the relationships of globular-cluster enable one to obtain photoelectric observations main sequences can be assembled. 170 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS

TABLE 1.—Globular clusters for which |6|>20c

Estimated modulus Distance from Clutter designation a 1950 t 1950 (M) Number of variables Integrated spectrum gal. nucleus mag kpc NGC 104 47 Tuc 00 2l!°9 -72 21 14. 5 8 8 288 00 50.2 -26 52 15. 7 2 15 362 01 00.6 -71 07 15. 7 14 12 1261 03 10.9 -55 25 1841 04 52. 5 -84 05 1851 05 12. 4 -40 05 3 F5 1904 M 79 05 22.2 -24 34 16. 3 5 F3 20 2419 07 34.8 +39 00 19.4 36 F5 65 4147 12 07. 6 +18 49 17.0 4 A5 24 4590 M 68 12 36.8 -26 29 16. 1 28 A6 13 5024 M 53 13 10.5 +18 26 16.7 42 A8 19 5053 13 13. & +17 57 16. 4 10 17 5272 M 3 13 39.9 +28 38 15.7 186 F2 13 5466 14 03. 2 +28 46 16. 4 18 17 5634 14 27.0 -05 45 17. 1 7 F4 19 5694 14 36.7 -26 19 17.7 0 A9 22 5824 15 00.9 -32 53 F5 5897 15 14. 5 -20 50 16.2 8 5904 M 5 15 16.0 +02 16 15.2 97 F7 7 6171 M107 16 29.7 -12 57 16.2 24 G2 6 6205 M 13 16 39. 9 +36 33 14. 8 15 F2 9 6218 M 12 16 44 6 -01 52 15.2 1 F7 4 6229 16 45.6 +47 37 17. 7 21 F6 28 6254 M 10 16 54.5 -04 02 15.2 2 GO 4 6341 M 92 17 15.6 +43 12 15.2 16 A5 10 6684 18 44. 1 -65 14 6752 19 06.4 -60 04 14.2 1 5 6809 M 55 19 36.9 -31 03 14. 5 2 4 6864 M 75 20 03.2 -22 04 18. 1 11 Gl 25 6934 20 31. 7 +07 14 16.9 51 F9 15 6981 M 72 20 50.7 -12 44 17.0 31 G2 16 7006 20 59. 1 +16 00 18.8 20 Fl 39 7078 M 15 21 27.6 +11 57 15. 9 66 FO 12 7089 M 2 21 30.9 -01 03 16. 2 17 FO 13 7099 M 30 2137.5 -23 25 15.5 3 A7 9 NGC7492 23 05.7 -15 54 17.9 9 32 IC4499 14 52.7 -82 02

The foregoing discussion of color-magnitude for the differences between their color-magni- diagrams is presented in terms of ordinary tude diagrams. two-color photometry. Much can be added, Multicolor photometry and spectrophotom- however, by extending observations to three or etry have not yet been applied to individual more colors. In particular, the addition of stars in globular clusters, and the results of ultraviolet enables one to compare the relative any multicolor exploration are likely to pay strengths of the Balmer regions in various handsome rewards. Although the useful range stars. This can be done for instance by of a photometer becomes limited to brighter plotting a diagram of U-B against B-V. objects when the color bands are made narrower, Results in M 13 show that the red giants and the bright stars in globular clusters are still have an ultraviolet excess of about within easy reach of conventional six-color 0.05 magnitude, while the very blue stars on observations with a telescope of modest aper- the horizontal branch have very pronounced ture. On the other hand, narrow-band spec- ultraviolet deficiencies. Rather oddly, there trophotometry will be quite difficult. is a lonesome B 2 star lying about two magni- tudes above the horizontal branch in M 13 Spectra which does not exhibit this ultraviolet defi- Spectroscopically, the individual stars in glob- ciency. In M3 and NGC 4147, the ultraviolet ular clusters are rather difficult objects; hardly excess of the red giants and subgiants is several any of them are brighter than 14th magnitude times greater than in M 13. Similar informa- in the blue region. Until recently, the spectra tion for other clusters would be very valuable; have consequently been of quite low dispersion. it might provide an important clue to the reason These have been suitable, however, for spectral NEW HORIZONS IN ASTRONOMY 171 classification and for detection of the grosser inferred, as it should be, from the ratios of line peculiarities of the cluster stars. It was already strengths instead of from the absolute strengths recognized a decade ago, for example, that themselves. cyanogen absorption is abnormally weak and A few particular ratios deserve mention. As that the hydrogen lines are a little more en- the excitation temperature increases, the ratio hanced than in normal stars of the same of X4254 to X4250 should decrease, whereas the absolute magnitude. ratio of X4247 to X4250 should increase. The During the last few 3rears, the fast new feature at 4254 A is a resonance line of neutral coud6 cameras at Palomar have made it pos- (0.0 volt excitation potential), the sible to reach some of the brighter cluster stars feature at 4247 A is a blend of scandium II with with a dispersion of 38 A/mm and, in two a high excitation line of neutral iron (3.3 cases, of 18 A/mm. Because of the very large volts), and the comparison line at4250A arises collimator-to-camera ratios for these disper- principally from a low level of neutral iron sions, the slit of the Palomar spectrograph can (1.6 volts). Another useful ratio which de- be made wide enough to receive nearly all of a creases with increasing temperature is Cr I star image during good "seeing" without causing 4275 to Fe I 4271, and another ratio which a loss of photographic resolution in the spec- increases with increasing temperature is Ti II trum. The first four such spectra at 38 A/mm 4394 to Fe I 4405. were obtained in 1952 with the idea of check- M 92 is apparently an extreme case of the ing spectroscopically on the photometrically peculiarities cited above, although M 15 runs observed difference between the red-giant it a close second. In some of the other clusters, branches of M3 and M92, which at that time such as M 3 and M 13, these effects are relatively were the only two clusters whose color-mag- mild. Within any one cluster, the spectra all nitude diagrams could be intercompared on a show roughly the same degree of peculiarity. standard system. These spectra not only con- In general, this degree of peculiarity appears to firmed the reality of an intrinsic difference be correlated with the redness of the red-giant between these two clusters but also showed branch (i. e., with series B) in the sense that quite dramatically that the red giants in M92 those clusters whose red-giant branches lie are very extraordinary objects. They are the farthest toward the right in the color- markedly different from any other known stars. magnitude plane are the least peculiar spectro- Additional spectra of the same dispersion were scopically. soon obtained in M92 and, subsequently, in All of the spectroscopic evidence suggests most of the globular clusters for which inter- that the weakness of the metallic lines in some comparable color-magnitude diagrams now of the cluster spectra is a bona fide abundance exist as well as in several others not yet photo- phenomenon. Quantitative abundance esti- metrically investigated. mates have been made on the basis of spectra The most striking peculiarity of the red- obtained at 18 A/nun in M 92 and M 13, which giant spectra in some of the clusters is the represent opposite extremes. Equivalent extreme weakness of the metallic lines. Super- widths of suitable metal lines in an M 92 ficially, these spectra appear to be of very spectrum are found to be about 60 percent of much earlier types than the color indices of the those in an M 13 spectrum, implying that the stars suggest. In M 92, for instance, the metals in M 92 are at least a factor 10 less metallic lines seem to have roughly the strengths abundant than in M 13. Line ratios in these usually found in F-type stars, whereas the color same spectra assure us that the excitation indices would normally be associated with temperatures in M 92 and M 13 are about early K stars. This disparity largely vanishes, the same in spite of the difference in the color however, if the spectral classification is based indices; perhaps the color difference is due to upon the ratios of certain metallic lines instead the fact that metallic lines chop different of^upon their absolute strengths. In other amounts out of otherwise similar continua. words, the excitation temperature is much less As pointed out earlier, however, the use of discordant with the color temperature if it is abundances to account for spectroscopic differ- 172 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS ences (and therefore series B) poses a dilemma, simply by comparing its color index with an because we also feel obliged to call upon unreddened mean. abundances to explain the various evolutionary The integrated spectral-energy distribution of features which behave according to series A. a stellar system such as a globular cluster is, of Possible ways around this dilemma, such as course, quite different from that of a single star allowing more than just the hydrogen/metals having the same color index. This means that, ratio to vary, have already been mentioned. regardless of data-reduction procedures, the ordinary color index of a cluster is rather sensi- Integrated properties tive to the actual color response of the photo- In addition to the classification of color- meter used. magnitude diagrams, there are several other The spectral energy distribution of a globular informative parameters which pertain to each cluster is also very different from that of an cluster as a system or as a unit. These include , indicating that the color- integrated luminosity or magnitude, integrated magnitude diagrams must be fundamentally color index, integrated spectral-energy distri- unlike unless the distribution of stars in the un- bution, integrated spectral class, mass, mass- evolved region of the cluster diagram is radically luminosity ratio, distance, reddening, position different from that for an elliptical. If the dia- in the galaxy, motion, spatial profile, and so on. grams are assumed to be of the same form, one Since the integrated luminosity and color can show that the ratio of the number of K stars represent the summed effects of all of the to the number of F stars on the main sequence members of a cluster, one should in principle of an elliptical galaxy would need to be literally be able to subtract the contributions of the hundreds of times greater than on the main se- resolved stars from the total observed lumi- quence of a globular cluster in order to account nosity, and thereby deduce how much light is for the observed disparity in their integrated contributed by members which are too faint spectral-energy distributions. Since the present to be individually resolved by present telescopes, data can be fitted equally well by assuming un- say, fainter than mPf=23. In practice, how- like diagrams, the identification of the pre- ever, uncertainties in the luminosity function dominant population in ellipticals remains an (i. e., in the numbers of stars per magnitude open question. The importance of further work interval) are much too large and the quantity on this problem has already been stressed. sought is much too small for such a procedure For much the same reason that the color index to succeed. Direct measurements of this is sensitive to the wavelength bands used for the unresolved background by itself appear to show observations, the estimation of the integrated some indications of a substantial white-dwarf spectral class of a stellar system such as a glob- population, but the uncertainties are presently ular cluster is dependent upon the spectral re- too large to justify much confidence. gion utilized. The blue stars on the horizontal As one might expect from the differences branch play a bigger role at short wavelengths between color-magnitude diagrams, the frac- than at long wavelengths. Integrated spectra tions of the total light contributed by the of 50 clusters have been obtained by various branches differ somewhat from cluster the clusters sufficiently far out of focus to smear to cluster. Typically, about half of the visual a blend of many stars uniformly over the spec- light comes from the red giants above Af,=0, trograph slit. The spectral types listed in table perhaps 20 percent each from the horizontal 1 were derived from integrated spectra obtained branch and from the subgiants, and the in this manner. Some of these spectra are cur- remainder from the main sequence downward. rently being reexamined in an effort to take In view of the broad range of color indices in better account of the peculiarities now recog- each cluster diagram, it is not surprising that nized; the present tendency is to assign them the integrated color indices of globular clusters later spectral types than were formerly esti- are not identical. One cannot correctly deter- mated. mine the reddening of an individual cluster An attempt has been made to determine the NEW HORIZONS EN ASTRONOMY 173 mass of M 92 from the dispersion in radial ve- absorption such as comparison of U, B, V data locities among red giants in its nucleus. This with the solar neighborhood or matching of was done by applying the to colors of the cluster-type variables are also open velocities measured on 23 Palomar spectro- to question. Consequently, the existence of grams of 15 stars. Some of the red giants for accurate photometric data would not by itself which two spectrograms were obtained show guarantee a precise distance for a cluster even evidence of intrinsic velocity variations, which if the absolute magnitude of the cluster-type is not surprising in view of the photometric variables were exactly known. variations already described. As a result of The fact that globular clusters differ signifi- these variations, the uncertainty in the true cantly from one another in their diagrams and velocity dispersion is disconcertingly large. in their over-all characteristics suggests that Nevertheless, with some recent refinements in they probably were not all formed under iden- theory the results yield an estimated mass of tical circumstances. It is therefore of interest 1.4 X 10s solar units, the upper limit being per- to consider what clues we have as to their times haps twice that value. The corresponding and places of origin during early epochs of the mass-luminosity ratio is of the order of unity, galaxy. When the integrated characteristics which is several times smaller than the average of globular clusters are examined in terms of for our own galaxy as a whole and is about a their positions in the galaxy, some fascinating hundred times smaller than that of an elliptical correlations stand out immediately. In par- galaxy. The interpretation of these compara- ticular, the number of cluster-type variables, tive ratios in terms of differing populations, which is an index for classification according to luminosity functions, and nonluminous content4' series A, appears to have a marked correlation are not yet clear. with the distance from the galactic plane. If The distances of globular clusters are con- we examine all clusters which have been ventionally estimated by assigning the cluster- searched for variables, which are not badly type variables a mean absolute magnitude of obscured, and whose distances are fairly well zero, or by comparing the brightest stars with known, we find that the variable-poor clusters those belonging to other clusters whose variables (say, those with 0 to 3 variables) favor regions have been calibrated. It should be remembered much closer to the galactic plane than the that the assignment of zero absolute magnitude others, hinting that they may have been formed to the cluster-type variables is only a rough ap- during a slightly later stage in the dynamic proximation which should not be taken too relaxation of the galaxy. The sample of clus- seriously when color-magnitude diagrams are ters in table 1 is not well suited for illustrating intercompared. In figure 1, for example, the this effect because of its cutoff at 20° latitude, cluster-type variables are a half magnitude apart but the effect can still be seen; the mean dis- because the diagrams were adjusted to place the tance from the galactic nucleus to the variable- evolutionary breaks from the main sequences at poor clusters listed in the table is 9 kpc, while the same bolometric magnitude. the mean distance for the variable-rich clusters Since most globular clusters inhabit regions is 19 kpc. well outside the absorbing material in the It follows, of course, that the various char- galactic disk, the photometric moduli (blue) acteristics which are correlated with the number should, on the average, be reduced by 0.25 of variables tend to show this same dependence esc b in order to estimate true distances. While on galactic distribution. For example, the in- a statistical correction of this sort is useful tegrated spectral type, which is influenced by when one is dealing with large-scale problems the relative shape and populousness of the involving many clusters, it does not provide a upper branches of the cluster diagram, reveals reliable estimate of absorption for individual an effect of this kind. Among the clusters clusters due to the probable patchiness of the listed in table 1, those possessing later spectral absorbing blanket in which we are imbedded. types (say, F 7 to G 2) have a mean distance Unfortunately, other criteria of reddening and of 11 kpc from the nucleus, while those pos- 174 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS sessing earlier types (say, A 5 to F 6) have a they travel in approximately circular orbits at mean distance of 21 kpc from the nucleus. constant radii from the nucleus. Although Again, the effect would be more distinct if arguments have been advanced for both models, clusters in the 10° to 20° zone were added to the present radial velocity data do not make a the sample. clear case for either. Probably the truth is If the centroid of the system of globular somewhere in between, and perhaps the situa- clusters in our galaxy is assumed to coincide tion is further complicated by the form of the with the galactic nucleus, one can check the potential function of the galaxy. In any zero point of the cluster-type variables against case, the time required for a cluster to make the present estimate of the sun's distance from one loop or one oscillation is of the same the nucleus (about 8 kpc). The variable-rich general order as the period of galactic rotation; clusters seem to be quite well centered on the hence, each cluster has passed through the nucleus if the conventional value of M=0 galactic plane quite a few times since its birth. is adopted for the variables, but the variable- The spatial distribution of the stars in a poor clusters appear to prefer a slightly brighter globular cluster has been the subject of several value, perhaps half a magnitude. While observational and theoretical studies. Obser- scarcely more than qualitative due to the vationally, the data consist of star counts or sur- smallness of the sample, this result is nicely face brightness measures as a function of radius consistent with figure 1. from the cluster center. To a first approxima- There are several observable globular clusters tion, these quantities vary as (constant+r2)""2, inhabiting the immediate vicinity of the and the corresponding integrated luminosity galactic nucleus. One or two of these may be rapidly approaches a well-defined asymptotic transients just passing through, but others are limit for large r. This two-dimensional profile suspected of being permanent residents and of is very different from that of an elliptical being quite peculiar. It would be especially galaxy, whose integrated luminosity continues interesting to see how they fit into the over-all to increase as log (constant+r) and which type-distribution picture described above. shows no tendency for a limit except that The individual space motions of globular imposed by the domain of its neighbors. clusters in the galaxy are unknown because Since the profile of a globular cluster is a they are too distant to have reliable proper manifestation of internal dynamic conditions, motions. Nevertheless, some general conclu- it is related to the numbers of stars having sions can be drawn from their radial velocities, various masses, to their mass-luminosity ratios, which range from —360 km/sec to +290 and to the manner in which energy is parti- km/sec. As already remarked, the globular tioned. We know that nearly all of the light clusters probably do not share to any large of a globular cluster comes from stars above its degree in the general rotation of the galaxy. main sequence and that their masses should all The present evidence suggests that the rota- be of the same order because they all evolved tional velocity of the cluster system is less than from the same region of the main sequence; a third of the rotational velocity of the disk. thus a single mass-luminosity ratio applies fairly It may be difficult to pin this figure down more well to the surface-brightness profile. This has closely due to the limited number of clusters been checked by noting that the difference in and to their large individual motions. Rel- integrated color between inner and outer regions ative to the clusters, the sun's orbital velocity of a cluster is generally small and that the radial is about 200 km/sec in roughly the same direc- spread of the red-giant population does not tion derived from differential rotation. The differ significantly from the spread of the sub- cluster system shows no indication of a general giants and the blue stars. However, since the contraction or expansion. exchange of energy in the central region of a Efforts have been made to determine whether cluster is believed to be sufficient for equi- globular clusters favor orbits of high ellipticity partition, one should expect the profile for the passing through the galactic nucleus or whether main-sequence membership to be broader than NEW HORIZONS IN ASTRONOMY 175 the profile for the brighter stars, and there is cosmic yardstick for everything beyond. A observational evidence in support of this. recent estimate based on the globular clusters Differing amounts of internal energy per unit associated with M 87 places the cluster mass from cluster to cluster probably account of galaxies at 10 megaparsecs. If new results for their differing degrees of over-all compact- for more distant clusters of galaxies are adjusted ness. to this scale value, one obtains a Hubble con- The origin of globular clusters is still shrouded stant of 150 km/sec per megaparsec, whose in mystery. It seems fairly evident that most of inverse is 6.5 billion years. Much more work them in our own galaxy were formed during an is needed on this problem, not only in connec- early epoch in its evolution. We do not know, tion with the photometry of globular clusters however, whether they condensed out of inter- in remote galaxies but also in connection with stellar material originally in the domain they establishing a better luminosity function by now inhabit or whether they were created in which the clusters belonging to different galaxies denser central regions and ejected due to the can be compared. We must also worry about high turbulent energy which may have existed whether the luminosity function for clusters is at that time. M 87, which is a very massive necessarily the same in all kinds of galaxies. elliptical in the of galaxies, pro- A fair amount of observational work has been vides a possible clue to this question. It has done on the globular clusters in the Andromeda two unusual features: there are an astonishing galaxy (M 31). In some respects they provide number of globular clusters associated with it, a better sample than those in our own galaxy and there is a peculiar blue jet streaming from for investigating colors, luminosities, compact- its nucleus with several pronounced globules nesses, motions, and distribution. The brightest along it. Is it possible that M 87, because of are around 15th apparent magnitude, which its unusual mass and energy, has a slow rate of corresponds to M~ — 9. They are close enough dynamic development and is still in the stage that we can even compare their profiles with of making globular clusters? Perhaps the those of clusters in our own galaxy to obtain an range of integrated color indices of the M 87 obscuration-free check on the distance of M 31. clusters will help to decide whether any of them The present paper outlines some of the can be young, but the problem is observa- problems in a very broad field. So many dif- tionally difficult. It is also of interest to note ferent workers have had a part in building our that the radial distribution of the globular present state of knowledge that it would have clusters in M 87 is approximately the same as been foolish to attempt the incorporation of the distribution of the nonclustered population (judged from the surface brightness profile). anything approaching a complete system of Since the relaxation time of a galaxy is theoret- bibliographical references into the present text. ically long compared with the average age of An excellent 80-page bibliography, by Helen B. these clusters, this result implies that the condi- Sawyer, already exists (Publ. D. Dunlap Obs., tions of formation have been such as to produce vol. 1, p. 383, 1947), but anyone actually plan- similar distributions since the beginning. ning to work in the field will also want to become acquainted with at least part of the In closing, we should not neglect to consider work done since that date. Recent publica- the value of globular clusters as tools in other tions and current programs include work by astronomical problems. Their usefulness in the following: Arp, Artiukhina, Baade, Baum, surveying the structure of our own galaxy is W. Becker, Belserene, Bernstein, Cuffey, already evident from earlier remarks, but their Deutsch, Eggen, Fatchikhin, Florya, Gaposch- importance in extragalactic studies also deserves kin, Gascoigne, Greenbaum, Greenstein, Hoag, comment. Globular clusters belonging to other von Hoerner, Huang, Idlis, Irwin, H. L. John- galaxies are convenient distance indicators. son, Kholopov, I. King, Kron, Kukarkin, Their luminosities can be used for intercom- Lohmann, Mayall, Melbourn, Minin, Morgan, paring the distances of galaxies as far as the Munch, Oort, Oosterhoff, Osterbrock, Parenago, Virgo cluster of galaxies, which serves as a 176 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TOUI

Reddish, Roberts, Roman, Rosino, Sandage, of these activities are summarized and refer- Savedoff, Sawyer-Hogg, Schurer, Schwarzschild, enced in the 1948, 1952, and 1955 reports of Shapley, Stebbins, Swope, Tayler, Walker, Commission 37 of the International Astronom- Wahlquist, Wallerstein, Whitford, Wilkins, ical Union (see Trans. Int. Astron. Union, O. C. Wilson, van Woerden, and others. Some vols. 7-9). Stellar Structure and Evolution

By M. Schwarzschild 1

The theory of the internal constitution of the processes will be needed from nuclear physics. stars entered a basically new phase in 1938 Similarly, for many star clusters color-magni- with the discovery of the particular nuclear tude diagrams of high precision at faint mag- processes that provide the main energy sources nitudes are needed from observational astron- for the stars, and, soon thereafter, with the omy. In both these fields, which are funda- fairly accurate determination of the reaction mental to the theory of stellar evolution, rates of these processes. progress has been very rapid and encouraging It had long been known that a star of given in recent years. There are, however, three mass and chemical composition has a unique other neighboring fields which should be equilibrium structure. But it was not possible actively cultivated if the investigation of stellar actually to determine this unique structure evolution in its broadest sense is to progress. as long as one of the fundamental relations, Computing facilities are our first concern. that for the rate of energy production by The basic problem of stellar structure and evolu- nuclear processes, was unknown. Now that tion can be formulated mathematically in the necessary nuclear data are available we terms of a fourth-order boundary-value prob- can not only determine the internal structure lem, whose solution depends on the stellar of a star in its initial state when it just starts mass and chemical composition and varies with burning its hydrogen fuel, but we can also time. To find the solution, numerical methods follow, step by step, in a quantitative non- have to be employed. Much useful prelimin- speculative manner, the evolutionary changes ary computation can be done with desk ma- caused by the nuclear transmutations. chines. The numerical work needed, however, Many possibilities are now within our grasp: to obtain a sufficient range of solutions with to interpret the Hertzsprung-Russell diagram satisfactory detail and accuracy is so extensive in terms of stellar evolution, to determine the that it cannot be tackled effectively with desk ages of star clusters and even of individual calculating machines; large electronic machines stars, to sort out the oldest stars and by analyz- are needed. At a small number of institutes ing them to gain insight into the composition such large machines have generously been made of our galaxy at its earliest phases, and to available for pure astronomical research. Astro- obtain clues to the origin of the elements by physical research absolutely requires the help a comparative analysis of the oldest and the of large electronic computers if it is to grow youngest stars. from its present promising beginning to the Only the very first steps have as yet been maturity the problems demand. taken in this large field of new and fundamental The second area we need to develop is photo- problems. The future speed and success of metric spectroscopy with high dispersion to de- the developments will depend not only on the termine the chemical composition of many sam- number of specialists working in the subject ple stars. This observational work, which is and on their ingenuity but also, much more basic to many branches of astrophysics but than in earlier decades, on the advances in which is very tedious and time consuming with neighboring scientific and technical areas. traditional techniques, has progressed very Clearly, more highly accurate determinations slowly in recent years despite its importance. of reaction rates for a larger variety of nuclear These investigations may expand as soon as it 1 Princeton University Observatory, Princeton, N. J. becomes practicable to use electronic image 177 178 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS tubes for this purpose. The advent of this new pheres would be highly valuable. This is, in instrumentation, however, will have its full rev- principle, possible for the photosphere of the olutionizing effect only if at the same time a sun. In practice, however, it is extremely diffi- larger number of technicians becomes available cult because of interference from the disturb- to build and maintain the new instruments and ances of the earth's atmosphere. In conse- to reduce the data. quence, very little progress has been made in The third field is that of the turbulent phe- this field during the past 50 years. Now, how- nomena in stellar atmospheres. Recent inves- ever, observations from high-altitude tigations have shown that these phenomena are appear to open up new possibilities; if they can of interest not only for the theory of stellar at- be realized, the final major difficulty in the mospheres but also play a decisive role in de- theory of the stellar interior may be resolved. terminations of the stellar radius for many types At present, no serious block seems probable in of stars; hence they are essential for the com- the necessary development of any of these fields. plete theory of stellar structure. Laboratory ex- Thus, research in the theory of stellar structure periments and theoretical investigations in hy- and evolution presents many unsolved, but solv- drodynamics have resulted in great progress, able, basic problems and may result in a great during the last decade, in our understanding of advance in answering objectively certain cosmo- turbulence. However, in this difficult subject logical questions. The rate of successful prog- extrapolation from the laboratory to the stars ress in this field will depend essentially on the will always involve some risk; therefore, direct ingenuity and the number of workers who apply observations of turbulence in stellar atmos- themselves to it. Radio Sources By John D. Kraus *

Perhaps you have seen a distant landscape telescope. By analogy, the antenna and re- compared in two photographs, one taken in ceiver may be called a radio telescope. ordinary light and the other in infrared. In the As observed with a radio telescope, the sky ordinary photograph the distant objects are looks strangely different. None of the familiar largely hidden by haze, but in the infrared stars of the night sky are observable, but in picture the distant hills and valleys stand out their places are many radio sources forming clearly. Infrared rays are longer than light totally unfamiliar . Some of these rays and more easily penetrate the haze. sources can be identified with supernova rem- Radio waves are millions of times longer than nants and other gaseous nebulosities and with light and have an even greater power of extragalactic nebulae. There are a couple penetration than infrared. Because of this dozen sources thus identified, at least tenta- fact, radio—applied to astronomy—cuts tively, but there are hundreds more not through the cosmic haze caused by the clouds identified with any optical object and the bulk of gas and dust in interstellar space and of them may never be. It is as though radio presents a view of vast regions that will always and optical telescopes "see" two different remain hidden to ordinary telescopes. Radio with only an occasional object has, in effect, opened a new window on the observable by both. And those objects that universe. can be detected by both appear to be of different It was about 1873 that Maxwell postulated sizes and shapes in the two telescopes. For the possibility of radio waves. About 15 example, the sun has a visual diameter of about years later Hertz produced them and demon- one-half degree, but at certain - strated their transmission over a distance of a lengths the diameter measures several degrees few feet. In 1901 Marconi received radio because the radio waves originate in the upper signals across the Atlantic and an era of world- regions of the corona or solar atmosphere. wide radio communication was begun. Thirty If our eyes were sensitive to radio waves years later detected radio waves from instead of to light, the winter sky would appear the center of our galaxy and initiated a new somewhat as suggested in figure 1. The method for exploring the cosmic universe. It contours are lines of equal surface brightness was 60 years from Maxwell to Jansky, from a showing the radio background. The back- man's idea to its cosmic application. And ground is a maximum along a ridge close to the today, some 20 years after Jansky's discovery, plane of our galaxy or Milky Way system. On this application of radio to astronomy is this background are dots and circles that show rapidly developing into a new science called the position of some of the strongest or brightest radio astronomy. radio sources. Their brightness is indicated by Radio observations of outer space are made a radio magnitude scale analogous to the visual with an antenna and a radio receiver. The magnitude scale. Some of the sources are of an antenna gathers and focuses the radio waves as extended type with diameters of more than 1°. does the lens of an , while the These are shown by the open circles. Others receiver detects and records the signal much as are more localized or pointlike, and these are does the photographic plate of an optical indicated by the solid dots. A few sources 1 Ohio State University, , Ohio. have been identified with optical objects and 179 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS

•TV N6C2244// /

-40

FIGURE 1.—Map of the radio sky at a wavelength of 124 cm made with the radio telescope at Ohio State University (KoandKraus, 19SS). for these the optical designation is given. Note One of the outstanding achievements of radio for example the Crab Nebula (Ml), which is a astronomy was the identification of the second , or the gaseous nebulae brightest radio source (other than the sun) designated IC 443 and NGC 2244. However, with a faint pair of colliding galaxies at a dis- many of the sources on the map have not been tance of some 200 million light-years. Team- identified with an optically observable object work by radio astronomers at Cambridge, and this is true for most radio sources, partic- England, and Sydney, Australia, and optical ularly the weaker ones not shown on the map. astronomers at Mount Palomar was required The identification of radio sources with for several years before the identification was optical objects is a challenging problem that made (Baade and Minkowski, 1954). If this requires close cooperation between radio and object were at 10 times the distance it would optical astronomers. Accurate radio position still be a readily observable radio source but and size determinations are vital in this work. would then be at or beyond the range of the NEW HORIZONS IN ASTRONOMY 181 largest optical telescope. It is this fact which of any optical telescope. Studies of these suggests that radio telescopes may open to ex- sources, individually and collectively, are bound ploration vast regions of space that lie beyond to supplement and alter our purely optical the range of any optical instruments. picture of the universe in a most significant The known radio sources are now numbered way. From Maxwell to Hertz, to Marconi, to in the hundreds, with lists published by Ryle, Jansky, to the outermost reaches of our universe Mills, Bolton, Brown, Haddock, and Ohio are big steps, but they do not end there and State University. However, those that are new advances are being made daily. Perhaps reliably established are much less numerous you may become the radio astronomer who and those identified with optical objects are will make the next big step. only a dozen or two (Pawsey, 1955). References The work of radio exploration has barely begun and great discoveries lie ahead as larger BAADE, W., AND MINKOWSKI, R. 1954. Astrophys. Journ., vol. 119, p. 206. radio telescopes are built and more radio sources Ko, H. C, AND KBAUS, J. D. are found. Many of these sources may emit 1955. Ohio State Univ. Radio Obs. Rep. No. 4, too little light for optical detection or may be June 22; in Sky and Telescope, vol. 14, p. 370. hidden by clouds of cosmic dust, while many PAWSEY, J. L. others may be at distances beyond the range 1955. Astrophys. Journ., vol. 121, p. 1. o