Russell, Henry Norris
RUSSELL, HENRY NORRIS
(b. Oyster Bay, New York, 25 October 1877; d. Princeton, New Jersey, 18 February 1957)
Russell was educated at home until the age of twelve before attending a preparatory school in Princeton, the home of his maternal grandparents. His father, Alexander Russell, a Scottish-Canadian immigrant to the United States, was a Presbyterian minister in Oyster Bay. Both Russell’s mother and maternal grandmother had had some advanced formal education and an outstanding gift for mathematics—a trait especially strong in Russell himself. He graduated insigne cum laude from Princeton University in 1897 and remained there to obtain his doctorate (1900). His dissertation was entitled “The General Perturbations of the Major Axis of Eros, by the Action of Mars.”
After completing graduate work, Russell suffered a serious breakdown of his health and spent much of the following year at Oyster Bay. In the fall of 1901 he returned to Princeton and the following autumn began a three-year stay at Cambridge University. During the first year he was a student at King’s College and also worked at the Cavendish Laboratory. The last two years were spent at the Cambridge observatory as a research assistant supported by the Carnegie Institution of Washington. There he worked in association with Arthur Hinks on a program of determining stellar parallaxes by photographic means. In September 1904 Russell was again taken seriously ill, and in his absence the remaining observations were made by Hinks.
In 1905 Russell accepted a post as instructor in astronomy at Princeton. In 1911 he was appointed professor of astronomy and, in 1912, director of the observatory, positions he held until his retirement in 1947. From June 1918 to early 1919 he was a consulting and experimental engineer in the Bureau of Aircraft Production of the Army Aviation Service. His chief responsibility was a study of problems in aircraft navigation, which included making observations in open aircraft at altitudes of up to 16,000 feet.
In 1921 Russell began his association with the Mt. Wilson Observatory, where he was a research associate until his retirement. In this capacity he usually spent two months of each year at the California observatory. Following his retirement he held research appointments at Lick and Harvard observatories.
Russell was a member of the American Astronomical Society (president, 1934–1937), the American Philosophical Society (president, 1931–1932), the National Academy of Sciences, the American Association for the Advancement of Science (president, 1933); an associate of the Royal Astronomical Society; a foreign member of the Royal Society; and a correspondent of the academies in Paris, Brussels, and Rome. He was president of the commissions of the International Astronomical Union on stellar spectroscopy and on the constitution of stars. Russell was awarded the Draper, Bruce, Rumford, Franklin, Janssen, and Royal Astronomical Society gold medals, and the Lalande Prize. In 1946 the American Astronomical Society established the annual Henry Norris Russell lectureship in his honor.
In the course of nearly sixty years of research, Russell concerned himself with most of the major problems of astrophysics. His principal contributions, however, can be summarized into four general categories. First, Russell presented (in 1912) the earliest systematic analysis of the variation of the light received from eclipsing binary stars; he later pointed out the importance of the motion of the periastron of the orbit in providing information about the internal structure of the component stars. Second, on the basis of his parallax studies, Russell developed a theory of stellar evolution that at the time was in good agreement with the known data. This work stimulated other astrophysicists, especially Arthur Eddington, and was the original context in which he introduced the Hertzsprung-Russell diagram. Third, in the 1920’s Russell began a series of quantitative investigations of the absorption-line spectrum of the sun that resulted in a reliable determination of the abundance of various chemical elements in the solar atmosphere. This work provided clear evidence of the predominance of hydrogen in the sun and, by inference, in most stars. Fourth, Russell carried out, with various co-workers, extensive analyses of the spectra of a number of elements, those of calcium, titanium, and iron being the more important. In this work he developed empirical rules for the relative strengths of lines of a given multiplet, and with F. A. Saunders he devised the theory of L-S coupling to explain spectra produced by atoms with more than one valence electron.
Russell’s first research papers were published while he was a student at Princeton. Several of them, and most notably his dissertation, dealt with problems in celestial mechanics and orbit determination. In light of his later work, however, the most interesting of these early studies was a short paper (1899) that showed how an upper limit to the densities of Algol-variable stars could be obtained. At this time the idea that these were binary stars, the components of which eclipsed each other, was not universally accepted. Only in a few particular cases had orbits been derived from the light variation. Recognizing that an upper limit to the sum of the diameters of the two stars (relative to the size of the orbit) could be determined from the duration of the eclipse, Russell derived limits for the mean densities of seventeen systems. After considering the possible systematic errors in his method, he concluded that Algol-variables were, as a class, much less dense than the sun—a determination that became important in his later work on stellar evolution.
At Cambridge, Russell learned about astrometric methods from Hinks, and in 1903 they embarked upon a program of photographically determining stellar parallaxes. Their two main objectives were to find the most suitable compromise between the amount of work done and the accuracy achieved, and to eliminate all known sources of systematic error. Because photographic techniques in astrometry were still plagued with difficulties and were generally considered inferior to visual observations, Hinks and Russell found it desirable to reconsider the entire observing procedure. The technique they developed was similar to that devised at about the same time by Frank Schlesinger in the United States and was one of the first modern parallax programs.
Russell completed the work of measuring all of the photographic plates and reducing the data in 1910. To him the most interesting result was the correlation between the spectral types and absolute magnitudes of different stars. For those of known parallax, the absolute magnitude decreased systematically from type B to type M—that is, from stars of high surface temperature to those of low surface temperature. This conclusion seemed contrary to the general opinion that many cool, red stars were at great distances and thus were of high luminosity. Russell pointed out that such stars were systematically excluded from parallax studies (since the parallax was undetectably small) and that all the known data could be accounted for under the assumption of two distinct groups of red stars.
To explain the existence of these two types of red stars, Russell adapted a theory of stellar evolution proposed by August Ritter and modified by Sir Norman Lockyer. According to this theory the stars first appeared as highly luminous cool objects that contracted and grew hotter until the high density of the gas caused a significant reduction in the compressibility. Thereafter the star decreased in brightness and in surface temperature. Thus the two kinds of red stars were representatives of the first and last stages of stellar evolution. This finding was in striking contrast with the prevalent view that stars evolved continuously from class B to class M.
To support his ideas, Russell returned to the study of binary stars. In several short papers presented at meetings of the American Astronomical Society between 1910 and 1912, he provided observational evidence that the basic distinguishing feature of the two groups of red stars was their density—as his theory demanded. He also saw other evidence for the correctness of his theory in the orbits of binary stars. Developing the ideas of G.H. Darwin (whose lectures he had attended at Cambridge) on the formation of binary stars from the fission of a single, rapidly rotating star (1910), Russell argued that the youngest stars were single and that binary stars with short orbital periods did not form until the density of the contracting star had become fairly high. The empirical evidence was that bright red stars were rarely members of binary systems and that when hot type-B stars were members of such systems the orbital periods were quire short. Thus Darwin’s picture of the development of binary systems agreed quite satisfactorily with Russell’s theory of the evolution of the individual stars.
The complete account of his theory of stellar evolution that Russell gave in December 1913 served to make his work more widely known. In this lecture he presented graphs plotting absolute magnitudes of stars against their spectral types (now known as Hertzsprung-Russell diagrams), which he used to illustrate the empirical evidence for his theory. It was also at this time that the terms “giant” and “dwarf” came into use, largely through his papers, to describe the two groups of stars, although it is not clear who actually coined them.
While developing his ideas on evolution, Russell began a systematic study of the interpretation of the variations in intensity of the light from eclipsing binary systems. In the decade since he had written his paper on the densities of Algol-variables, two developments had made such a study desirable: the availability of data of much greater precision and completeness—primarily from the observations of Raymond S. Dugan at Princeton and from the photoelectric observations of Joel Stebbins at the University of Illinois—and Russell’s need for the densities of individual stars as further supporting evidence for his theory of stellar evolution. These densities could be obtained only by detailed analysis of eclipsing binaries.
In the first of four papers (1912) on the subject, Russell stated his objective of determining both the orbital elements of the system and the dimensions and brightnesses of the component stars from the observed light curve. The problem was first reduced to one of the simplest cases—two spherical stars, seen as uniformly illuminated disks, moving in a circular orbit. Russell then showed under what circumstances a complete solution could be obtained and gave tabular values for the special functions that were required. His solution emphasized the importance of accurate observations of the binary system at all phases, not only at the primary eclipse.
The remaining papers showed how the basic solution could be extended to more realistic representations of the binary system. In the second paper Russell introduced the refinements of elliptical orbits of small eccentricity and of stars distorted into ellipsoids by their mutual gravitational attraction. Russell handled the latter effect by what he called “rectification” —a transformation of the observed light curve to removed the effects of ellipticity and to reduced the problem to the previously studied case of spherical stars.
At the end of this paper Russell briefly outlines a means of handling the problem of limb darkening—the decrease in brightness of the stellar disk near its apparent edge. The extensive calculations for this part of the theory were assigned to Harlow Shapley, who had recently arrived at Princeton as a graduate student. The results of their collective efforts were presented in the third and fourth papers of the series, in which the systematic treatment of eclipsing binaries was extended to those with limb-darkened components. Under Russell’s direction Shapley later applied the new techniques to ninety eclipsing binaries, thereby providing a large number of new density determinations.
For nearly thirty years the standard techniques of dealing with eclipsing binary systems were essentially those introduced by Russell, and much of his nomenclature and notion became a permanent part of the subject. The wide acceptance of this work can in part be attributed to the very practical manner in which the analysis was presented. There were references to how much time certain calculations required and comments that certain refinements were not worth the work. Indeed, this was a characteristic feature of many of Russell’s papers, in which a balance was struck between ease of computation and precision of results.
From 1914 to 1921 Russell worked on various subjects, some of which were a continuation of his study of stellar evolution. He published several papers on the orbits of visual binaries and the determination of the masses of the component stars. He continued a project on the photographic determination of the position of the moon, carried out jointly by the Harvard. Yale, and Princeton observatories. A review of the determination of the albedoes and magnitudes of planets and satellites was also conducted. In 1921 Russell showed that the age of the earth’s crust was about 4 X 109 years, basing his statement on the radioactive decay of uranium and the abundance of its end products, lead and thorium.
Two developments in 1921 marked a major shift in Russel’s career. The first was the publication of M.N. Saha’s theory of the ionization of atoms in stellar atmospheres; the second was Russell’s appointment as a research associate at the Mt. Wilson observatory, which brought him into close association with Walter S. Adams and other astronomers and spectroscopists in California.
During his first visit to Mt. Wilson, in the summer of 1921, Russell investigated the application of Sah’s formula to the sun. To do so it was necessary to generalize Saha’s original theory, which described a gas composed of a single atomic species. Russell pointed out that in a mixed gas, such as the sun’s atmosphere, the ionization relationships were more complex because one of the products of the ionization process, the free electron, was common to the ionization reactions of all atoms. Thus the equilibrium state of ions and electrons could be determined only for all elements simultaneously. The result of this analysis was that the degree of ionization of an atom, but also upon the relative abundances of other atomic species and their ionization potentials.
Russell undertook a critical test of the expanded theory by comparing the spectrum of the normal solar photosphere with that of sunspots. Since the temperatures of both the photosphere and the sunspots were well known and the pressures, although not well determined, were assumed to be equal, the relative strengths of absorption lines in the two spectra provided the desired comparison of theory and observation.
The most exact comparisons were for the alkali metals, since their ionization potentials were known. In particular Saha had predicated that the lines of neutral potassium would be stronger in the spot spectrum and that lines of neutral rubidium would be faintly visible in the sunspots, although they had not been detected in the normal solar spectrum. Russell’s examination of spectra taken at Mt. Wilson confirmed these predictions and, in general, supported the Saha theory.
This success made Russell keenly aware of the tremendous possibilities that the new theory offered. The spectra of stars could now be used to give quantitative information about the state of the atmosphere where the lines were formed. In concluding his paper he wrote:
The possibilities of the new method appear to be very great. To utilize it fully, years of work will be required to study the behavior of [the alkali earths, scandium, titanium, vanadium, manganese, and iron] and others, in the stars, in laboratory spectra, and by direct measurement of ionization, but the prospect of our knowledge, both of atoms and of stars, as a result of such researches, makes it urgently desirable that they should be carried out [“Theory of Iionization.” pp. 143–144].
Although he continued to investigate problems relating directly to the atmospheres of stars, such as the theoretical determination of the atmospheres of stars, such as the theoretical determination of the pressure at the solar photosphere (carried out jointly with John Q. Stewart), Russell soon turned to the determination of atomic structure through the study of spectra. His first major effort in this direction was a study, in collaboration with the Harvard spectroscopist F.A. Saunders, of the spectra of the alkali earths—calcium, scandium, and barium (1925). In 1923, when this work was being carried out, atomic theory was unable to explain “complex” spectra—the spectra of elements other than hydrogen, helium, and the alkali metals. Of the remaining elements, the spectra of the alkali earths were partially understood in terms of the Bohr atom: but it was clear that energy levels existed that did not fit into the regular series of terms.
Russell and Saunders had found several groups of lines in the ultraviolet spectrum of calcium that led to the identification of three new “anomalous” triplet terms for this element. With the additional data they were able to find some systematic relationships among the anomalous terms. The most important discovery was that the energies of some of these terms were greater than the ionization potential of the atom, a fact also recognized by Gregor Wentzel at about the same time. They interpreted this result as evidence that the anomalous terms were produced by an excitation of both optical electrons. This idea explained not only why an atom can absorb energy greater than its ionization potential, but also why the alkali metals—which have only one valence electron—do not have any anomalous terms.
From this basic concept of excitation of more than one electron (which had also occurred to Bohr), Russell and Saunders proceeded to extend Alfred Landé’s vector model to account for the quantum numbers and energies of the anomalous terms. Landé’s model identified the azimuthal quantum number with the orbital angular momentum of the electron, the multiplicity of the spectroscopic term with the angular momentum of the rest of the atom (the Rumpf), and the inner quantum number with the vectoral sum of the two. Russell and Saunders assumed that the quantized angular momentum of the individual excited electrons could be combined first, and the resultant combined with that of the Rumpf. This technique of handling complex spectra, which later became known as L-s coupling, proved quite successful in predicting both the energy levels for the anomalous terms and the observed transitions to those levels resulting from the excitation of only a single electron.
The final section of Russell’s and Saunders’ paper is of some interest in the history of spectroscopy for its attempt to introduce uniformity into the chaotic state of spectroscopic notation. The proposed system became the basis for the modern notation, although it was late refined by Russell, Allen G. Shenstone, and L.A. Turner (1929).
Russell subsequently turned to the problem of finding formulas that could represent the relative intensities of the spectral lines of a particular atom (1926). He accomplished this for the lines of a given multiplet, that is, all the lines arising from transitions between the various levels of two spectroscopic terms. (It is an indication of the vigorous activity in this field in 1926 that two other spectroscopists, R. Kronig and Sommerfeld, derived similar formulas at the same time.) Determined without any detailed theory of atomic structure, the intensity formula was based upon Bohr’s correspondence principle and a rule for the sum of the intensities of the lines having a given initial or final level. Thus, this work was carried out entirely within the framework of the “old” quantum mechanics, as was essentially all of Russell’s spectroscopic work.
The motivation for these spectroscopic studies was, as Russell had indicated in 1921, not only atomic but also astrophysical. By 1928 he was able, in collaboration with Walter Adams and Charlotte Moore, to bring this new knowledge to bear on stellar spectra. The first problem to be solved was one of calibration: how could one deduce the number of atoms in the solar atmosphere that were responsible for producing a particular absorption line? Two methods appeared possible – a measurement of contours of the lines, followed by a theoretical interpretation in terms of atomic physics, or a direct calibration of the empirical Rowland intensity scale of spectral lines in terms of numbers of atoms. The latter method was chosen. Russell, Adams, and Moore assumed that the intensity of the lines, as derived from Russell’s multiplet formulas, was proportional to the number of atoms acting to produce the line. The problem then reduced to that of calibrating the Rowland scale in terms of the theoretical one. Since the Mt. Wilson observatory was then revising Rowland’s table of absorption lines in the solar spectrum, abundant information was available. A comparison of these data with the multiplet intensity formulas gave a relative scale of intensities—relative in the sense that although the shape of the curve was determined, the zero point was not; the zero point, in fact, proved to be different for each multiplet.
Yet even this relative calibration was of considerable interest. Adams and Russell used the results in conjunction with the Saha-Boltzmann relationship to compare the atmospheres of different stars (1928). Using the sun as a standard, they analyzed seven stars on the basis of high-dispersion spectra taken with the 100-inch telescope. In calculating the relative populations of excited states of atoms in different stars, they found–as expected–that in hotter stars the population of higher states was greater . In the cooler stars, however, the dependence on excitation potential was not what was expected (the Adams-Russell effect), leading them to believe that the atmospheres of cool stars were not in thermodynamic equilibrium. Their temperature determinations of these stars were in substantial agreement with the results of other methods, and the values they found for the partial pressure of free electrons emphasized the extremely low densities in red giant stars.
Russell continued this analysis in 1929 with the help of Albrecht Unsöld’s measurements of line profiles. Unsöld’s work provided an absolute calibration of the number of atoms involved in producing an absorption line. Because of the amount of work involved in his procedure, only a relatively small number of lines had been so analyzed. Using this work to provide the zero point for his own scale, Russell developed an absolute calibration scale for the Rowland intensities. He then showed that the total abundances of elements in the sun could be calculated by taking into account the atoms in various states of excitation and ionization. In this manner Russell determined the abundances of fifty-five elements and several molecules in the solar atmosphere. In many instances the abundance ratios between elements was similar to that in the earth’s crust—with one notable exception. Hydrogen proved to be by far the most abundant element. This discovery was not completely unexpected, for Cecilia Payne had earlier found high abundances of hydrogen in giant stars but had dismissed the numerical values as “spurious.” Russell’s analysis had proceeded on more solid footing, however, and he was also able to show that the high abundance of hydrogen actually removed several other apparent difficulties in the analysis of the sun. Thus, there was clear evidence of the dominant abundance of hydrogen in the sun and, therefore, in most stars. It is difficult to overestimate the importance of this result in the development of astrophysics, since much of the subsequent progress has depended upon recognizing the predominant role of hydrogen in astrophysical processes.
Although this discussion of Russell’s work in the 1920’s might suggest that he proceeded single mindedly toward the goal of a quantitative analysis of stellar spectra, Russell also investigated related matters. In atomic spectroscopy he carried out detailed analyses of several elements (most notably titanium, iron, and scandium); and on the basis of Friedrich Hund’s theory of complex spectra, he found systematic similarities in the spectra of the elements of the iron group (those in the periodic table from potassium through zinc).
Eddington’s discovery (1924) that the ideal gas law was applicable to the interiors of stars on the main sequence led Russell back to his old theory of stellar evolution, which had now been rendered untenable. Recognizing the chief problem to be the source of energy, he postulated highly temperature-sensitive processes of transforming matter to energy (1925). Thus, as a star contracted and grew hotter, the energy source would become activated and contraction would cease. The main sequence and giant branch were thus interpreted as stages in which different processes were active. As one of these processes proceeded, mass was converted into energy and the total mass of the star decreased. With the decrease, the temperature and luminosity changed as, consequently, did the position of the star in the Hertzsprung-Russell diagram. The main sequence and giant branches were again seen to be the evolutionary paths taken by stars. This theory was challenged by James Jeans, who claimed that Russell’s stars would be unstable and who preferred an energy source the rate of which was independent of temperature or pressure. No substantial progress in unraveling the complexities of stellar evolution occurred, however, until the identification, fifteen years later, of the particular nuclear reactions occurring inside stars.
Russell also contributed to the theory of stellar structure. He suggested that instead of calculating models with a specific opacity formula or equation of state, an attempt should be made to postulate only very general principles and to search for distinctive relationships among stars (1931). The most important result of this approach, now known as the Vogt-Russell theorem, was that on very general grounds the properties of a star can be expected to be completely determined by its mass and chemical composition. Heinrich Vogt had derived a similar result, but apparently his work was not well known in England and the United States, where most of the research in stellar structure was being conducted.
Eddington’s work had made the question of mass distribution inside a star an important one, and Russell realized that there was an empirical method of estimating the ratio of the mean density to the central density of a star (1928). In close binary systems the interaction of the distorted stars would result in an advance of the periastron of the orbit, and this advance could be detected from the light curve of an eclipsing binary system. Although Russell’s initial results were not very satisfactory, his method was sound; and later investigators were able to derive better results.
During the 1930’s and 1940’s, Russell continued to work on most of the subjects that had occupied him in the previous twenty years. He made detailed analyses of the spectra of several more elements. His study of the orbits of visual binary stars and the masses derived therefrom led to the publication, with Charlotte Moore, of a monograph on the subject (1940). Russell enlarged his work on the chemical composition of the sun to include a study of molecular abundances (1933). He also returned to the analysis of eclipsing binary systems, extending his methods to more complicated systems and considering further the effects of the internal structure of the stars upon the advance of the periastron (1939, 1942).
While Russell continued to make significant contributions to astrophysical research during the latter part of his career, his role as an adviser and consultant to other astronomers became increasingly important. His yearly trips from Princeton to Mt. Wilson afforded many opportunities to visit other American observatories; and because his own interests covered such a wide range, he was often quite familiar with the problems on which others were working.
Russell’s role as a critic and reviewer of contemporary research began at least as early as 1919, when, at the request of the National Academy of Sciences, he wrote a comprehensive review of current research in sidereal astronomy that pointed out to his colleagues some of the most important problems to be solved. Another indication of his interest in analyzing the work of others is his review of the dispute between Jeans and Eddington over the mass-luminosity relation for stars (1925). By placing their arguments within the framework of a more general theory, Russell resolved the apparent discrepancy between their results.
Russell’s interest in teaching began fairly early in his career. By 1911 he had started a revision of the general astronomy textbook written by his predecessor at Princeton, Charles A. Young, but other work delayed its completion. With the collaboration of his Princeton colleagues Raymond Dugan and John Q. Stewart, it was finally published in 1926. The first volume was the originally projected revision of Young’s book; the second contained largely new material, most of it written by Russell. This textbook was widely used for thirty years, and many American astronomers trained in this period were introduced to the subject through studying it.
After 1930 especially, Russell gave a number of lectures in which he reviewed the progress in various areas of astrophysics. Although new results were rarely presented in these talks, they did serve the important function of summarizing and organizing recent work.
These lectures, and even his textbook, were aimed principally at the scientist or potential scientist. To reach a larger, more general audience, Russell wrote a monthly article in Scientific American. Beginning in 1900 as a column describing the appearance of the evening sky for the coming month, the articles soon included information on recent research in astronomy. By 1911 Russell was regularly including a short essay on some astronomical subject, and this section of the article soon came to be the dominant feature. By 1943, when the last one appeared, he had written 500 short articles discussing all phases of astronomy.
Although in the vast majority of his writings Russell kept strictly to scientific matters, he did on several occasions discuss his ideas concerning science and religion. The fullest exposition was in a series of lectures given at Yale University in 1925 and published two years later as Fate and Freedom. The title reflects one of his central concerns: the conflict between the concept of a deterministic universe and the belief in free will. Although Russell concluded that the universe was completely mechanistic, he felt that the observed behavior of men should be considered a kind of statistical phenomenon and that consequently free will was as real as (to use his analogy) statistical phenomena in physics, such as the pressure of a gas. These ideas were formulated prior to the introduction into quantum mechanics of the uncertainty principle, which Russell in his later writings does not seem to have considered of central philosophical importance.
Over a period of fifty years Russell’s work showed a continuous effort to provide a clear understanding of the physics of stars. Early in his career he focused on stellar evolution and the related problems of determining masses, radii, temperatures, luminosities, and densities of stars. A result of this effort was his series of investigations of eclipsing binary systems. His interest later turned to stellar atmospheres, the problems of determining pressures and temperatures, and the quantitative measurement of chemical abundances. An outgrowth of this work was his extensive work in the theory of atomic spectra.
I. Original Works. Bibliographies of Russell’s published writings are in Poggendorff, V, 1081–1082, and VI, 2249–2250, and at the end of the biographical essays by Shapley and Seaton (see below). The most extensive is that following Shapley’s article, although it does not include abstracts of certain papers presented at meetings of the American Astronomical Society or Russell’s articles in Scientific American.
Russell’s principal publications include “The Densities of the Variable Stars of the Algol Type”, in Astrophysical Journal, 10 (1899), 315–318; “The General Perturbations of the Major Axis of Eros, by the Action of Mars”, in Astronomical Journal, 21 (1900), 25–28; “On the Origin of Binary Stars”, in Astrophysical Journal, 21 (1900), 25–28; “On the Origin of Binary Stars”, in Astrophysical Journal, 31 (1910), 185–207; Determinations of Stellar Parallax (Washington, D.C., 1911): “On the Determination of the Orbital Elements of Eclipsing Variable Stars”, in Astrophysical Journal, 35 (1912), 315–340, and 36 (1912), 54–74; “On Darkening at the Limb in Eclipsing Variables”, ibid., 36 (1912), 239–254, 385–408, written with H. Shapley: “Relations Between the Spectra and Other Characteristics of the Stars”, in Nature, 93 (1914), 227–230, 252–258, 281–286; and “Some Problems of Sidereal Astronomy”, in Proceedings of the National Academy of Sciences of the United States of America, 5 (1919), 391–416.
See also “A Superior Limit to the Age of the Earth’s Crust”, in Proceedings of the Royal Society, 99A (1921), 84–86: “The Theory of lonization and the Sun-Spot Spectrum”, in Astrophysical Journal, 55 (1922), 119–144: “New Regularities in the Spectra of the Alkaline Earths”, ibid., 61 (1925), 38–69, written with F. A. Saunders; “The Intensities of Lines in Multiplets”, in Proceedings of the National Academy of Sciences, 11 (1925), 314–328; “Note on the Relations Between the Mass, Temperature, and Luminosity of a Gaseous Star”, in Monthly Notices of the Royal Astronomical Society, 85 (1925), 935–939; Astronomy, a Revision of Young’s Manual of Astronomy, 2 vols. (Boston, 1926–1927), written with R. S. Dugan and J. Q. Stewart: Fate and Freedom (New Haven, 1927); “On the Advance of Periastron in Eclipsing Binaries”, in Monthly Notices of the Royal Astronomical Society, 88 (1928), 641–643; “A Calibration of Rowland’s Scale of Intensities for Solar Lines”, in Astrophysical Journal, 68 (1928), 1–8, written with W. S. Adams and C. E. Moore: “Preliminary Results of a New Method for the Analysis of Stellar Spectra”, ibid., 9–36, written with W. S. Adams: “On the Composition of the Sun’, Atmosphere”, in Astrophysical Journal. 70 (1929), 11–82; “Notes on the Constitution of the Stars”, in Monthly Notices of the Royal Astronomical Society, 91 (1931), 951–966, and 92 (1931), 146; The Solar System and Its Origin (New York, 1935): and The Masses of the Stars, With a General Catalog of Dynamical Parallaxes (Chicago, 1940), written with C. E. Moore.
II. Secondary Literature. The most extensive biographical essays are those by F. J. M. Stratton in Biographical Memoirs of Fellows of the Royal Society, 3 (1957), 173–191: and by Harlow Shapley in Biographical Memoirs. National Academy of Sciences, 32 (1958), 354–378. Obituary notices include those by Donald H. Menzel in Yearbook. American Philosophical Society (1958), 139–143: and Otto Struve, in Publications of the Astronomical Society of the Pacific, 69 (1957), 223–226. See also Axel V. Nielsen, “Contributions to the History of the Hertzsprung-Russell Diagram”, in Centaurus, 9 (1964), 219–253; and the following articles in Vistas in Astronomy, 12 (1970); Katherine G. Kron, “Henry Norris Russell (1877–1957); Some Recollections”, 3–6; Bancroft W. Sitterly, “Changing Interpretations of the Hertzsprung-Russell Diagram, 1910–1940: A Historical Note, “357–366; and R. Szafraniec, “Henry Norris Russell’s Contribution to the Study of Eclipsing Variables”, 7–20.
Bruce C. Cogan