“Two roads diverged in a yellow wood,
And sorry I could not travel both …
I took the one less traveled by,
And that has made all the difference.”
Robert Frost (1916)1
A Lifelong Princetonian with Cosmic Power
Henry Norris Russell was born on October 25, 1877, at Oyster Bay, Long Island, half of Puritan and half of lowland Scots stock. His father, Alexander Russell, was a pastor of the local First Presbyterian Church, who had graduated from the Princeton Theological Seminary in New Jersey. Henry’s mother, Eliza Norris, met his father when he was staying at her family home in Princeton. The couple had two other sons, Gordon and Alexander, born in 1880 and 1883, respectively.
Henry spent most of his school years in Princeton, where he stayed at the Norris family home, and returned each summer to Oyster Bay. At age 12 he entered the Princeton Preparatory School and enrolled in the local College of New Jersey a month shy of 16. He graduated at age 19 — at the head of his class and with extraordinary honors; by that time the college had changed its name to Princeton University.
The bright young man immediately joined the fledging graduate program at Princeton University, which led to his Ph.D. in mathematical astronomy in 1900 with a thesis on the way Mars perturbs the orbit of the asteroid Eros. But it wasn’t easy. Henry spent all his strength in completing his degree, and had to take two years off to rest. He recuperated from the breakdown in trips with his mother to the island of Capri and other parts of southern Italy.
It was a privileged and sheltered life. Money was never an issue. During his student days, Henry did not earn his own way, even in part. Housing and meals with the Norris family were free, and because he was a minister’s son tuition at Princeton University was waived. His Aunt Ada Norris provided for any other fees, and legacies to his mother from her parents had made their family economic position fairly comfortable, at least with careful spending which came by tradition from the New England housewives who “feared dirt, debt, and the Devil, and nothing else.”2
College life fitted well with Henry’s evangelical beliefs. Most of the Professors were church members, chapel attendance was compulsory, and the students were happy and knew it.3 As an adult, he did not drink coffee or tea, smoke, or indulge in any alcoholic drinks beyond a bit of sherry before dinner. It is even said that his thesis about the asteroid Eros caused him embarrassment, owing to the name Eros, the Greek god of love.
The Princeton Professor Charles A. Young stimulated Russell’s interest in astronomy through his lectures and his Manual of Astronomy. Young also connected his astronomical and religious views, writing that astronomy “reveals the glory and majesty of the Creator, the eternal, omniscient and all pervading God.”4 He also quoted from the Bible, stating in his lectures that: “The Heavens declare the glory of God; and the firmament sheweth his handiwork.”5
After graduation from Princeton, Russell spent nearly two years at the Cambridge University Observatory, where he photographed eclipsing binary stars and helped use them to determine their physical properties.
In 1905 Russell returned to Princeton University, after he persuaded Woodrow Wilson, then President of Princeton, that they needed a good astrophysicist just like himself to discover how stars evolve, how the Universe reached its present form, and what will become of it.6
And that is precisely what Russell built his career on. He became a Full Professor of Astronomy in 1911, and Director of the Princeton University Observatory a year later. He spent nearly his entire professional life at Princeton University, living in the same old Norris-family house from 1890 to 1957.
Like his mentor Professor Young, Henry Norris Russell was a devout Christian, and in later life he also wrote about the importance of non-scientific questions. Human beings, he declared, can be described by their average behavior, with an understanding of any individual that defies complete scientific determination. That is because:
“Science answers only the question: ‘How do these things come to pass?’ ...But the methods of mathematical analysis are no longer of prime importance for the question which the child asks first of all, “Why are things so?”7
Russell wrote that personal freedom and responsibility, and our capacity for good and evil, have to be taken into account when discussing our uncertain human existence, but that we should always keep in mind that “man was created to glorify God and enjoy him forever.”8
Russell believed the Universe exists because there is a Power behind it, and as human beings “we can assuredly have relations of some sort with the Cosmic Power. .... No one feels more keenly than the student of Nature the greatness and splendor of that Power.”9 This transcendent Deity, our God, is “clear outside those limitations of space and time within which the material Universe, and we as parts of it, have our being ... a Being who is the only reason why there is any Universe.”10
For this God, the distinctions of before and after do not exist. The human soul, Russell thought, outlasts death and belongs to this immortal realm. As an astronomer, he likened this soul to starlight that can persist long after a star has ceased shining. “Even if a star is dead and gone, its light lives on — undiminished, individual, immortal.” The immortal human soul may also “survive indefinitely, through an unlimited time, retaining its full individuality, never becoming merged with any other personality or lost in some vague undifferentiated whole.”11
Some of the most distant stars and galaxies may no longer exist, but their light can survive unchanged. As long as it encounters no matter, starlight can travel within empty space forever. Moreover, there is no interference between lights traveling through space from different stars. Thus, even when viewing thousands of stars, a telescope can focus on the light of any particular star and form an individual image of it.
Upon death, Russell believed, we commend our spirit to God. It’s something we can stake our lives on.
Russell’s student Harlow Shapley has written a fine biographical memoir about his mentor’s accomplishments in stellar astronomy,12 and David H. DeVorkin’s book-length biography is filled with rich detail about Russell’s personal and scientific life, and his efforts to integrate theory into mainstream American astronomy.13
For 43 years, starting in 1900, Henry contributed monthly popular articles about the discoveries, status, and progress of astronomy in The Scientific American. This opportunity was due to Princetonian ties, through the owners of the magazine who had attended the University. Russell also wrote a widely used, two-volume textbook of Astronomy, coauthored with two Princeton colleagues R. S. Dugan and J. Q. Stewart. This revision of Young’s Manual of Astronomy was first published in 1926–1927, with supplements and revisions over the next two decades. In 1935, Russell additionally wrote an influential book about The Solar System and Its Origin, and in 1940 he published an important monograph on The Masses of the Stars, with Charlotte E. Moore.
Henry Norris turned down the directorship of the Harvard College Observatory, and was content with wielding more influential power through endless correspondence and visits, from Harvard to the Lick and Mount Wilson Observatories in California, and through his Astronomy textbook and Scientific American columns.
Russell became President of the American Astronomical Society, the American Association for the Advancement of Science, and the American Philosophical Society. He was also awarded the gold medal of the Royal Astronomical Society of England, two medals of the French Academy, five other medals of American scientific societies, and numerous honorary degrees. Altogether, Henry Norris Russell was quite an eminent, influential, versatile, and accomplished astronomer, the ultimate Princetonian.
In spite of his accomplishments, in later life Russell was also tormented by deep depression, self-doubt, and indecision, and suffered from frequent periods of collapse from exhausting overwork and nervous, restless activity.
His insight to the path of stellar life depended on the dedicated work of other astronomers, many of them women.
How Stars Age
Working under the direction of Edward C. Pickering, astronomers at the Harvard College Observatory examined the photographic spectra of hundreds of thousands of stars, and determined their dominant spectral lines and temperatures. These astronomers were mainly young ladies who had studied physics or astronomy at nearby colleges for women, such as Wellesley and Radcliffe. Harvard did not educate females at that time and did not permit them on its faculty.
One of these faithful, stalwart workers was Annie Jump Cannon, who classified the spectra of roughly 400,000 stars between 1918 and 1924.14 She distinguished the stars on the basis of the absorption lines in their spectra, and arranged most of them in a smooth and continuous spectral sequence. The hottest stars, with the bluest colors, were designated as spectral type O, followed in order of declining visible disk temperature by spectral types B, A, F, G, K and M (Table 10.1). Stars that displayed spectral lines of highly ionized elements, for example, were relatively hot because high temperatures are required to ionize atoms. Stars that displayed spectral lines of unionized hydrogen atoms would be cooler, and those with molecular lines cooler still. These spectral classifications eventually led to an understanding of the way stars evolve.
The probable course of a star’s aging might be inferred by assuming it is a large, hot gaseous sphere whose heat is sustained by its contraction. August Ritter, a Professor of Mechanics at the Polytechnic University of Aachen, Germany, published a series of eighteen important papers that described such a process and proposed how it could be related to a star’s spectral type. These papers did not attract much attention until George Ellery Hale, the editor of the newly formed Astrophysical Journal initiated the publication of an English version of the sixteenth paper in 1898.15 Here Ritter introduced a classification of stars with rising temperature as well as a falling one. A star would become hotter while undergoing gravitational contraction, and after reaching the acme of its brilliance, the star would cool and evolve toward extinction. The peak luminous output would occur just before a star has been compressed so much that it could no longer behave as a gas. At this critical density the star could not supply more heat by further contraction, and it would just cool down and radiate the leftover heat away.
Table 10.1. The spectral classification of starsa
aAn H denotes hydrogen, He is helium, Ca is calcium, and TiO is a molecule. The Roman numeral I denotes a neutral, unionized atom, the number II describes an ionized atom missing one electron, and the temperatures are in degrees kelvin.
In the early 20th century it was also proposed that stars might change over time in ways suggested by their spectral lines.16 This set the stage for Henry Norris Russell’s display of stellar luminosity as a function of the disk temperature inferred from spectral lines. Russell’s diagram became a primary tool for tracing the paths of stellar evolution, which is still used today.
Russell’s Diagram
In 1914, Russell published his famous plot of the stellar luminosities against spectral class or disk temperature (Fig. 10.1),17 which displays the trajectory of a star’s life and demonstrates how its properties change with age. For many years this figure was called the Russell Diagram, but eventually astronomers realized that the Danish astronomer Ejnar Hertzsprung had previously published, in 1911, a less extensive diagram relating stellar luminosity and color for the nearby star clusters, the Pleiades and Hyades.18 It therefore became known as the Hertzsprung-Russell diagram, or H-R diagram for short.
Because he had access to the parallax measurements, and therefore distances, of many stars, Russell’s diagram included rare and very luminous red stars of spectral class M, as well as faint red stars of the same class M. According to Russell, the noteworthy aspect of his diagram was that most stars form a continuous slanting progression from the luminous B to faint M stars. The other notable aspect of the diagram was that the red stars were either very faint or very luminous with none observed in between at intermediate luminosity.
Russell predicted that if his survey was extended to many thousands of stars, the diagram would be represented by two lines: one descending diagonally from B to M and the other starting also at B and running almost horizontally.
Figure 10.1 Russell’s diagram In 1914 Henry Norris Russell published this diagram of the luminosity (vertical axis) plotted as a function of spectral class (top horizontal axis) for four moving star clusters. The two diagonal lines mark the boundaries of Ejnar Hertzsprung’s observations of the Pleiades and Hyades open star clusters in 1911; this is now known as the main sequence along which most stars, including the Sun, are located. [Adapted from Henry Norris Russell: “Relations Between the Spectra and Other Characteristics of Stars,” Popular Astronomy 22, 275–294 (1914).]
Another way of expressing this two-fold division, Russell stated, is that there are two stellar classes — giant stars of great luminosity of about one hundred times the Sun’s luminosity, and varying very little in luminosity from one class of spectrum to another, and dwarf stars of lower luminosity, which fall off very rapidly in luminosity with increasing redness.19
Giant and Main-Sequence Stars
As Ejnar Hertzsprung subsequently asserted: “The giant stars are indeed more luminous because they are ‘swollen.’”20 Russell’s study of eclipsing binary stars showed that the luminous giant stars have much larger diameters than other fainter stars of similar spectral class, while the masses of both kinds of stars are comparable. This indicated to him that the very luminous red stars have a smaller mass density and larger size when compared to the faint red stars. This is also just what you would expect from the thermal radiation of a hot, gaseous sphere, whose luminosity at a fixed temperature increases as the square of the radius.
The large size of at least one giant star was fully confirmed by direct measurement of the diameter of the red star Alpha Orionis (Betelgeuse) by Albert Michelson and Francis G. Pease in late 1920. Michelson, who was then located at the University of Chicago, had pioneered the interference method of measuring the angular sizes of sources that are too small to be resolved by a single telescope. The radiation from the source is observed with two connected mirrors that act as an interferometer, or interference-meter, whose changing mirror-separation produces interference fringes with an effective angular resolution comparable to a single telescope with a diameter equal to the separation of the two mirrors.
George Ellery Hale, a friend of Michelson, suggested that he come out to Mount Wilson and use his interferometer to measure the size of Betelgeuse, whose large angular size had been predicted by A. S. Eddington. Michelson and Pease took the long train ride to California, and mounted their mirrors on a 20-foot (6-meter) steel beam placed at the end of the open tube of the 100-inch (2.5-meter) telescope on Mount Wilson. By measuring the mirror separation when the interference fringes disappeared, they concluded that Betelgeuse has an angular diameter of 0.047 seconds of arc.21 Current observations indicate a very similar angular diameter of 0.055 seconds of arc, which means that Betelgeuse is 1,180 times the size of the Sun, and has a radius comparable to the distance between the Sun and Jupiter.
A century of increasingly accurate and extensive observations have confirmed the initial characteristics of the H-R diagram (Fig. 10.2). The majority of stars, including the Sun, lie in a band that extends diagonally from the upper left to the lower right, or from the high-luminosity, high-temperature, blue stars to the low-luminosity, low temperature red stars. Russell dubbed these more numerous stars dwarfs, since they are smaller than the giants, but the designation is confusing. There is, for example, no observable difference between the size and luminosity of the hottest dwarf and giant stars. Astronomers now retain the designation giant stars, but use the term main-sequence stars for the other ones, a name suggested by A. S. Eddington in the 1920s. The stars on the main sequence are the most common type of star in the Milky Way, constituting 90 percent of its stars.
How do Stars Begin their Lives?
Before astronomers could use the Hertzsprung-Russell, or H-R, diagram to trace out a star’s life, they had to determine how stars are born and where they first appear on that diagram. Henry Norris Russell had one interpretation, but he was dead wrong. In his proposal, the stars begin their lives as vastly extended, low-temperature giant stars of spectral type M. Under the influence of gravity they collapse, grow smaller and hotter, and move along the top of the H-R diagram as they age. When the star reaches the upper left-hand corner of the H-R diagram, further compressibility is no longer possible. Cooling begins and the star moves from top left to bottom along the main sequence. It decreases in luminosity and temperature, and ends up as a faint star of spectral type M.
Russell’s interpretation of the H-R diagram had to be abandoned when Eddington showed in 1924 that a single mass-luminosity relation applies to all stars,22 which led Russell to abandon his interpretation of the main-sequence stars. He instead suggested that a star’s mass may be an important factor in solving the great problem of stellar evolution.23
Figure 10.2 H-R diagram for nearby stars The luminosity (left vertical axis), in units of the Sun’s absolute luminosity, L
, plotted against the effective temperature of the star’s disk in degrees kelvin, designated K (bottom horizontal axis) for 22,000 stars in the catalogue of the HIPPARCOS satellite. Most stars, including our Sun, lie along the main sequence that extends from the upper-left to the bottom-right sides of the diagram. Stars of about the Sun’s mass evolve into helium-burning red giant stars, located in the upper-right side of the diagram. (Data points courtesy of ESA/HIPPARCOS mission.)
The Danish astronomer Bengt Strömgren subsequently concluded that it is mass and hydrogen content that determine a star’s position in the H-R diagram. In 1933, he additionally proposed that studies of stars in clusters would help determine the evolutionary interpretation of the diagram.24
Figure 10.3 Gravitational collapse The collapse of an interstellar cloud of gas and dust (left), compresses the cloud, heats it and produces radiation (right). The arrows are pointed in the direction of motion of the gas atoms, and the lengths of the arrows denote their speed of motion.
As it turned out, a star begins it life on the main sequence. Once an interstellar cloud of gas and dust has become sufficiently massive, or after external compression has been provided, it must collapse to make stars, in much the same way that the Sun came into being. The mutual gravitational attraction of its parts will overcome the internal gas pressure and cause this cloud to start collapsing. As this protostar falls inward from all directions, the gas gains energy; a dropped stone similarly gains energy when it moves down to a pool of water. Some of the star’s energy is converted into heat as the gas particles fall inward and collide with each other (Fig. 10.3).
When observations of young stellar clusters were combined with the theoretical studies of the Japanese astrophysicist Chiushiro Hayashi, the premain-sequence evolution of protostars of different masses was deciphered.25 Their tracks in the H-R diagram initially move straight down, and subsequently turn to the left and continue that way until the protostar arrives on the main sequence (Fig. 10.4). The outward pressure of the hot gas, which is now heated by nuclear fusion reactions, prevents the star from collapsing further. It has settled down for a long, rather uneventful life as a mainsequence star, the longest stop in its life history.
So stars begin their lives on the main sequence and spend the majority of their time there. The giant stars belong to a subsequent and shorter-lived part of a star’s evolution. As it turned out, stars do not move along the main sequence, but stay on it for many millions to billions of years.
Figure 10.4 Protostars Evolutionary tracks of protostars of various mass in the Hertzsprung-Russell, or H-R, diagram, ending with their arrival on the main sequence when stars have begun burning hydrogen in their core. High-mass stars, which have greater luminosity than low-mass stars, are found at higher points on the main sequence and take a shorter time to arrive there. The luminosity, L, is in units of the solar luminosity, L, the mass in units of the Sun’s mass, M, and the effective visible disk temperature is in units of degrees kelvin, designated K.
When all the hydrogen fuel has been used up in the core of a star, it can no longer support itself under the crush of gravity. It collapses inside, which increases the central temperature and density and opens up a new source of energy not previously available, while the surrounding stellar atmosphere expands to produce a red giant star.
The Way Stars Become Giants
It was the Estonian astronomer Ernst Öpik who in 1938 proposed how main-sequence stars might become giant stars.26 He suggested that both the giant and main sequence stars shine by thermonuclear processes that follow a well-defined sequence of increasing core temperature. As a star evolves, the nuclear fuels burn from the center outward, and successively new nuclear fuels begin to burn at the center. For the giant stars, the exhaustion of the first processes begins earlier and the central temperatures rise to open up a new source of energy not available to main-sequence stars.
The course of the star’s trajectory into the realm of the giants was mapped out in the H-R diagrams of globular star clusters, just as Strömgren had proposed, which brings us to Allan Sandage, who spent much of his life looking at the stars from the dark, cathedral-like domes of the largest telescopes.
Sandage was born in Iowa City, Iowa, on June 18, 1926, the only child of a Professor of Advertising and a homemaker mother. He began his undergraduate studies in 1943 at the Miami University in Oxford, Ohio, because his father was on the faculty there at the time, but two years later Allan was drafted into the Navy, where he spent 18 months training to be an electronics technician’s mate. When he was discharged from military service, his father was moving to the University of Illinois, so Allan transferred there to finish his undergraduate degree in physics and mathematics.
In September 1948, he joined the first class of students to begin formal graduate studies in astronomy at the California Institute of Technology, as a selfdescribed “hick who fell off the turnip truck.” Allan became an observing assistant for Edwin Hubble, but since Hubble had no specific tasks suitable for a doctoral thesis, the German astronomer Walter Baade, who was working at the Mount Wilson Observatory, became Sandage’s advisor and guided him in using the 60-inch (1.5-meter) telescope to detect faint stars in globular star clusters.
Sandage’s investigations of the star cluster M 3 showed that its main sequence does not contain the hot, luminous O and B stars (Fig. 10.5)27 This indicated that it has an age of at least 5 billion years, which was the time for these stars to consume all their core hydrogen fuel. That was roughly comparable to the geological and radioactive-dating estimates for the age of the Earth. [Later more precise estimates indicated an age of 11.4 billion years for M 3.]
When Hubble died in 1953, Sandage, a fresh Ph.D. of just 27 years old, inherited the task of using the 200-inch (5.08-meter) telescope on Palomar Mountain in California to measure the age and fate of the expanding Universe. After five years of careful measurements of the distances of galaxies, he obtained a new value for the rate of expansion of the Universe, which meant that the Universe is seven times older than was previously thought.28 His measurements indicated an age of about 13 billion years if the Universe has been expanding at a constant rate, which is close to today’s accepted value for the time since the Big Bang.
Sandage also continued to investigate the paths of stellar life by describing how the observed properties of both open and globular star clusters depend on their age. Stars within a star cluster are all of the same approximate age, within a few million years of each other, dating back to the formation of the cluster many billions of years ago. They also began with the same initial composition of material, and exhibit a full range of stellar mass. And because the stars in a given cluster are all at the same distance from the Earth, we can obtain direct observations of their relative luminosity without knowing the distance.
Figure 10.5 H-R diagram for M 3 The high mass stars in the globular star cluster M 3 have left the main sequence and evolved into the red giant branch (top right). Allan Sandage included this diagram in his 1953 doctoral thesis at the California Institute of Technology, entitled A Study of the Globular Cluster M 3. The vertical axis is the apparent photographic magnitude, mpv, and the horizontal axis is the color index, CI.
All of the stars in a cluster begin shining when they arrive on the main sequence of the H-R diagram. As time goes on, the more luminous and massive stars evolve into the next phase of stellar life and the main sequence disappears from the top down. Very massive stars at the upper left of the main sequence become supergiants, and those with intermediate masses comparable to the Sun become red giants. A cluster H-R diagram can therefore be used as a clock, dating the age of the cluster and the stars in it by the place of their turnoff from the main sequence to become supergiants or giants. The lower the luminosity and temperature of the turnoff point, the older the star cluster is.
Sandage’s investigations of stellar evolution in star clusters also involved theoretical calculations of just how long a main-sequence star’s central fuel supply can last, and models of what happens when that fuel is used up. Martin Schwarzschild, the son of the German astronomer Karl Schwarzschild, was one of the first to examine these aspects of stellar evolution. After emigrating to the United States, Martin used theoretical models and primitive computers, developed by his Princeton University colleague John von Neumann, to chart the evolutionary trajectory of a star and compare it to the various kinks, bends, and gaps of missing stars in the observed H-R diagrams of globular star clusters.
Martin teamed up with Allan Sandage, whose observations included faint stars that connected the main sequence to the red giants.29 They found that when the internal energy source is depleted in a main-sequence star of roughly a solar mass, the shrinking stellar core heats up and causes the star as a whole to swell up into a bloated red giant (Fig. 10.6). The internal energy released within the star is then spread over a much larger area, resulting in a lower disk temperature and a shift of the visible starlight into the red part of the spectrum. This accounts for the red giant branch in the H-R diagram of globular star clusters, and also permits determination of their ages that are between 10 billion and 14 billion years old (Fig. 10.7).
Figure 10.6 Formation of a giant star When a main-sequence star consumes the hydrogen in its core, the inside of the star contracts and heats up, while the outside expands and cools down. The center of this giant star eventually becomes hot enough to burn helium and stop the core collapse.
Figure 10.7 How old is a globular star cluster? These observations of the stars in the southern globular star cluster 47 Tucanae, also designated NGC 104, indicate it has an age of between 12 and 14 billion years. (Courtesy of James E. Hesser.)
Allan Sandage also contemplated the mysterious reasons for life. By middle age, he had become plagued by two questions: “What is the purpose of life?” and “Why do we exist?” He concluded that these questions could not be answered by science, and that they required belief in the supernatural. “There is a mystery out there,” he declared, “and it’s outside the realm of science. Science can only answer the question of how, when and where, and perhaps what, but not why. The question of why is outside the scientific purview; but that’s still part of the whole picture.”30 In this distinction between the how and the why, he was agreeing with Henry Norris Russell.
Sandage eventually became deeply spiritual in his outlook on life, the Universe, and the practice of astronomy, writing: “If there is no God, nothing makes sense.... And if there is a God, he must be true both to science and religion.” The Big Bang, for example, might account for how the expanding Universe began, but “knowledge of the creation is not knowledge of the Creator, nor do any astronomical findings tell us why the event occurred. It is truly supernatural and by this definition a miracle.”31
In order to explain the biggest why, and understand the mystery of existence, Allan Sandage became devoutly religious and appealed to something bigger, outside and beyond both observational astronomy and his own existence.