Kirchhoff was mad enough to try, and by 1861 he had identified sodium, calcium, magnesium, iron, chromium, nickel, barium, copper, and zinc in the sun. A link had been found between the physics of earth and the stars, and a path blazed to the new sciences of spectroscopy and astrophysics.
In London, a wealthy amateur astronomer named William Huggins learned of Kirchhoff’s and Bunsen’s finding that Fraunhofer lines were generated by known chemical elements in the sun, and saw at once that their methods might be applied to the stars and nebulae. “This news came to me like the coming upon a spring of water in a dry and thirsty land,” he wrote.5 Huggins fitted a spectroscope to the Clark telescope at his private observatory, on Upper Tulse Hill in London. By carefully studying each spectrum until he could make sense of their many overlapping lines, he succeeded in identifying iron, sodium, calcium, magnesium, and bismuth in the spectra of the bright stars Aldebaran and Betelgeuse. This was the first conclusive evidence that other stars are made of the same substances that we find here in the solar system.
With mounting excitement, Huggins turned his telescope to a nebula. His journal for the year 1864 records the feeling “of excited suspense, mingled with a degree of awe, with which, after a few moments hesitation, I put my eye to the spectroscope. Was I not about to look into a secret place of creation?” He was not disappointed:
I looked into the spectroscope. No spectrum such as I expected! A single bright line only! … The riddle of the nebulae was solved. The answer, which had come to us in the light itself, read: Not an aggregation of stars, but a luminous gas. Stars after the order of our own sun, and of the brighter stars, would give a different spectrum; the light of this nebula had clearly been emitted by a luminous gas.6
Because this first nebula Huggins observed with his spectroscope happened to be gaseous, he was led to the erroneous conclusion that all nebulae, the ellipticals and spirals included, were gaseous and that none was composed of stars.
But life is seldom simple, and misleading evidence for the nebular hypothesis continued to accumulate. The positions of hundreds of spiral nebulae were charted, and they were found to be most numerous in the parts of the sky that lie well away from the Milky Way—to “avoid” the Milky Way, in astronomical jargon. The avoidance effect suggested that the spiral nebulae were associated with our galaxy. (Actually, avoidance results from the fact that dark clouds along the plane of our galaxy obscure our view of the other galaxies, so that we see mostly those that lie away from the galactic plane.) The nebular hypothesis was strengthened on the theoretical front as well, when the astrophysicist James Jeans demonstrated, with considerable mathematical rigor, that a collapsing cloud of gas would tend to assume a disk shape much like that of the spiral nebulae. Jeans even managed to coax his model into generating spiral arms like those seen in the astrophotographs.
By now the nebular hypothesis was so successful that a bandwagon syndrome took over and astronomers began seeing what they thought they ought to see. One announced that he had measured the parallax of the Andromeda spiral. (Parallax is detectable only out to a few hundred light-years; the Andromeda galaxy is over two million light-years away.) Another found that by examining older photographs he could detect signs of circular motion in spiral nebulae. (In reality, galaxies are so large that to see a galaxy turn by as much as the second hand on a clock moves in one second would require taking two photographs separated by an interval of fully five million years.)
As the twentieth century began, then, several of the most stupefying aspects of the closed, pre-Copernican cosmology had been resurrected on a galactic scale. The sun was widely thought to be located at or near the center of a stellar system—the Milky Way—which embraced every star and nebula in the telescopic sky, and which, therefore, constituted nothing less than the entire observable universe. Beyond our galaxy might lie an infinite void, but this question remained as purely academic as had been the nature of space beyond the outer sphere of stars in Aristotle’s model.
But there is a self-correcting mechanism to science, and by the turn of the century it had begun to assert itself. The first cracks in the facade of the nebular hypothesis appeared on the theoretical side, when a fatal defect was identified in the Jeans theory of how the solar system had condensed. Were the hypothesis correct, the mathematicians calculated, the sun should have retained most of the angular momentum of the solar system, and be spinning very rapidly; instead, the solar “day” lasts a leisurely twenty-six days at the sun’s equator, and the planets harbor 98 percent of the angular momentum of the solar system.* The observational evidence began to turn against the nebular hypothesis as well. Huggins took a spectrum of the Andromeda nebula in 1888 but found it hard to interpret. Nine years later, Julius Scheiner in Germany published a spectrum of the Andromeda nebula, noting that the spectrum was not gaseous but starlike. Undoubtedly, at least some spiral nebulae were made of stars.
Exploding stars then came to the astronomers’ aid, as they had centuries earlier for Tycho, Kepler, and Galileo. Two or three supergiant stars explode in an average major galaxy every century, with such brilliance that they can be seen across the reaches of intergalactic space. Since thousands of galaxies (or elliptical and spiral nebulae, as they were then being called) lay within the reach of existing telescopes and cameras, it was only a matter of time before supernovae began to be detected in photographs of other galaxies. The first such extragalactic supernova to be noticed, in Andromeda in 1885, happened to be near the center of the spiral, and so could be explained away as the sputtering of a Laplacian protosun. But then, in 1917, George Ritchey, an optician at Mount Wilson, and Heber Curtis, an astronomer at Lick, announced that they had found several novae in old file photographs of spirals. Other astronomers started ransacking their plate files, and found scores more. The novae were not central, but occurred primarily in the spinal arms. This was extremely damaging to the notion that all nebulae were gaseous: Dozens of exploding stars in galaxies full of stars made sense; in Laplacian gas disks, they did not. As Curtis commented, “The novae in spirals furnish weighty evidence in favor of the well known ‘island universe’ theory.”7
The stage was set for the discovery of galaxies. What remained was the most expansive surveying project in the history of our planet—to chart the location of the solar system in the Milky Way, and to determine the distances of the other galaxies, if such they were, beyond.
The champion of this cause was the founder of observational astrophysics, George Ellery Hale. Hale’s early career reenacted the progression of spectroscopy from the sun to the stars. He became enchanted by the sun as a boy growing up in the Chicago suburbs, built a backyard observatory where he observed solar spectra, and by the age of twenty-four had invented the spectrohelioscope, a device that made it possible to examine the solar atmosphere in one wavelength of light at a time. Captivated by the realization that, as he kept repeating all his life, “the sun is a star,” he then turned his attention to the depths of space. He was responsible for building four telescopes, each in its day the world’s largest—the 40-inch refractor at Yerkes Observatory in Wisconsin and, in southern California, the 60- and 100-inch reflectors at Mount Wilson and the 200-inch reflector at Palomar. Mount Wilson in particular stood as a monument to Hale’s dual passions in spectroscopy: There, solar telescopes recorded the spectra of the sun by day, while by night giant reflecting telescopes were employed to probe the multitude of other suns scattered through the Milky Way and beyond.
Hardworking even by the hard-boiled standards of the opticians and astronomers of the day, Hale rode mules up the rocky, twisting road from Pasadena to Mount Wilson’s peak, and when no mules were available simply ran up the side of the mountain. He did a lifetime’s worth of research of his own and managed simultaneously to act as the observatory director, raising funds for ever larger telescopes and recruiting some of the world’s leading astronomers to Mount Wilson. One of the cleverest of his recruits was Harlow Shapley.
Shapley h
ad studied at Princeton Observatory under Henry Norris Russell, where he specialized in Eclipsing Binaries. These are double stars, so close together in the sky that they look like single stars even through the most powerful telescopes, that happen to be oriented in space in such a way that they periodically eclipse each other. The resulting variations in the total brightness of the system bear a superficial resemblance to genuine variable stars, which change their brightness owing to internal pulsations. In this fashion Shapley came to study variable stars as well. The knowledge he gained in this somewhat backhanded way came in handy, for a class of variable stars—the Cepheid variables—were to provide astronomy with a means of measuring distances across interstellar space and even intergalactic. Thanks to the Cepheids, Shapley was to earn a place in history as the first human being to establish the location of the sun in the Milky Way galaxy.
Cepheids—as Shapley was the first to propose—pulsate, varying in brightness as they change in size. Astrophysically speaking, they are giant stars, three or more times the mass of the sun, undergoing a period of instability that occurs when they run low on hydrogen fuel and begin burning helium. The wonderful thing about them is that the period of each Cepheid—i.e., the time it takes to go through a cycle of variation in brightness—is directly related to its intrinsic brightness (i.e., its absolute magnitude). Once the absolute magnitude of any star is known, it is a simple matter to compute its distance: All the astronomer has to do is measure its apparent magnitude and then apply the formula that brightness decreases by the square of the distance. If, for instance, we have two Cepheid variables with the same period, we may assume that they have about the same absolute magnitude. If the apparent magnitude of one is four times that of the other, we conclude (barring complications such as the interference of an intervening interstellar cloud) that the dimmer star is twice as far away.
The relationship between the periodicity and the absolute magnitude of Cepheid variable stars was discovered in 1912 by Henrietta Swan Leavitt, one of a number of women hired at meager wages to work as “computers” in the Harvard College Observatory office in Cambridge, Massachusetts. Leavitt spent her days examining photographic plates taken through the twenty-four-inch refracting telescope at the Harvard station in Arequipa, Peru. One of her tasks was to identify variable stars. This involved comparing thousands of pinpoint star images on plates taken on different dates, looking for changes in brightness. It was painstaking toil, considered too menial to claim the time of a full-fledged astronomer. Leavitt spent thousands of hours at it, and in doing so acquired an unusual degree of familiarity with the southern sky.
She happened to be assigned to a region that includes the Magellanic Clouds. So named because they attracted the attention of Magellan and his crew on their voyage around the world, the Magellanic Clouds are two large, shaggy patches of softly glowing light that resemble detached swatches of the Milky Way. We know today, as Leavitt and her contemporaries did not, that the Clouds are nearby galaxies, and that the stars in each Cloud therefore all lie at about the same distance from us, like fireflies in a bottle viewed from across a field at night. This means that any significant difference in the apparent magnitudes of stars in a Magellanic Cloud must result from genuine differences in their absolute magnitudes and not from the effect of differing distances. Thanks to this happy circumstance, Leavitt in studying Cepheid variable stars in the Magellanic Clouds was able to notice a correlation between their brightness and their period of variability—the brighter the Cepheid, the longer its cycle of variation. The period-luminosity function Leavitt discovered was to become the cornerstone of measuring distance in the Milky Way and beyond.
Shapley, out to chart the Milky Way galaxy, seized on the Cepheids with great enthusiasm. Using the big sixty-inch Mount Wilson telescope, he photographed globular star clusters—spectacular assemblages of hundreds of thousands to millions of stars —identified Cepheid variable stars in each of them, then employed the Cepheids to calibrate the distances of the clusters. “The results are continual pleasure,” he wrote the astronomer Jacobus Kapteyn in 1917. “Give me time enough and I shall get something out of the problem yet.”8 The payoff came sooner than Shapley had hoped, and within a matter of months he could write, to the astrophysicist Arthur Stanley Eddington: “Now, with startling suddenness and definiteness, they [the globular clusters] seem to have elucidated the whole sidereal structure.”9
Shapley had found that the globular clusters are distributed across a spherical expanse of space, as if they were part of an enormous metaglobular cluster, and that the center of this sphere is nowhere near the sun, but lies far away to the south, past the stars of Sagittarius. In a superbly daring intuitive leap, Shapley then conjectured—accurately, as it turned out—that the center of the realm of the globular clusters was also the center of the galaxy itself. As Shapley put it, “The globular clusters are a sort of framework—a vague skeleton of the whole Galaxy—the … best indicators of its extent and orientation.” If so, the sun lies far from the center of things: “The solar system can no longer maintain a central position,” Shapley asserted.10
Shapley’s triumph was marred only by problems with his calculation of distances. The diameter of the Milky Way galaxy previously had been reckoned—by various investigators, Shapley among them—at some fifteen to twenty thousand light-years. Now, with his Cepheid variable work in hand, Shapley concluded that the correct figure was three hundred thousand light-years—more than ten times larger than the dimensions entertained by his contemporaries, and three times the most generous estimates we have today.*
Various errors contributed to Shapley’s inflated picture of the Milky Way galaxy. Like many of his contemporaries, he underestimated the extent to which clouds of interstellar gas and dust dim the images of distant stars, making them appear farther away than they really are. Moreover, he assumed that the Cepheid variable stars he observed in globular clusters were essentially identical to those Henrietta Leavitt had found in the Magellanic Clouds; actually, as Walter Baade and other astrophysicists were to find, the cluster variables are less massive and intrinsically less bright, and therefore by implication less distant, than a straightforward comparison of their periods with those of their younger cousins would imply. Inaccuracies of this sort are routine on the cutting edge of science, but they had the dolorous effect of misleading Shapley into thinking that the Milky Way, rather than being but one galaxy among many, was a system of unique grandeur. He began to think of the Milky Way as more or less the entire universe, and to regard the spiral nebulae but its subjects or its satellites.
For these and perhaps for subtler psychological reasons as well, Shapley came to take a proprietary interest in defending what he called “the enormous, all-comprehending” dimensions of the galaxy that he had charted. He called this view his “big galaxy” hypothesis. Those who agreed with him tended to think of the word “big” in terms of its Norse etymology, from bugge, meaning “important.” Those who disagreed preferred to emphasize the word’s Latin etymology, from buccae, for “puffed up.”
Among the dissidents was Heber Curtis of Lick Observatory, an advocate of the “island universe” theory. Shapley reacted to Curtis’s arguments with the abhorrence of a patient contemplating surgery: Curtis, he noted, “must shrink my galactic system enormously to have much luck with island universes.”11 The issue was formally debated by Shapley and Curtis under the auspices of the National Academy of Sciences, in Washington, D.C., on April 26, 1920. Shapley was generally judged the loser, but, as is usually the case in science, the debate settled little and the last word belonged not to men but to the sky.
The hypothesis defended by Curtis, that the spiral nebulae were galaxies of stars, would be confirmed if a spiral could be unambiguously resolved into stars. That vital step was accomplished in 1924 by Shapley’s colleague and nemesis, Edwin Hubble. A tall, elegant, and overbearing man with a highly evolved opinion of his potential place in history, Hubble made everything he did look
effortless—he had been a track star, a boxer, a Rhodes scholar, and an attorney before turning astronomer—and one of the things he did most effortlessly was to infuriate Shapley. Hubble took scores of photographs of M33 and its neighbor M31, the Andromeda spiral, and found there what he later called “dense swarms of images in no way differing from those of ordinary stars.”12
Whether the pinpoints of light on Hubble’s photographic plate really were stars, however, was open to contention; Shapley dismissed them as curds in a Laplacian nebula. Here, again, Henrietta Leavitt’s Cepheid variable stars provided the needed mileposts. Cepheids are bright enough to be discernible across intergalactic distances. Using the new one-hundred-inch telescope at Mount Wilson, Hubble photographed the spirals again and again, comparing the plates to find stars that had varied in brightness. His efforts soon bore fruit, and on February 19, 1924, he wrote Shapley, who by then had left Mount Wilson to become director of Harvard College Observatory, a laconic note containing one of the most momentous findings in the history of science: “You will be interested to hear that I have found a Cepheid variable in the Andromeda Nebula.”13
Hubble deduced that Andromeda lies about one million light-years away, an estimate half the distance of later ones but clearly sufficient to establish that the spiral was well beyond even Shapley’s “big galaxy.” Shapley replied sourly that he found Hubble’s letter to be “the most entertaining piece of literature I have seen for a long time.”14 Later he complained that Hubble had given him insufficient credit for his priority in using Cepheid variables to chart distances. But the game was over. Hubble’s paper announcing that he had found Cepheids in spirals—read (in his Olympian absence) at a joint meeting of the American Astronomical Society and the American Association for the Advancement of Science in Washington, D.C., on New Year’s Day, 1925—initiated the final decline of the nebular hypothesis, the ascendancy of the island universe hypothesis, and humankind’s realization that we live in one among many galaxies.
Coming of Age in the Milky Way Page 17