The spectrum of the average spiral nebula is indistinguishable from that given by a star cluster. It is such a spectrum as would be expected from a vast congeries of stars.
—Heber Curtis
On a spring day in 1920, Shapley found himself strolling along the train tracks somewhere in Alabama talking about flowers and classics and probably looking for ants. His companion was Heber Curtis, and they had agreed for now to avoid the touchy subject of astronomy. Each, unbeknownst to the other, had booked passage on the same train from California to Washington, D.C., where, in the halls of the National Academy of Sciences, they were scheduled to debate whether there was anything in the universe beyond the Milky Way.
The event had been arranged at the prompting of Shapley’s boss, George Ellery Hale, one of the most revered astronomers of the time. Hale’s father had made a fortune selling hydraulic elevators to the builders of Chicago’s soaring skyscrapers. The son set his sights still higher, studying astronomy and exploiting his family’s financial connections to fund the development of some of the best telescopes in the world. When the old man died, a series of lectures was endowed in his name. The younger Hale thought that the one in 1920, at the National Academy’s annual meeting, should be devoted to a currently hot cosmological issue—either relativity or island universes.
Harlow Shapley
(Harvard University Archives)
The first topic struck the academy’s secretary as too esoteric. (He personally thought that Einstein’s theory should be banished “to some region of space beyond the fourth dimension, from whence it may never return to plague us.”) Nor was he keen on island universes, fearing that “unless the speakers took pains to make the subject very engaging the thing would fall flat.” He proposed instead a presentation on glaciers or “some zoological or biological subject.” In the end, Hale had the final word, and Shapley and Curtis were picked to present their opposing views on “The Scale of the Universe,” and, more pointedly, on whether it consisted of more than a single galaxy.
Were such an event to take place today, it would be videotaped and perhaps transcribed. You could probably download it from the Web. The encounter between Shapley and Curtis can be pieced together only from scraps of evidence—a typescript of Shapley’s talk, annotated with his scribbles, some of Curtis’s slides (he misplaced his script soon after the event), and letters the two exchanged before and after what came to be called the Great Debate, mostly by people who had not been there.
Curtis himself was eager for a scrap. He imagined the two astronomers going after each other with “hammer and tongs,” then shaking hands like gentlemen. Shapley, however, was worried that he might lose. Not that he thought his theory was wrong. But persuading an audience of geologists, biologists, and scientists of other nonastronomical persuasions required the skills of an orator. Right or wrong, Curtis, thirteen years older and the more polished debater, might best Shapley at the podium.
He expected for example that Curtis would pick on the tiny handful of stars (“my eleven miserable Cepheids,” Shapley called them in a letter) from which he had extrapolated such enormous distances. What seemed to some like a brilliant analysis might strike others as a house of cards. Many astronomers were much less confident than Shapley about the usefulness of Henrietta Leavitt’s yardstick. Curtis had made it clear that he thought Shapley’s Milky Way was ten times too large. If he could shrink it back by an order of magnitude, the island universe theory might be easier to uphold.
Shapley also had an ulterior motive. He was certain he was being considered for the directorship of Harvard Observatory—Edward Pickering had just died—and he expected an emissary from Observatory Hill to be in the audience. He dreaded making a bad impression.
For weeks Shapley worked to bolster his evidence while maneuvering for a more advantageous position. Through sheer obstinacy, he managed to get the debate downgraded to a discussion—“two talks on the same subject from our different standpoints”—and the talks themselves reduced in length. While Curtis wanted forty-five minutes for each presentation, Shapley wanted thirty-five. They split the difference, settling on forty. To blunt the impact further, no time would be allowed for rebuttals, just a general discussion at the end. Finally Shapley made certain that Henry Norris Russell, his former teacher and ally, would be in the audience to support his position. He wasn’t taking any chances.
2
The evening began at an excruciating pace with awards followed by long testimonials. There was a tribute to the Prince of Monaco for oceanography, Shapley later remembered, and another to some “noble human antique,” honored for combating hookworm. Many years later in a memoir, Shapley recalled a bored Albert Einstein sitting in the audience, whispering to his companion that he now had a new theory of eternity. It made a good story, but actually Einstein was in Germany then, fending off the first stirrings of Nazi denunciations of his “Jewish physics.” He made his first visit to the United States the following year.
Once the main event finally began, Shapley went first, easing in slowly with a long introductory tutorial on astronomy. A third of the way through his allotted time, he had gotten only as far as the definition of a light-year. So far the presentation was pure popular science. One can imagine Curtis glancing at his watch, wondering when Shapley would say something he could dispute.
Then he surprised Curtis again with a promise to spare the audience “the dreary technicalities of the methods of determining the distance of globular clusters,” the way stations he had used to map the galaxy. Maybe that was a sensible approach for an audience of nonspecialists. But Curtis, who was still itching for a fight, had prepared a meticulous pointby-point deconstruction of Shapley’s every assumption and logical inference. There was still nothing for him to rebut.
The biggest surprise was still to come. Skipping over Cepheids entirely, Shapley described an entirely independent method of establishing the enormity of the galaxy (and by implication, undermining the case for island universes).
Astronomers had uncovered what appeared to be a relationship between a star’s temperature and its inherent brightness. (These had been plotted on a chart famous to astronomers as the Hertzsprung-Russell diagram after Henry Norris Russell and Ejnar Hertzsprung.) The result was another kind of measuring stick. A survey of nearby “B-type stars,” identified by their bluish sheen, had found them to be on average 200 times brighter than the sun. This suggested, to Shapley anyway, that these blue giants could be used as standard candles. No matter how much a B star’s light is dimmed on the journey to Earth, it could be assumed, with a small leap of faith, to have the same intrinsic brightness as its cousins closer to home. And so the giant’s distance could be calculated from the inverse square law. If it was nine times dimmer than a nearby B star, it must be three times farther away.
Shapley found what he took to be blue giants in the Milky Way cluster called Hercules. Exceedingly dim—of the fifteenth magnitude—the stars, he reckoned, would have to be 35,000 light-years away. Then he extrapolated further. Assuming that the smaller, dimmer clusters had the same overall brightness as Hercules, he circled around to his original conclusion: that they lie at the fringes of a galaxy 300,000 light-years in diameter, with the sun shoved off to one side.
He briefly mentioned how another kind of yardstick, stars called red giants, also supported his measurements. Then he tried to fend off any criticism of his “miserable” Cepheids by removing them from the debate: Professor Curtis, he said, “may question the sufficiency of the data or the accuracy of the methods.…But this fact remains: we could discard the Cepheids altogether, use instead the thousands of B-type stars upon which the most capable stellar astronomers have worked for years, and derive just the same distance for the Hercules cluster, and for the other clusters, and obtain consequently the same dimensions for the galactic system.”
Before the debate the red and blue stars had been no more than footnotes to Shapley’s argument—secondary checks on his primary measuring tool, Henrietta Leavitt’s Cepheid variables. Now figure and ground had been reversed, leaving Curtis to aim at a moving target and leading one to wonder who really was the wilier debater.
As for the nature of the spiral nebulae, the original focus of the discussion, Shapley dismissed them in a few sentences:
I shall leave the description and discussion of this debatable question to Professor Curtis. We agree, I believe, that if the galactic system is as large as I maintain, the spiral nebulae can hardly be comparable galactic systems; if it is but one-tenth as large, there might be a good opportunity for the hypothesis that our galactic system is a spiral nebula, comparable in size with the other spiral nebulae, all of which would then be “island” universes of stars. On one other point I think we also agree, or at least we should agree, and that is that we know relatively so little concerning the spiral nebulae … that it is professionally and scientifically unwise to take any very positive view in the matter just now.
Even if the spiral nebulae were not firmly inside the boundaries of the Milky Way, he believed, they probably lay on the outskirts, small gas clouds encountered during the galaxy’s drift through endless nothingness.
3
There is no way to re-create the tone of Curtis’s presentation from the skeletal talking points left behind on his typewritten slides. It is clear that, undeterred by Shapley, he marshaled a strong defense of island universes. As expected, he challenged the reliability of Shapley’s calibration of the Cepheid yardstick, joining those who suspected that the more rapidly oscillating cluster variables used to measure out the Milky Way were different in nature from the slower ones Leavitt had found in the Magellanic Clouds. Why assume they had precisely the same relationship between period and inherent brightness, if there was indeed any such relationship at all? If Shapley’s variables were actually much dimmer to begin with, then all the clusters would lie closer in. The perimeter of the galaxy would contract and the Milky Way would shrink from continent to island, one member of a great archipelago.
Herber Curtis
(Lick Observatory)
Curtis was equally unimpressed with the giant blue stars, arguing that far too little was known to trust them as standard candles. He proposed what he considered a more reliable measuring device—the yellow-white stars like the sun that seemed to make up most of the galaxy. It was reasonable, Curtis proposed, that the sunlike stars in the far reaches of the Milky Way shine, on average, with the same brightness as those nearby. Like Shapley, he was assuming the uniformity of nature, and his conclusion was that the galaxy can be only about 30,000 light-years across. For the clusters to be as distant as Shapley believed, these stars would have to be far brighter than those in our own neighborhood. A different physics would prevail. “While it is not impossible that the clusters are exceptional regions of space [with] a unique concentration of giant stars, the hypothesis that cluster stars are, on the whole, like those of known distance seems inherently the more probable.”
With the Milky Way knocked down in size, evidence for island universes seemed compelling. Curtis recalled the familiar argument that the color patterns produced by the spirals were indistinguishable from that of starlight. “It is such a spectrum as would be expected from a vast congeries of stars.” And the novae that appeared within them “seem a natural consequence of their nature as galaxies, incubators of new stars.” Used as standard candles the novae put Andromeda half a million light-years from Earth and other spirals 10 million or more light-years away. “At such distances, these island universes would be of the order of size of our own Galaxy of stars.”
Finally he noted a curious phenomenon that had been puzzling astronomers for years: the spiral nebulae appeared to be concentrated at the two “poles” of the Milky Way—the regions directly above and below the galaxy’s central bulge. None was found in the galactic plane where most of the stars reside. If the spirals were small clouds within or near our galaxy, then why were they not evenly distributed? It was as though they were being repelled by some mysterious force.
It was far more plausible, Curtis argued, that this “zone of avoidance” was an illusion—that the spirals lay far beyond the Milky Way, in every direction, with those along the galactic plane hidden from our view. Many spirals appeared to be surrounded by a thick ring of “occulting” matter, a halo of interstellar dust. The same might be true of the Milky Way. When astronomers aimed their telescopes directly into this dust storm, spirals in that direction were blocked from view. Of the millions of spirals, the only ones we could see were those that happened to lie above and below. Each, he proposed, was a world as vast and shining as our own.
4
Each man left the lecture hall certain that he had won. “Debate went off fine in Washington,” Curtis wrote to his family, “and I have been assured that I came out considerably in front.” Shapley, for his part, attributed any points Curtis may have scored to his rhetorical skills. “Now I would know how to dodge things a little better,” he said years later, a comment that seems strange since Shapley spoke first. Maybe he was referring to the discussion that followed the talks, during which his mentor Russell had, as planned, come forth with a strong endorsement of Shapley’s Big Galaxy theory. The champions of island universes surely responded just as vigorously. “Curtis did a moderately good job,” Shapley recalled. “Some of his science was wrong, but his delivery was all right.”
Both men fleshed out their arguments in papers published the following year in the Bulletin of the National Research Council. (In some early historical accounts these published papers are treated as the actual substance of the debate.) The texts contain no fundamentally new arguments. Shapley bolstered his case with more data, whose certainty Curtis continued to question. What is most striking is how two of the world’s smartest astronomers could take the same trove of astronomical observations and come up with two such very different pictures of the universe, a reminder that science lies not in the facts themselves but in their arrangement.
For Curtis, the zone of avoidance (it sounds like something from a Superman comic) was strong evidence for seeing the spirals as island galaxies. In Shapley’s hands, the phenomenon seemed to support the argument that spirals are little wisps of stellar gas: they would have to be small and light to be repulsed somehow by the Milky Way. Also open to conflicting interpretations were the novae that appeared now and then inside the spirals. For Curtis their existence showed that spirals were indeed galaxies. Where else would one expect to observe stars being born? For Shapley each nova represented “the engulfing of a star by [a] rapidly moving nebulosity.”
The most memorable passage in either paper is a paragraph by Shapley on the perils of extrapolating too boldly from a limited set of data. He meant the statement as a criticism of Curtis’s measurements involving the average magnitude of sunlike stars. But it could be taken just as easily as a humbling reminder about how much the entire enterprise of astronomical measurement rests on a few vulnerable assumptions. (It is also the inspiration for the story about the villagers in the canyon, which appears in the prologue of this book.)
“Suppose,” Shapley began, “that an observer, confined to a small area in a valley, attempts to measure the distances of surrounding mountain peaks.” He can use parallax for the nearby hills, but since he cannot leave his narrow valley, his baseline is too small to triangulate any farther. He needs another kind of measuring stick. Seeing through his telescope that there is plant life on the mountaintop, he makes the simplifying assumption that it is approximately the same as the plant life on the valley floor—averaging about a foot in height. Thus from the apparent size of the foliage, he can judge how far the mountain is.
His calculation would be wrong. “If, however,” Shapley noted, “he had compared the foliage on the nearby, trigonometrically-measured hills with that on the remote peaks, or had used some method of distinguishing various floral types, he would not have mistaken pines for asters and obtained erroneous results for the distances of the surrounding mountains. All the principles involved in the botanical parallax of a mountain peak have their analogues in the photometric parallax of a globular cluster.”
Mistaking pines for asters, and asters for pines. It was an occupational hazard that would plague Shapley as much as anyone.