“Impersonal monsters, namely, Immensities. Until a person has thought out the stars and their inter spaces, he has hardly learnt that there are things much more terrible than monsters of shape, namely, monsters of magnitude without known shape. Such monsters are the voids and waste places of the sky.”
—Thomas Hardy, Two on a Tower
For years after Henrietta Swan Leavitt’s death in 1921, her presence lived on, not just in her discovery about Cepheid variables but in a ghost story that circulated around Observatory Hill. Cecilia Payne, the young Harvard astronomer (and later department chairman) who had inherited Henrietta’s old desk, was amused to hear rumors “that Miss Leavitt’s lamp was still to be seen burning in the night, that her spirit still haunted the plate stacks.” More likely, she concluded, someone had seen Payne herself as she burned the midnight oil, sometimes working on theories about variable stars.
There is something almost ghostlike in the scant traces Leavitt left in the public record—biographical ectoplasm to be shaped according to one’s need. After years of so little recognition, there has followed an almost reflexive rush to mythologize her. A planetarium has been named after her, albeit a virtual one residing only on the World Wide Web. While Harlow Shapley had an entire cluster of galaxies named in his honor, Leavitt (and Annie Cannon) got a crater on the moon.
In brief hagiographies scattered across the Web, the same few scraps of data about Leavitt’s life, and often the same sentences, are repeated again and again, all traceable to a few stable sources. She has been turned into a standard bearer by people less interested in her astronomy than in the fact that she was a woman, and deaf. She has been included, absurdly, in a roster of “The World’s Greatest Creation Scientists,” for no apparent reason other than that she believed in God. She probably would be appalled.
Among the factoids that ricochet through the infosphere is that she was nominated for a Nobel Prize. What happened is that in 1925 Gösta Mittag-Leffler, an elderly Swedish mathematician, heard something about Leavitt’s work from a colleague and was impressed enough to write her a letter. He didn’t know that she had died.
“Honoured Miss Leavitt,” he began. “What my friend and colleague Professor von Zeipel of Uppsala has told me about your admirable discovery of the empirical law touching the connection between magnitude and period-length for the S. Cephei-variables of the Little Magellan’s cloud, has impressed me so deeply that I feel seriously inclined to nominate you to the Nobel prize in physics for 1926, although I must confess that my knowledge of the matter is as yet rather incomplete.” His expertise was in analytic function theory, not astronomy.
He asked for more information and vowed to handle the matter “with the greatest discretion and in the way that seems to me most likely to further my plan.” He also promised to send, for her perusal, a treatise he had written on Sonja Kowalewsky, a stunning young Russian mathematician, and her correspondence with Karl Weierstrass, the older mentor who had furthered her career. Maybe Mittag-Leffler was hoping Henrietta would become his Sonja.
When the letter arrived at the observatory, it was forwarded to the director, Harlow Shapley. It is hard to know quite what to make of his reply:
“Miss Leavitt’s work on the variable stars in the Magellanic Clouds, which led to the discovery of the relation between period and apparent magnitude, has afforded us a very powerful tool in measuring great stellar distances.”
Led to the discovery of the relation between period and luminosity? His phrasing suggests that he would deny her credit for her one breakthrough, relegating her back to the role of human computer, the diligent manipulator of data.
The next sentence continues in this vein—faint praise weighed with subtle condescension:
“To me personally [the discovery] has also been of highest service, for it was my privilege to interpret the observation by Miss Leavitt, place it on a basis of absolute brightness, and extending it to the variables of the globular clusters, use it in my measures of the Milky Way.”
It is clear whom Shapley would like to nominate for the prize.
UNTIL THE END, Leavitt’s title continued to be “assistant.” Although Solon Bailey, in his history of Harvard College Observatory, does give her credit for the period-luminosity relationship, he describes the work only in passing and notes, a little belittlingly, “The number of variables included in Miss Leavitt’s discussion was unfortunately rather small, but the data have been much increased since that time, especially by the studies of Shapley.”
Leavitt probably would have been surprised by how much fuss would later be made over her delightfully simple observation, and by how far Shapley and then Hubble were able to go with it. Given the opportunity—better health, better times— maybe she would have joined them at the forefront. Or maybe not. Barring the discovery of a lost cache of letters, we may never know.
In time, someone else would have discovered Henrietta’s law. It is the discovery not the discoverer that matters. Miss Leavitt may have understood this in a way that a Shapley or a Hubble never could. She seemed content to be a small part of a greater thing called science.
In January 1920, the year before she died, a census taker encountered her for the last time in the apartment with her mother on Linnean Street. Among their neighbors were teachers, a salesman for a candy company, a bank clerk, an auditor. Asked to state her occupation, Miss Leavitt replied, honestly and perhaps a bit defiantly, “Astronomer.”
2
On a spring afternoon in 1996, seventy-six years to the day since the Great Debate, astronomers gathered at the same lecture hall in Washington for a presentation called, once again, “The Scale of the Universe.” The cosmos was seventy-six years older and, if you believed the big bang theory, seventy-six light-years larger in every direction than it had been in 1920.
It was a real debate this time, with opening arguments, rebuttals, and closing statements presented by two celebrated astronomers, Gustav A. Tammann and Sidney van den Bergh. While Tammann argued for a Hubble constant of around 55, van den Bergh put it at about 80. Plugged into the equations of universal expansion, that narrow difference would translate into a universe ranging, depending upon other factors, somewhere between 10 billion and 15 billion light-years in radius.
Large as it seemed, the spread had narrowed in recent years. The lowest values for the constant, around 50, continued to be championed by Allan Sandage, who had taken over where Hubble left off, an opportunity and a burden he once compared to having been Dante’s assistant and inheriting The Divine Comedy. (The legacy included the incomplete—barely begun, in fact—Hubble Atlas of Galaxies.) Then just as the number seemed set in stone, or at least wet cement, an astronomer named Gerard de Vaucouleurs came along and doubled it back to 100, cutting the size of the universe in half. The ensuing controversy became known as the “Hubble Wars.”
As Sandage’s collaborator and protégé, Tammann was an ideal person to carry on the fight for an older, larger universe, and van den Bergh showed himself to be a formidable opponent. As the debate unfolded, each man in turn challenged the other’s choice of standard candles and the manner in which he interpreted them. Watching from the auditorium, members of the audience wore colorful buttons, “Hubble Meters,” on which they displayed their own guesses about the constant’s value. One might have come away from the debate with the impression that the size of this most fundamental parameter was a matter of opinion.
For all the constant’s inconstancy, Harlow Shapley and Heber Curtis both would have been impressed by how far the craft of intergalactic measuring had come. Edward Pickering and Henrietta Leavitt would have been astonished. What is now the world’s largest telescope, the Keck, perched atop the 13,800-foot Mauna Kea volcano in Hawaii, collects light with a mirror 10 meters wide. That’s 394 inches, or almost twice the size of the 200-inch telescope at Mount Palomar, which was twice the size of the one that Hubble had used to show that Andromeda is really a galaxy. Just when it seemed that mirrors had become as large as physically possible, on the verge of buckling under their own weight, computer and robotics technology stretched the limits further. The mirror on the Keck telescope was made from thirty-six hexagonal sections nudged back and forth by precision pistons so that the whole thing acts like one enormous reflector. Its shape can be constantly adjusted, a few nanometers at a time, to compensate for atmospheric distortion, a technique called adaptive optics. The glass molds itself to the sky.
In fact, by the time of the second debate, there were two Keck telescopes sitting side by side on the mountaintop soon to be linked by an optical interferometer, a computerized device that would combine light from both mirrors, along with that from several smaller scopes, into a single image. The result would be as powerful as a telescope with a mirror 85 meters, or 3,346 inches, across.
For even more acute observations, the Hubble Space Telescope, launched in 1990, was orbiting 380 miles above the Earth, electronically beaming back pictures of deepest space. Among its tasks was looking for Cepheids.
According to the picture that has emerged from these investigations, Andromeda, two million light-years away, now appears to be twice the size of the Milky Way. Both these nebulae mark the far edges of the constellation of galaxies called the Local Group, which also includes Triangulum, the Magellanic Clouds, and several dozen dwarf galaxies.
Nearby are other groups called Sculptor, Maffei, Canes I, Canes II, Dorado … so many (more than 150) that most are just given numbers. In addition to the groups are larger “clusters” like Fornax, with 49 galaxies, and Eridanus with 34. Largest of all in this tiny corner of the universe is the Virgo Cluster, which includes another 200 galaxies. Put all these together and you have the Virgo Supercluster—a galaxy of galaxies, numbering in the thousands, spanning 200 million light-years. One of them is the Milky Way.
Farther beyond are the neighboring superclusters: Coma, Centaurus, Hydra, Pavo-Indus, Capricornis, Horologium, Shapley, Sextans—80 of them within a billion light-years. As might be expected, our own Virgo Supercluster turns out to be on the smallish side.
Altogether there are believed to be tens of millions of galaxies just within a billion light-year radius of our solar system— more galaxies than there once were stars.
For all the technological progress astronomy has made, the basic approach to measurement has remained fundamentally the same: use redshift to gauge the recessional velocity of a galaxy or a galactic cluster, then divide that number by the Hubble constant to get the distance.
To calibrate the Hubble scale, Cepheids are still the standard of choice, when you can find them. The Hubble Space Telescope has been spotting them in distant clusters that once remained beyond reach. By the end of its mission in 1993, the European Space Agency’s High-Precision Parallax Collecting Satellite, or Hipparcos, had measured parallactic shifts as tiny as one milli-arc-second—1/1,000 of one second of one degree. The thousands of stars whose distances it gauged included a number of Cepheids.
It would be comforting to report that Hipparcos had been able to directly measure the trigonometric parallax of at least one Cepheid in as straightforward a manner as Hipparchus himself had measured the moon. The whole celestial distance scale, the ladders piled on top of ladders, would stand on firmer ground. But the nearest Cepheids are still too distant for even the orbiting satellite to fix any one of them very accurately. A statistical analysis of the better measurements suggested to some astronomers that the whole cosmic distance scale might have to be corrected by 10 percent.
The interpretation is open to dispute. It is probably inevitable that the more astronomers study Miss Leavitt’s stars, the less simple they appear. The pulsations of some, including Polaris, include subtle “overtones,” secondary rhythms that can throw off the beat. Debates periodically erupt over whether Cepheids of various colors and chemical content have significantly different period-luminosity curves.
In tweaking the Hubble constant, astronomers also now rely on other kinds of pulsating stars, like the RR Lyraes (Shapley’s old “cluster” variables) and the Miras. In addition are a wide assortment of secondary candles, those of lesser reliability that are calibrated using Cepheids for measuring beyond where it is possible to pick out individual stars.
During the 1920 debate, one of the arguments against the existence of island universes had been the extreme brightness of a certain nova in Andromeda. Unless the nebula was nearby, the nova would have to be unusually powerful, way off the scale. By the time of the 1996 debate, astronomers spoke comfortably of “supernovae,” intense bursts of light coming from exploding stars. A variety of supernovae called Type Ia has been calibrated for use as standard candles. Because of their extreme brightness they have been detected billions of light-years away.
When whole galaxies must be used as standard candles, astronomers can draw on something called the Tully-Fisher method: the larger a spiral galaxy, the faster it will spin. Bigger galaxies are also brighter, so their intrinsic luminosity can be estimated from their rotational speed, which is gauged by measuring Doppler shifts.
The details of this and other methods can quickly become esoteric, but the underlying idea is the same: if you can devise a theory that relates an observable characteristic—of a star, a supernova, a galaxy, a cluster—to its inherent brightness, then you can use it as a standard candle.
The measurements remain fraught with uncertainty. In addition to the universe’s outward expansion, galactic clusters are also pulled gravitationally toward one another. These “peculiar motions” result in kinks in the Hubble expansion that must be corrected for. Our Local Group is believed to be gradually falling into the massive Virgo Cluster, a phenomenon called virgocentric flow.
Astronomers must also guard against selection effects, giving too much weight in their calculations to the stars, galaxies, and clusters that are easiest to see. The most well-known of these is the so-called Malmquist bias: The stars you can pick out in a cluster are necessarily the brightest. If you rely on them to compute the average luminosity, the answer will be skewed.
Even with so many ways to go wrong, astronomers have been moving toward a consensus that the universe is a little less than 14 billion years old. Looking out from any vantage point, an observer will be at the center of a bubble extending that many light-years in every direction. It is still a hard idea to get used to. No one is at the center yet everyone is. Wherever you stand, you can see no farther than light has been able to travel since the big bang, the explosion that occurred everywhere and nowhere, that created space and time.