The Glass Universe

by Dava Sobel

  “You will not I hope be annoyed at my criticism,” Mrs. Draper added, “but I feel in publishing any of Dr. Draper’s work that I want his opinions represented as nearly as possible, now that he is not here to explain them himself.”

  The Drapers had met William and Margaret Huggins while visiting London in June 1879, at the Hugginses’ home observatory on Tulse Hill. Mrs. Draper recalled Mrs. Huggins as a petite woman with short, unruly hair that stuck straight out from her head as though galvanized. She was half the age of her husband, but a full participant in his studies, both at the telescope and in the laboratory.

  The two couples seemed destined to become either rivals or intimates. William gave Henry the benefit of his lengthier experience by offering helpful advice about spectroscope design. He also recommended a new type of dry, pretreated photographic plate that had lately come on the market. There was no need to paint liquid emulsion on these plates just prior to exposing them, and consequently they allowed for much longer exposure times. Before leaving England, the Drapers purchased a supply of Wratten & Wainwright’s London Ordinary Gelatin Dry Plates, which proved a boon indeed. They were particularly sensitive to the ultraviolet wavelengths of light, beyond the range of human vision. Unlike the old wet plates, the dry ones created a permanent record suitable for precision measurement. The dry plates gave the Drapers the wherewithal to photograph the spectra of the stars.

  • • •

  THE PAPER ANNOUNCING the stellar spectra findings, “by the late Henry Draper, M.D., LL.D.,” appeared in the Proceedings of the American Academy of Arts and Sciences in February 1884. Pickering mailed copies to prominent astronomers everywhere. By return mail dated March 12, he received William Huggins’s indignant reaction. Huggins found some of Pickering’s measurements “very wild,” the letter said with emphasis. “I should be glad if you could see your way to look into this, because it would be better that you should discover the error & publish the correction, than that the matter should be pointed out by others. . . . My wife unites in kind regards to you and Mrs. Pickering.”

  Pickering was certain he had not erred. And, as Huggins had never explicated his measurement procedures, Pickering stood firmly by his own. As they traded charges, Pickering forwarded Huggins’s letters to Mrs. Draper.

  Now it was her turn to grow indignant. “I felt very sorry,” she wrote Pickering on April 30, 1884, “that you should have been subjected to such an ungentlemanly attack, through your interest in Dr. Draper’s work.” Before returning the letters to Pickering, she took the liberty of copying one, since “it is worth preserving as a curiosity of epistolatory literature.”

  During this same time, Pickering was seeking assistants who might help Mrs. Draper advance her husband’s work to the next stage. He considered former director Joseph Winlock’s son, William Crawford Winlock, currently employed at the U.S. Naval Observatory, to be a very likely prospect, but Mrs. Draper rejected him. To her regret, she could not induce her preferred candidate, Thomas Mendenhall, to leave his professorship at Ohio State University. She channeled some of her frustration into the creation of the Henry Draper gold medal, to be awarded periodically by the National Academy of Sciences for outstanding achievements in astronomical physics. She gave the Academy $6,000 to endow the prize fund, and spent another $1,000 commissioning an artist in Paris to fashion a medal die featuring Henry’s likeness.

  The spring of 1884 brought Pickering new money worries. The successful five-year subscriptions from generous astronomy enthusiasts had run their course, ending the accustomed annual stipend of $5,000. The director was covering various operating expenses out of his own salary, and even so was forced to let go five assistants. In a touching show of solidarity, observatory colleagues took up a collection to retain one of those who had been dismissed, and furnished “part of the required sum,” Pickering told his circle of advisers, “from their own scanty means.” He appreciated the “extraordinary efforts on the part of the observers, who have performed without assistance the work in which they were previously aided by recorders. This has required an increase in the time spent in observation, and has rendered the work much more laborious. While this evidence of enthusiasm and devotion to science is most gratifying, it is obvious that it cannot long be continued without injury to health. Indeed, the effects of over-fatigue and exposure during the long, cold nights of last winter were manifest in more than one instance.”

  The motto on the Pickering family coat of arms, “Nil desperandum,” plus the lifelong habit of his own thirty-seven years, obliged the director to substitute resourcefulness and resilience for despair. He began formulating a means of combining Mrs. Draper’s wishes and wealth with the capabilities and needs of his observatory.

  “I am making plans for a somewhat extensive piece of work in stellar photography in which I hope that you may be interested,” he informed her in a letter of May 17, 1885.

  Pickering intended to redirect most of the observatory’s projects along photographic lines. His predecessors the Bonds had recognized the promise of photography, and achieved the first photograph of a star in 1850, but the limitations of the wet plates had impeded further attempts. With the new dry plates, possibilities multiplied. Determinations of stellar brightness and variability would surely prove easier and more accurate on photographs, which could be examined, reexamined, and compared at will. A methodical program for photographing the entire sky would transform the painstaking process of zone mapping. As a bonus, these photographs would reveal untold numbers of unknown faint stars, invisible even through the world’s biggest telescopes, because the sensitive plate, unlike the human eye, could gather light and aggregate images over time.

  Pickering’s younger brother, William, a recent graduate of MIT, was already teaching photographic technique there and testing the limits of the art by trying to photograph objects in motion. The twenty-seven-year-old William had consented to assist Edward in a few photographic experiments with the Harvard telescope. One of their pictures yielded 462 stars in a region where only 55 had been previously documented.

  The part of Pickering’s plan with the greatest potential interest for Mrs. Draper concerned a new approach to photographing stellar spectra. Rather than focus on one target star at a time, à la Draper or Huggins, Pickering anticipated group portraits of all the brightest stars in a wide field of view. To achieve these, he envisioned a new instrument setup combining telescope and spectroscope with the type of lens used in the studios of portrait photographers.

  “I think there will be no difficulty in carrying out this plan without your aid,” he assured Mrs. Draper. “On the other hand, if it commends itself to you, I am confident that we could make it conform to such conditions as you might impose.”

  “Thanks for your kindness,” she replied on May 21, 1885, “in remembering my desire to be interested in some work with which Dr. Draper’s name could be associated, and his memory kept alive. I will be glad to cooperate, if I can, in what you suggest, for its bearing on stellar spectrum photography appeals to me very strongly.” More than two years had passed since Henry’s death. Still unable to make his observatory productive, she saw no harm in lending his name to Harvard.

  Pickering proceeded slowly and with caution, apprising her of his progress until he could send her some sample images of stellar spectra taken through his new apparatus. She found them “exceedingly interesting.” On January 31, 1886, she said, “I would be willing, if the plan could be carried out satisfactorily, to authorize the expenditure of $200 a month or somewhat more if necessary.” Pickering thought more would be needed. They settled terms on Valentine’s Day for the Henry Draper Memorial—an ambitious photographic catalogue of stellar spectra, gathered on glass plates. Its goal was the classification of several thousand stars according to their various spectral types, just as Henry had set out to do. All results would be published in the Annals of the Harvard College Observatory.

  On Febr
uary 20, 1886, Mrs. Draper sent Pickering a check for $1,000, the first of many installments. Pickering publicized the new undertaking in all the usual places, including Science, Nature, and the Boston and New York newspapers.

  Later that spring Mrs. Draper decided to increase her already generous gift by donating one of Henry’s telescopes. She visited Cambridge in May to make the arrangements. Since the instrument needed a new mounting—something Henry had meant to build himself—she asked George Clark of Alvan Clark & Sons to fabricate the parts, at a cost of $2,000, and to oversee the transfer of the equipment from Hastings to Harvard. Once arrived, it would require its own small building with a dome eighteen feet in diameter, and Mrs. Draper meant to cover that expense as well. Together with the Pickerings, she strolled among the plantings of rare trees and shrubs around the observatory to select a site for the new addition.

  CHAPTER TWO

  What Miss Maury Saw

  THE INFUSION OF FUNDS for the Henry Draper Memorial made the Harvard College Observatory hum with new people and purpose. Construction of the small building to house Dr. Draper’s telescope started in June 1886 and continued through the summer while Mrs. Draper toured Europe. In October the instrument was mounted in the new dome. Now there were two telescopes outfitted for nightly rounds of spectral photography—the Draper 11-inch and an 8-inch purchased with a $2,000 grant from the Bache Fund of the National Academy of Sciences. The illustrious Great Refractor, through which the first-ever photograph of a star had been taken in 1850, later proved unsuitable for photography. Its 15-inch lens had been fashioned for visual observing; that is, for human eyes most attuned to yellow and green wavelengths of light. The lenses of the two new instruments, in contrast, favored the bluer wavelengths to which photographic plates were sensitive. The 8-inch Bache telescope also boasted a wide field of view for taking in huge tracts of sky all at once, rather than homing in on single objects.

  In less than a decade at the helm, Edward Pickering had shifted the observatory’s institutional emphasis from the old astronomy, centered on star positions, to novel investigations into the stars’ physical nature. While half the computing staff continued to calculate the locations and orbital dynamics of heavenly bodies, a few of the women were learning to read the glass plates produced on-site, honing their skills in pattern recognition in addition to arithmetic. A new kind of star catalogue would soon emerge from these activities.

  The earliest known star counter, Hipparchus of Nicaea, catalogued a thousand stars in the second century BC, and later astronomers enumerated the content of the heavens to ever better effect. The projected Henry Draper Catalogue would be the first in history to rely entirely on photographs of the sky and to specify the “spectrum type,” as well as the position and brightness, for myriad stars.

  Dr. and Mrs. Draper had gathered their spectra one by one, using a prism at the telescope’s eyepiece to split the light of each star. Pickering and his assistants, eager to increase the pace of operations, altered the Drapers’ approach. By installing prisms at the objective, or light-gathering end of the telescope, instead of at the eyepiece, they were able to capture group portraits containing two or three hundred spectra per plate. The prisms were large, square sheets of thick glass, wedge-shaped in cross section. “The safety and convenience of handling the prisms,” Pickering found, “is greatly increased by placing them in square brass boxes, each of which slides into place like a drawer.” Harvard’s picture gallery grew apace. When Mrs. Draper paid another visit soon after Thanksgiving, Pickering assured her that any star visible from Cambridge appeared on at least one of the glass plates.

  Toward the end of December 1886, just when the staff had smoothed out most of the difficulties with the new procedures, Nettie Farrar’s beau proposed. Pickering was all in favor of marriage, of course, but he hated to lose Miss Farrar, a five-year veteran of the computing corps whom he had personally trained to measure spectra on the photographic plates. On New Year’s Eve, he wrote to inform Mrs. Draper of Miss Farrar’s engagement, and also to name Williamina Fleming, the former maid, as her replacement.

  Since returning from Scotland in 1881, Mrs. Fleming had been assisting Pickering with photometry. Often she took the director’s penciled notations from the nightly observations with his assistants and applied the formulas he specified to compute the stars’ magnitudes. By 1886, when the Royal Astronomical Society awarded Pickering its gold medal for this work, he had already embarked on a parallel approach to photometry via photography. This change required Mrs. Fleming, well accustomed to reading lists of numbers scribbled in the dark, to judge magnitudes from fields of stars on glass plates.

  Mrs. Fleming had let Pickering know that photography ran in her pedigree. Her father, Robert Stevens, a carver and gilder praised for his gold-leaf picture frames, had been the first in the city of Dundee to experiment with daguerreotyping, as the process was called in her childhood. She was still a child, only seven, when her father died suddenly of heart failure. Her mother and older siblings tried, for a time, to keep the business running without him, but without success. One by one, her older brothers sailed away to Boston, where she eventually followed them. Now, at twenty-nine, she had a seven-year-old child of her own to care and provide for. Edward would soon arrive; her mother was booking passage with him on the Prussian out of Glasgow.

  Miss Farrar dutifully introduced Mrs. Fleming to the plates of stellar spectra, and taught her how to measure the hordes of tiny lines. Mrs. Fleming could have taught Miss Farrar a thing or two about marriage and childbirth, but on the subject of the spectrum she had everything to learn.

  • • •

  THE YOUNG ISAAC NEWTON coined the word spectrum in 1666, to describe the rainbow colors that arose like ghostly apparitions when daylight passed through cut glass or crystal. Although his contemporaries thought glass corrupted the purity of light by imparting color to it, Newton held that colors belonged to light itself. A prism merely revealed white light’s component hues by refracting them at different angles, so that each could be seen individually.

  The microscopic dark lines within the stellar spectra, to which Mrs. Fleming now directed her attention, were called Fraunhofer lines, after Joseph von Fraunhofer of Bavaria, their discoverer. A glazier’s son, Fraunhofer had apprenticed at a mirror factory and gone on to become a master crafter of telescope lenses. In 1816, in order to measure the exact degree of refraction in different glass recipes and lens configurations, he built a device that combined a prism with a surveyor’s small telescope. When he directed a beam of light from the prism through a slit and into the instrument’s magnified field of view, he beheld a long, narrow rainbow marked by many dark lines. Repeated trials convinced him that the lines, like the rainbow colors, were not artifacts of passage through glass, but inherent in sunlight. Fraunhofer’s lens-testing apparatus was the world’s first spectroscope.

  Charting his finds, Fraunhofer labeled the most prominent lines with letters of the alphabet: A for the wide black one at the rainbow’s extreme red end, D for a dark double stripe in the orange-yellow range, and so on through the blue and violet to a pair named H, and ending farther along the violet with I.

  Fraunhofer’s lines retained their original alphabetical designations through the decades following his death, gaining greater importance as later scientists observed, mapped, interpreted, measured, and depicted them with fine-nib pens. In 1859 chemist Robert Bunsen and physicist Gustav Kirchhoff, working together in Heidelberg, translated the Fraunhofer lines of the Sun’s spectrum into evidence for the presence of specific earthly substances. They heated numerous purified elements to incandescence in the laboratory, and showed that each one’s flame produced its own characteristic spectral signature. Sodium, for example, emitted a close-set pair of bright orange-yellow streaks. These correlated in wavelength with the dark doublet of lines that Fraunhofer had labeled D. It was as though the laboratory sample of burning sodium had colored in those partic
ular dark gaps in the Sun’s rainbow. From a series of such congruities, Kirchhoff concluded the Sun must be a fireball of multiple burning elements, shrouded in a gaseous atmosphere. As light radiated through the Sun’s outer layers, the bright emission lines from the solar conflagration were absorbed in the cooler surrounding atmosphere, leaving dark telltale gaps in the solar spectrum.

  Astronomers, many of whom had considered the Sun a temperate, potentially habitable world, were shocked to learn of its inferno-like heat. However, they were soon placated—even soothed—by the revelatory power of spectroscopy to expose the chemical content of the firmament. “Spectrum analysis,” Henry Draper told the Young Men’s Christian Association of New York in 1866, “has made the chemist’s arms millions of miles long.”

  Throughout the 1860s, pioneering spectroscopists such as William Huggins discerned Fraunhofer lines in the spectra of other stars. In 1872 Henry Draper began photographing them. While the number of spectral lines in starlight paled in comparison to the rich tapestry of the Sun’s spectrum, several recognizable patterns emerged. It seemed that the stars, which had for so long been loosely categorized by brightness or color, could now be further sorted according to spectral features hinting at their true nature.

  In 1866 Father Angelo Secchi of the Vatican Observatory divided four hundred stellar spectra into four distinct types, which he designated by Roman numerals. Secchi’s Class I contained brilliant blue-white stars such as Sirius and Vega, whose spectra shared four strong lines indicating the presence of hydrogen. Class II included the Sun and yellowish stars like it, with spectra full of many fine lines identifying iron, calcium, and other elements. Classes III and IV both consisted of red stars, differentiated by the patterns in their dark spectral bands.