On a Tuesday morning in October 1922, in a brick building on Garden Street in Cambridge, Massachusetts, a woman named Antonia Maury walked into the second-floor classification room at the Harvard College Observatory and asked Annie Jump Cannon how long it would take to finish.
Cannon was at her desk, classifying spectra at her habitual rate of about three stars a minute. She looked up and said, without consulting her notes, that she expected to complete the last volume by the end of 1924.
She was off by about three months. The ninth volume of the Henry Draper Catalogue, covering southern declinations from minus 30 to minus 60 degrees, was published in March 1925. The previous eight volumes had appeared between 1918 and 1924. The work, which had begun in earnest in 1911, took fourteen years.
The catalogue contains spectral classifications for two hundred and twenty-five thousand three hundred stars, brighter than ninth magnitude, across the entire sky. It uses the Harvard spectral sequence, which is the OBAFGKM system every astronomy student still memorises.
The sequence itself is largely Cannon's. She arranged the classes, in 1901, into the order based on surface temperature rather than the alphabetical order earlier used by Williamina Fleming. She added the subdivisions from zero to nine within each class. She defined the criteria for distinguishing one class from the next.
The work was funded by Anna Draper, widow of the physician and amateur astrophotographer Henry Draper, who had pioneered the photography of stellar spectra before his death in 1882. Anna Draper's bequest, beginning in 1886, supported the program for forty years.
The bequest paid for plates, chemicals, salaries, and the time of the Boyden Station in Arequipa, Peru, which photographed the southern sky and shipped the glass plates north by steamship. By 1911, when systematic classification began, the Harvard archive held over two hundred thousand plates.
Cannon worked from objective-prism plates: thin glass coated with photographic emulsion, exposed at the telescope, then developed and stored. A single plate could contain spectra of several hundred stars, each appearing as a small streak with absorption lines visible at the appropriate wavelengths.
She used a magnifying eyepiece. She held the plate in her left hand and turned it slowly. She spoke the classification aloud to a recorder, who wrote it down. The recorders, in different years, included Edith Gill, Florence Cushman, and a young woman named Margaret Walton, who later married and left the observatory.
Cannon's rate of three stars per minute, sustained over several hours at a stretch, is something approaching legend in the field. Modern automated classification programs, run on stellar spectra of similar quality, achieve a comparable accuracy. They do not, however, achieve it any faster than she did.
The accuracy of the catalogue is itself remarkable. Comparisons with modern classifications by, among others, the MK system of Morgan, Keenan, and Kellman (1943) and the more recent LAMOST spectral survey, show that Cannon's classifications are within one subclass for ninety-four percent of stars. They are within two subclasses for ninety-nine percent.
The work was not solitary. The Harvard College Observatory's program of stellar classification, begun under Edward Pickering in the 1880s, employed by 1911 a staff of roughly thirty women working at any one time on various aspects of the photographic archive.
They were, in the language of the period, computers. The term referred to people, not machines, and described any worker engaged in routine mathematical calculation. At Harvard the computers worked at long tables in a basement classroom and were paid roughly half what a male assistant would have earned for similar work.
Pickering, who died in 1919, has been generously credited and somewhat blamed for the arrangement, depending on the historian. The salaries were low; the work was difficult; the access to telescope time and to publication was uneven. Cannon herself published forty-seven papers under her own name during her years at the observatory.
What the women produced was, by any standard, extraordinary. In addition to the Henry Draper Catalogue, the Harvard program issued the Draper Catalogue Extension (1925-1936, adding forty-six thousand fainter stars), Williamina Fleming's catalogue of two hundred and twenty-two variable stars, Henrietta Leavitt's period-luminosity relation for Cepheid variables, and Cecilia Payne's 1925 doctoral thesis on stellar composition.
The Henry Draper Catalogue's HD numbers remain in use. A star catalogued as HD 209458, for instance, is the same star that hosts one of the best-studied transiting exoplanets. Astronomers writing about it in 2026 still call it by its 1923 designation.
Cannon herself was awarded the Henry Draper Medal of the National Academy of Sciences in 1931. She was the first woman to receive it. She was made an honorary professor at Harvard in 1938, at the age of seventy-five, and the first woman to receive an honorary doctorate from Oxford, in 1925.
She died in 1941, in Cambridge, at the age of seventy-seven. Her ashes are buried at Mount Auburn Cemetery. The Harvard plate archive, which preserves the original spectra she classified, is held today in a climate-controlled basement at the Center for Astrophysics and is being digitised at a rate of about twelve thousand plates per year.
The catalogue itself is freely available through the VizieR online service, hosted by the Strasbourg Astronomical Data Centre. It is, by current count, the seventh most-cited stellar catalogue in the history of the discipline.
Annie Jump Cannon's name is on it. So are, in the introduction, Edward Pickering's and the names of the women who recorded the classifications. The names of the plate developers are not recorded. Neither are the names of the porters who carried the plates up from the steamship docks in Boston.
It is the way of catalogues that some names survive and others do not. The names that survive are usually the ones written on the title page. The names that do not are the ones whose work made the title page possible.
