Red-Shift Spectroscopy and Stellar Classification

Glass Plate photonegative c10213

With the Spectroscope...

Using a magnifying loupe and spectra image, Maury set to the task of parsing out the stars (Fig. 2 & 3). The first challenge she set for herself at HCO was to unpack the so-called "spectroscopic binary star," binary star systems whose structure and dynamics are not visible with the naked eye, but only emerge through changes in the spectra (Fig. 4). To capture a spectrum, starlight was focused through a telescope, then a prism, spreading it into a rainbow. This was directed onto a glass plate covered with light sensitive chemicals to capture a negative. Lines occur in a star’s spectrum when its light passes through clouds of atoms, as in a star’s corona or a dust cloud, and particular wavelengths are absorbed by those atoms’ electrons. 

Explanation of Beta Lyrae system

Photograph of the spectrum of a Cygni          Red shift explanatory diagram


To tell if a star is in motion, and if that motion is towards or away from earth, Maury used redshift calculations. She began by measuring and noting the locations and intensities of spectral lines in two images of the same star; here her notions are shown for a spectra of α Cygni (Fig. 5). She then compared those with standard element spectra measured on Earth, and calculated how much each line was shifted right (example shown in Fig. 6). She could then determine the velocity of a star with the following maxim: the faster a star is moving away, the more red shifted its spectrum. This is due to the same Doppler effect as on Earth when a siren passes, wee-ing towards you and wobble-ing away. But Maury's observations were not so visceral: an excerpt from one of Maury’s notebooks shows the abstract calculations of the wavelength shift of three Hydrogen spectral lines on plate x7720, using the Hartmann formula for calculating exact wavelengths (Fig. 7). A summary table only gestures at more complexity (Fig. 8) -- the addition of rates of change (velocities), error margins, and special notes amounts to static noise, but noise that reveals the presence of two stars where one is seen. The cyan boxes at the bottom of Fig. 8 highlight that these calculations are relevant to the analysis of β Lyrae and α Cygni, two spectroscopic binaries discovered by Maury.

 Notebook page showing red shift calculationsSummary page from Maury’s notebook referencing both a Cygni and b Lyrae

...Classifying Stars

In addition to her analysis of specific stars, Maury also developed a classification system for stellar spectra, based on examination of ~4,800 photos of the spectra of 681 stars. Fig 9 shows the first page of a 16 page table, the twelfth of which to group similar stars (Fig. 9: 1897AnHar..28....1M). Maury's classification system was notoriously complex, so much so that the astronomy community preferred the simpler styles of Annie Cannon. But Maury's system appealed to a few notable scientists for its rigor and innovative groupings. She began with a standard set of groupings, taking 22 Roman numerals, and subdividing these classes into 3-4 subgroups, sparse in population but indicative of different qualities. For example, many of her subclassifications in the and ac groups contained stars that were brighter, but otherwise identical, to other in that numeral group. Danish astronomer Ejnar Hertzsprung in particular favored this classification, also suspecting a unique quality for these brighter stars. 

Table 8 from 1897 AM's Spectra of Bright Stars

In a 1911 paper, Hertzsprung used Maury's data to plot the relative brightness of a star in the Hyades cluster against its color, expressed as wavelengths (Fig. 10: 2013ASPC..471..205G). The result shows two distinct columns of stars: a thick group of bluer stars evenly dispersed across brightnesses, and a thiner group of redder stars clustered around specific magnitudes. Because Hertzsprung used stars in the Hyades cluster, he could assume that they were all the same distance away, and thus he could observe their intrinsic (as opposed to relative) brightness. With such an absolute scale, Hertzsprung concluded that these stars were fundamentally of the same type, just changed by the passage of time. With help from the astronomical community (especially his colleague Henry Noris Russell), Hertzsprung forumlated a diagram that showed the age of a star to be a function of its color and brightness (Fig. 11). Affirmed for almost a century, the GAIA satellite from the ESA assembled in 2018 a color-luminosity diagram from over 4 million observation points, confirming that Fig. 11 matches the observed universe (Fig. 12). 


Hertzsprung 1911 luminosity-color diagram

Theoretical HR-Diagram

Gaia HR-Diagram

Maury's science lives on in other scientific missions to this day. The Sloan Digital Sky Survey, begun in 2000, uses redshift and parallax observations to deduce the structure of the universe, in much the same way that Maury used those methods to identify and analyze spectroscopic binary systems (Fig. 13). While the new digital computations allow for more precise measurements across greater magnitudes of space and time, their basis in methods pioneered by Antonia Maury demonstrate her lasting influence on the discipline.

Sloan Digital Sky Survey 20th Anniversary Image

Image Sources

  1. (1887). Senior portrait of Antonia Maury [Photograph], (Identifier Ph.f 11.24), Archives and Special Collections, Vassar College Library, Poughkeepsie, NY.
  2. (1899). Plate c12013. Center for Astrophysics | Harvard & Smithsonian, Photographic Glass Plate Collection, Cambridge, MA.
  3. Adapted from “Beta Lyrae Low Resolution Eclipse Spectroscopy - Alpy 600 Spectrograph” by Jim Ferreira, 2014.
  4. Adapted from “Beta Lyrae Low Resolution Eclipse Spectroscopy - Alpy 600 Spectrograph” by Jim Ferreira, 2014.
  5. (1899). Plate c11989. Center for Astrophysics | Harvard & Smithsonian, Photographic Glass Plate Collection, Cambridge, MA.
  6. (2011). Absorption lines in the optical spectrum of a supercluster of distant galaxies as compared to those in the optical spectrum of the Sun. Retrieved from
  7. Maury, A. (1918). Formulae for reduction, Phase in HA84.Page 90. Project PHaEDRA (Box 42), Center for Astrophysics | Harvard & Smithsonian, Cambridge, MA.
  8. Maury, A. (1918). Formulae for reduction, Phase in HA84.Page 60. Project PHaEDRA (Box 42), Center for Astrophysics | Harvard & Smithsonian, Cambridge, MA.
  9. Maury, A. (1897). Spectra of bright stars photographed with the 11-inch Draper telescope as part of the Henry Draper Memorial. Annals of the Astronomical Observatory of Harvard College, 28(1), p. 107.
  10. Adapted from [Figure 3] from “Über den Zusammenhang von Helligkeit und Spektraltypus in den Plejaden” by H. Rosenberg, 1910, Astronomische Nachrichten, 186(5), p. 76.
  11. European Southern Observatory. (2007). Hertzsprung-Russell diagram [Image]. Retrieved from
  12. ESA, Gaia, DPAC. (2018). Gaia’s Hertzsprung-Russell diagram [Image]. Retrieved from
  13. Belokurov, V., Blanton, M. R., Bonaca, A., Fan, X., Geha, M. C., Lupton, R. H., the SDSS Collaboration. (2018). Map showing some of what the Sloan Digital Sky Survey has discovered over the last twenty years [Image]. Retrieved from