The Dundee Maid Who Classified the Stars

Harvard and Edward Pickering

The famous story — that Pickering, exasperated by his male assistants' sloppy work, declared that his Scottish maid could do better — is frequently repeated but is regarded by historians as at least partly apocryphal. What is documented is that Pickering's wife Elizabeth recognised Fleming's talents, and that from 1881 Pickering employed her at the Harvard College Observatory, teaching her to analyse stellar spectra.

This was the dawn of a revolution. Photography had transformed astronomy: rather than squinting through an eyepiece, astronomers could now capture the spectra of hundreds of stars on a single glass plate using an objective prism, then study them by daylight at leisure. The problem was scale — there was far more data than there were people to analyse it. Pickering's solution was to hire women as "computers", the human calculators and classifiers who processed the plates. The group — later dubbed the Harvard Computers — grew to more than eighty women over Pickering's tenure and included Annie Jump Cannon, Henrietta Swan Leavitt and Antonia Maury. As Smithsonian Magazine records, they were "putting in six-day weeks poring over photographs, and earning 25 to 50 cents an hour (half what a man would have been paid)."

Fleming began with clerical and copying work and rose to become the observatory's leading classifier and, eventually, a manager supervising a dozen or more women.

The Pickering–Fleming System

When Mary Anna Draper endowed the Henry Draper Memorial in 1886 to fund the systematic photography and classification of stellar spectra, Fleming was placed in charge of the project. The existing system — Angelo Secchi's five visual classes — was too crude for the detail visible on the new photographs. Fleming devised a new scheme, subdividing Secchi's classes and assigning each a letter of the alphabet, A through Q, ordered principally by the strength of the hydrogen absorption lines in each spectrum, with type A having the strongest. The letter Q was used for spectra that fitted no other category.

This became the basis for the first Draper Catalogue of Stellar Spectra, published in 1890, containing spectral classifications for 10,351 stars — the majority classified by Fleming herself. It is important to be honest about the collaborative nature of the work: Pickering directed the overall project and is often credited as author, and the system is sometimes called the "Pickering–Fleming system". But the practical classification labour, and much of the scheme's design, was Fleming's.

Her system was empirical and had a flaw: the alphabetical order did not produce a smooth physical sequence. Antonia Maury and then Annie Jump Cannon refined it, dropping most letters and reordering the survivors into the now-famous O, B, A, F, G, K, M sequence, which turned out to track stellar surface temperature. Cannon's system, applied in the full Henry Draper Catalogue published 1918–1924 — classifying 225,300 stars, later extended to 359,083 — remains the foundation of stellar classification today.

The Horsehead Nebula (1888)

In 1888, examining a plate designated B2312 — taken in early February 1888 with the 8-inch Bache telescope in Cambridge, in a photograph associated with Pickering's brother William Henry Pickering — Fleming noticed something unusual near Orion's Belt. She described "a semicircular indentation" in the bright nebulosity, later catalogued as IC 434, south of Zeta Orionis. This was the first recorded detection of what we now call the Horsehead Nebula (Barnard 33), a dark cloud of dust silhouetted against glowing gas roughly 1,500 light-years away in Orion.

Fleming was not initially credited. When J. L. E. Dreyer compiled the first Index Catalogue in 1895, he attributed the Harvard discoveries simply to "Pickering". By the second Index Catalogue in 1908, Harvard's women were prominent enough to receive proper credit for later discoveries — but the Horsehead attribution was never fully corrected in that era. Today the Horsehead is among the most photographed and iconic objects in all of astronomy.

40 Eridani B — The Claim, Handled Carefully

This is the claim most often garbled, so precision matters.

The star. 40 Eridani B — also written Omicron² Eridani B, or o² Eri B — is a component of a triple star system about 16.3 light-years away in Eridanus. It is a white dwarf of apparent magnitude about 9.5, with roughly 57% of the Sun's mass packed into a body only about 1.4% of the Sun's radius, a little larger than Earth, and an effective temperature near 16,500–17,000 K. Its modern spectral type is DA2.9. It is the second-brightest white dwarf and by far the easiest to observe, because unlike the companions of Sirius and Procyon it is not drowned out by a dazzling primary.

The observation history. The B/C pair was first resolved by William Herschel on 31 January 1783, and observed again by Friedrich Struve in 1825 and Otto Struve in 1851. None of them recognised anything strange about it.

What Fleming actually did. In 1910 the American astronomer Henry Norris Russell was measuring stellar parallaxes — distances — and found 40 Eridani B to be intrinsically very faint. Pickering offered to have the Harvard plates checked. As Russell recounted decades later, Pickering telephoned Mrs Fleming, who reported within an hour or so that the star's spectrum was Class A: hot and blue-white. This was the puzzle. An A-type star should be luminous, yet this one was extremely faint. The only resolution was that the star had to be extraordinarily small and dense. Fleming's specific contribution, then, was the precise spectral classification — identifying 40 Eridani B as anomalously hot for so faint a star. She did not, and could not, describe it as a "white dwarf".

Who recognised the anomaly, and when. The realisation of the physical implication — that here was a new class of tiny, dense star — belongs to Russell, with Pickering, around 1910. In his 1944 note "Notes on White Dwarfs and Small Companions" in The Astronomical Journal (vol. 51, pp. 13–17), Russell recalled that he was "puzzled" and "crestfallen" at the discrepancy, and that "Pickering smiled and said, 'It is just such discrepancies which lead to the increase of our knowledge.'" In the same note, Russell wrote: "The first person who knew of the existence of white dwarfs was Mrs. Fleming; the next two, an hour or so later, Professor E. C. Pickering and I."

A note of caution. Some popular sources state that "in 1910 Fleming published her discovery of white dwarf stars." This appears to be a distortion. There is no 1910 publication by Fleming announcing white dwarfs; the concept and the term did not yet exist. The episode is known chiefly through Russell's 1944 recollection. What is accurate is that Fleming's 1910 spectral classification was the observational key that unlocked the eventual identification.

Sirius B, for comparison. Sirius B, the companion of the brightest star in the sky, was first seen by Alvan Graham Clark on 31 January 1862. But its true nature was not understood until Walter Adams obtained its spectrum at Mount Wilson in 1915, finding it hot and white like 40 Eridani B — thereby confirming a whole new class of underluminous stars. In terms of which white dwarf was recognised first in the historical record, 40 Eridani B (1910) preceded Sirius B (1915).

The term. "White dwarf" was coined by Willem Luyten in 1922, to describe these faint, hot, white A-type stars. Before that, astronomers had no single term — they spoke of anomalous or "subluminous" faint stars with early-type spectra. The theoretical understanding — degenerate electron-gas matter, from Fowler in 1926, and later the Chandrasekhar mass limit — came later still. Fleming, who died in 1911, never heard the words "white dwarf".

Other Discoveries and Honours

Across her career Fleming discovered 59 gaseous nebulae, 10 novae, and over 310 variable stars — she published a list of 222 in her 1907 study. She also established the first photographic standards of stellar magnitude. She published the Draper Catalogue of Stellar Spectra (1890), A Photographic Study of Variable Stars (1907), Spectra and Photographic Magnitudes of Stars in Standard Regions (1911), and, posthumously, "Stars Having Peculiar Spectra" (1912).

In 1898 the Harvard Corporation appointed her Curator of Astronomical Photographs — the first woman to hold such a title at Harvard. In 1906 she was elected an honorary member of the Royal Astronomical Society of London, the first American woman so honoured; women could not be elected full Fellows until 14 January 1916. Shortly before her death she received the Guadalupe Almendaro medal of the Astronomical Society of Mexico. A lunar crater, Fleming — on the Moon's far side, shared with Alexander Fleming and named by the IAU in 1970 — and the asteroid 5747 Williamina bear her name.

Pay, Character and Her Own Words

Fleming raised her son Edward alone, putting him through MIT. Her diary, quoted from her 1900 journal, records her frustration over pay. After raising the subject with Pickering, she wrote: "I am immediately told that I receive an excellent salary as women's salaries stand… Does he ever think that I have a home to keep and a family to take care of as well as the men?… And this is considered an enlightened age!" Her annual salary as a senior computer and curator was $1,500 around 1900 — while equivalent male roles at the observatory could earn up to $2,500.

She examined an enormous number of plates over her career, and her reputation as a strict, exacting supervisor lived after her. She became a US citizen in 1907, listing her occupation as "astronomer" — as does her gravestone in Mount Auburn Cemetery, Cambridge. She died of pneumonia in Boston on 21 May 1911, aged 54. She was one of eleven women charter members of the American Astronomical Society, and was succeeded as curator by her protégée, Annie Jump Cannon.

Legacy

Fleming's classification work is the direct ancestor of the system astronomers use today. The Harvard scheme she helped build evolved through Cannon into the OBAFGKM sequence and ultimately the modern MK system. The Henry Draper Catalogue's HD numbers remain in daily use for identifying stars — Sirius, for instance, is HD 48915. The broader Harvard Computers programme she helped anchor produced Henrietta Swan Leavitt's period–luminosity relation for Cepheid variables, which gave astronomers their first reliable cosmic yardstick and, later, made possible the measurement of the expanding universe.

Fleming also did something less quantifiable but just as important: she demonstrated, in an age that doubted it, that a woman could perform rigorous, front-rank scientific analysis. She used her platform to say so — at the 1893 World's Fair in Chicago she delivered "A Field for Woman's Work in Astronomy", arguing that with real opportunities women could become men's equals in the discipline.

Scotland can rightly claim her. Born and educated in Dundee, she carried a hard-won education across the Atlantic and turned adversity into one of the most productive careers in the history of astronomy — while the institution that employed her paid her half a man's wage for it.

Timeline

1. 1857

   Williamina Paton Stevens born at 86 Nethergate, Dundee, 15 May

   One of nine children of Robert Stevens, a carver and gilder, and Mary Walker

2. c.1864

   Her father dies when she is seven

   The family falls on hard times

3. 1871

   Becomes a pupil-teacher at 14

   Teaching while continuing her own studies in Dundee's public schools

4. 26 May 1877

   Marries James Orr Fleming in Dundee

   An accountant and widower some sixteen years her senior

5. 1878

   Emigrates to Boston, Massachusetts

   Shortly after arriving, her husband abandons her — pregnant and destitute

6. 6 Oct 1879

   Her son Edward Pickering Fleming is born

   She takes work as a maid in the household of Professor Edward Charles Pickering

7. 1881

   Hired at Harvard College Observatory

   Pickering's wife Elizabeth recognised her talents; Pickering teaches her to analyse stellar spectra

8. 1886

   Mary Anna Draper endows the Henry Draper Memorial

   Fleming is placed in charge of the systematic photography and classification project

9. 1888

   Detects the Horsehead Nebula on plate B2312

   'A semicircular indentation' in the bright nebulosity later catalogued as IC 434, south of Zeta Orionis

10. 1890

    The Draper Catalogue of Stellar Spectra published

    10,351 stars classified A through Q; the majority by Fleming herself

11. 1893

    Delivers 'A Field for Woman's Work in Astronomy' at the Chicago World's Fair

    Argues that with real opportunities women could become men's equals in the discipline

12. 1895

    Dreyer's first Index Catalogue attributes Harvard discoveries to 'Pickering'

    Fleming is not credited for the Horsehead

13. 1898

    Appointed Curator of Astronomical Photographs by the Harvard Corporation

    The first woman to hold such a title at Harvard

14. 1900

    Her diary records her frustration over pay

    'Does he ever think that I have a home to keep and a family to take care of as well as the men?'

15. 1906

    Elected an honorary member of the Royal Astronomical Society

    The first American woman so honoured; women could not be elected full Fellows until 14 January 1916

16. 1907

    Publishes A Photographic Study of Variable Stars

    Includes a list of 222 variables; she becomes a US citizen the same year, listing her occupation as 'astronomer'

17. 1910

    Classifies 40 Eridani B as spectral type A

    Russell, measuring parallaxes, had found it intrinsically very faint; Pickering telephoned Fleming, who reported back within an hour or so

18. 1911

    Dies of pneumonia in Boston, 21 May, aged 54

    Succeeded as curator by her protégée Annie Jump Cannon

19. 1915

    Walter Adams obtains the spectrum of Sirius B at Mount Wilson

    Hot and white like 40 Eridani B — confirming a whole new class of underluminous stars

20. 1918–24

    Cannon's full Henry Draper Catalogue published

    225,300 stars, later extended to 359,083, on the OBAFGKM sequence descended from Fleming's scheme

21. 1922

    Willem Luyten coins the term 'white dwarf'

    Eleven years after Fleming's death

22. 1944

    Russell publishes 'Notes on White Dwarfs and Small Companions'

    The Astronomical Journal 51: 13–17 — the recollection through which the 1910 episode is known

Myths & Facts

Did You Know?

• Fleming arrived in America as an abandoned, pregnant immigrant working as a maid — and went on to discover the Horsehead Nebula and help classify the heavens.
• She discovered the Horsehead Nebula in 1888 from a single glass photographic plate; it is now one of the most iconic images in astronomy.
• She was paid 25 to 50 cents an hour at Harvard, about half a man's wage; even as curator her $1,500 salary trailed the up-to-$2,500 paid for comparable male roles.
• She classified over 10,000 stars for the Henry Draper survey — 10,351 in the 1890 catalogue.
• The term 'white dwarf' wasn't coined until 1922, by Willem Luyten, eleven years after Fleming died — yet her 1910 spectral classification of 40 Eridani B made the identification possible.
• She was elected an honorary member of the Royal Astronomical Society in 1906; women couldn't become full Fellows until 1916.
• Her father was among the first in Dundee to experiment with photography — the medium his daughter would spend her life reading.

Honest Caveats

The "my maid could do better" anecdote is likely apocryphal. It is presented here as legend, not as documented fact.

"Fleming discovered or published white dwarfs in 1910" is a popular distortion, repeated even on Wikipedia. The accurate account is that she classified 40 Eridani B's spectrum as type A, and Russell recognised the anomaly. The episode is documented via Russell's 1944 Astronomical Journal note, not a 1910 Fleming paper.

Object counts vary slightly between sources — 52 versus 59 nebulae, for example — and her novae are occasionally mislabelled "supernovae". The most commonly cited figures, from the Cannon obituary, are 59 gaseous nebulae, 10 novae and over 310 variable stars.

Her appointment as curator is dated 1898 by the Harvard Corporation records and 1899 by some biographies; 1898 is the better-sourced date.

The lunar crater Fleming honours both Williamina and Alexander Fleming jointly, which is a frequent source of confusion between two unrelated Scots.