The women of Harvard's computers: Cannon, Leavitt, Fleming
Hired as calculators, paid less than the assistants, and they built the framework the field still uses.
4 min read · Published 28 August 2026
The short answer
From the 1880s Harvard College Observatory employed women to process astronomical photographic plates — work then classified as computation rather than research. Williamina Fleming developed an early classification scheme and discovered hundreds of variable stars. Annie Jump Cannon created the OBAFGKM spectral sequence still used today. Henrietta Swan Leavitt discovered the period-luminosity relationship for Cepheid variables, which made measuring cosmic distances possible.
What the work was
Photographic astronomy produced glass plates in enormous quantities, each holding hundreds of stellar spectra. Someone had to measure, classify and catalogue them.
Edward Pickering, the observatory director, hired women for this work. The reasons were partly practical and partly economic: women could be paid substantially less than men for the same hours, and a fixed budget bought more labour.
The role was defined as computation — mechanical processing of data — rather than research, which meant the people doing it were not expected to interpret what they found.
Several of them did anyway, and produced results that reorganised the field.
Annie Jump Cannon and the spectral sequence
Cannon classified stellar spectra at a rate nobody has matched — hundreds of thousands of stars over her career, at times more than three a minute.
More importantly, she reorganised the classification. Earlier schemes had ordered classes alphabetically by hydrogen line strength. Cannon recognised that reordering and merging them produced a sequence that was physically meaningful, corresponding to temperature.
The result is OBAFGKM, still the foundation of stellar classification and the horizontal axis of the Hertzsprung–Russell diagram. Its non-alphabetical order is a direct fossil of that reordering.
The Sun is a G-type star because of her scheme. Cannon was deaf from early adulthood, and colleagues noted it may have helped her concentrate through the noise of a working observatory.
Henrietta Leavitt and the measurement of the universe
Leavitt was assigned to identify variable stars in photographic plates of the Magellanic Clouds, and catalogued 1,777 of them.
She then noticed something nobody had asked her to look for. Among Cepheid variables, the brighter ones pulsed more slowly, and the relationship was regular. Because all the stars in a Magellanic Cloud are at effectively the same distance, differences in apparent brightness reflected genuine differences in luminosity.
That relationship — now Leavitt's Law — turns a Cepheid into a standard candle. Measure its pulsation period, derive its true brightness, compare with how bright it appears, and you have its distance.
This is the foundation of the cosmic distance ladder. Hubble used Cepheids to establish that other galaxies exist and to measure the expansion of the universe. Neither was possible before her work.
Leavitt died of cancer in 1921. A Swedish mathematician attempted to nominate her for the Nobel Prize in 1926, unaware she had died four years earlier; the prize cannot be awarded posthumously.
The conditions, stated plainly
- Pay. They were paid substantially less than men doing comparable work, and less than some unskilled male staff at the observatory.
- Title. 'Computer' was a job description denoting mechanical processing, which is why the research they produced was often published under a supervisor's name or with minimal attribution.
- Access. They were largely excluded from operating the telescopes whose plates they analysed.
- Advancement. Cannon was not given a formal Harvard appointment until 1938, decades after her defining work and shortly before her death.
It matters that this is stated as fact rather than as background colour. The stellar classification system in use today and the method that established the scale of the universe were both produced under those conditions, by people whose job description said they were not doing research.
The pattern recurs — Jocelyn Bell Burnell's omission from the 1974 Nobel Prize for pulsars is the best-known later instance.
Common questions
- Who were the Harvard Computers?
- Women employed at Harvard College Observatory from the 1880s to process astronomical photographic plates. The role was classified as computation rather than research, and they were paid substantially less than men doing comparable work.
- What did Annie Jump Cannon discover?
- She created the OBAFGKM stellar spectral sequence still used today, by reorganising earlier alphabetical schemes into an order corresponding to temperature. She classified hundreds of thousands of stars, at times more than three a minute.
- What was Henrietta Leavitt's discovery?
- The period-luminosity relationship for Cepheid variable stars — brighter Cepheids pulse more slowly, in a regular relationship. This made Cepheids standard candles for measuring distance, and became the foundation of the cosmic distance ladder.
- Why is the stellar classification sequence not alphabetical?
- Because Cannon reordered and merged earlier classes, which had been lettered by hydrogen line strength, into a sequence corresponding to temperature. The resulting OBAFGKM order is a fossil of that reorganisation.