In 1925, 24-year-old graduate student Cecilia Payne calculated that the Sun and other stars were composed overwhelmingly of hydrogen and helium—a result so contrary to accepted astronomy that senior astronomer Henry Norris Russell called it “clearly impossible,” leading her to write in her thesis that the enormous abundances she had derived were “almost certainly not real.” Four years later, Russell reached the same conclusion by another method and acknowledged that Payne had got there first.

Cecilia Payne did not discover the composition of the stars by finding a previously invisible element. Every clue she needed was already printed across starlight as a pattern of dark absorption lines. Her achievement was to realise that astronomers had been translating those lines incorrectly.

In 1925, at 24, Payne was working at the Harvard College Observatory toward a doctorate awarded through Radcliffe. Her calculation indicated that hydrogen and helium were not merely present in the atmospheres of the Sun and other stars. They were overwhelmingly abundant, reducing the heavier elements familiar on Earth to a small part of the mixture.

Henry Norris Russell, the influential Princeton astronomer advising on the thesis, told her that such a result was “clearly impossible.” Payne retained the calculation but stepped back from its meaning. Near the end of the thesis, she wrote that the enormous abundances she had derived for hydrogen and helium were “almost certainly not real.”

Four years later, Russell’s own analysis supported hydrogen’s dominance. He acknowledged that Payne’s work had reached the territory first. The history is therefore more complicated than either a clean story of instant scientific correction or a simple story of stolen credit.

SpaceDaily’s earlier account of Payne’s thesis followed the broad discovery and the institutional imbalance surrounding it. This companion piece stays closer to the calculation itself: why spectral lines misled astronomers, why the correct answer looked so implausible in 1925, and what Russell actually added in 1929.

A dark line does not count every atom

Pass sunlight through a spectroscope and its colours are crossed by narrow dark lines. Atoms in the Sun’s outer layers absorb particular wavelengths because their electrons can move between only certain energy states. Each element therefore leaves a recognisable set of marks.

By the early twentieth century, those marks had identified hydrogen, iron, calcium, sodium and many other elements in the Sun. The difficult step was turning a line’s darkness into a reliable estimate of abundance.

A strong absorption line does not necessarily mean that an element is plentiful. It means that many atoms capable of making that particular transition absorbed the relevant light. The number of capable atoms can be a tiny and rapidly changing fraction of the element’s total population.

Temperature can remove electrons from atoms, producing ions with different spectral signatures. It can also redistribute the remaining electrons among energy levels. Pressure and electron density matter too. An element may be abundant yet produce a surprisingly weak line because few of its atoms occupy the required state. A scarcer element can produce a conspicuous line if conditions favour the transition being observed.

Earlier abundance estimates could therefore mistake visibility for quantity. Strong solar lines from elements common in Earth’s crust seemed to reinforce the accepted belief that the Sun and Earth shared broadly similar elemental proportions.

The missing correction came from atomic physics

Payne brought a new interpretive tool to Harvard’s enormous archive of stellar spectra. Indian physicist Meghnad Saha had developed a theory relating the ionisation of atoms to temperature and pressure. Ralph Fowler and Edward Milne recast and extended the approach through statistical mechanics, making it more usable for stellar atmospheres.

Applied carefully, this physics allowed Payne to estimate how many atoms were invisible to a given spectral line because they had been ionised or because their electrons occupied another energy level.

Hydrogen’s familiar Balmer absorption lines provide the decisive example. They are produced when an electron begins in hydrogen’s first excited state. The American Physical Society’s history of Payne’s work notes that in the solar atmosphere only about one hydrogen atom in 200 million occupies that state. If such a tiny fraction can still make a strong line, the total hydrogen reservoir must be immense.

The correction also explained the Harvard sequence of stellar classes, O, B, A, F, G, K and M. A-type stars have especially strong Balmer lines, but not because they alone contain extraordinary stores of hydrogen. At their atmospheric temperatures, a favourable share of hydrogen atoms occupies the state that creates those lines. Hotter stars ionise more of the hydrogen. In cooler stars, more electrons remain in the ground state. The line weakens on either side even when hydrogen remains abundant.

Payne had shown that a classification based on appearance was primarily a temperature sequence. Beneath the visible differences, ordinary stellar atmospheres could possess broadly similar mixtures.

Why the right answer resembled a failed calculation

The modern result can make Russell’s objection look absurd. It did not look absurd in January 1925.

Saha’s theory was only a few years old. Laboratory values for many atomic transitions were incomplete. Stellar pressures and temperatures were imperfectly constrained. Payne repeatedly warned that some of her numerical results should be trusted only in order of magnitude.

The old composition also seemed to work. Astronomers had found a rough similarity between the ordering of strong lines in the solar spectrum and the abundance of elements in Earth’s crust. Arthur Eddington had built successful models of stellar structure using an Earth-like elemental profile. The premise was woven into several lines of reasoning at once.

Payne’s hydrogen value did not propose a modest adjustment. In a January 1925 letter, Russell objected that it was “clearly impossible” for hydrogen to be a million times more abundant than the metals. Here, as usual in astronomy, “metals” means all elements heavier than helium, not only metallic substances such as iron.

Smithsonian historian David DeVorkin’s study of the Payne-Russell episode is useful because it does not flatten the dispute. Russell gave detailed reasons for doubting a method that appeared to generate an extraordinary conclusion. He also supported Payne professionally in other ways. Yet he knew that a senior astronomer could advance a speculative result more safely than a graduate student could, and he advised younger researchers to moderate claims that outran the accepted strength of their methods.

That standard was not neutral in its effect. Payne carried the weaker institutional position, and Russell was the authority evaluating her work. His disbelief shaped what she felt able to say about her own calculation.

The thesis both withdrew and preserved the result

The familiar version says Payne was forced to remove her discovery. The surviving document shows something more subtle.

Her abundance tables remained. In the published text of Stellar Atmospheres, the large values for hydrogen and helium are followed by the statement that their enormous abundance was “almost certainly not real.” Elsewhere, she called the actual values spurious.

The caveat denied confidence without erasing the numerical trail. Anyone reading closely could see that the method had produced a result unlike the terrestrial mixture. DeVorkin argues that this preserved Payne’s priority whether the answer eventually proved right or wrong.

There is an important scientific boundary in the same pages. Payne wrote that her abundance measurements referred to stellar atmospheres and could not by themselves establish a star’s internal composition. The light she analysed escaped from the outer layers. Moving from atmosphere to entire star required additional physics and evidence.

That caution remains worth keeping even though later research supported the broader conclusion. The thesis contained a radical result, but it also recorded exactly where the measurement stopped.

Russell returned with a different route in 1929

Between 1925 and 1929, atomic spectroscopy developed rapidly. Albrecht Unsöld and other researchers improved methods for relating the shape and strength of an absorption line to the number of atoms contributing to it. Russell then undertook a long quantitative analysis of the solar atmosphere.

His paper, On the Composition of the Sun’s Atmosphere, appeared in The Astrophysical Journal in July 1929. Russell calibrated estimates of solar line intensity and calculated abundances for dozens of elements. A recent historical review in Astronomy & Geophysics describes how this more quantitative treatment led him back to high hydrogen and helium abundances.

Russell did not present the finding as though Payne had never existed. He described her thesis as the most important previous astrophysical determination of elemental abundance and said his results agreed remarkably well with hers, including the great abundance of hydrogen.

That acknowledgement closes off one common simplification. Russell did not copy her table, conceal the source and claim the whole discovery as his own. His 1929 study involved substantial independent and confirmatory work with methods that had matured since her thesis.

It does not close off the problem of credit. Russell was the senior figure who had persuaded Payne to distrust the right result. When he later endorsed hydrogen’s dominance, his reputation helped move the conclusion into accepted astronomy. Subsequent writing often cited the 1929 paper while overlooking the earlier thesis, even though Russell himself had pointed readers toward it.

From the solar atmosphere to a hydrogen universe

Payne’s calculation concerned line-forming layers, and Russell’s paper was explicitly about the Sun’s atmosphere. The modern statement that stars are mostly hydrogen and helium rests on a much wider structure assembled afterwards.

Nuclear physics explained how stars use hydrogen as fuel and build helium in their cores. Models of stellar evolution connected composition to luminosity, lifetime and internal structure. Helioseismology tested the Sun’s interior through its oscillations. Solar neutrinos supplied evidence from the fusion reactions themselves.

Modern estimates put the Sun’s visible surface at roughly 74 percent hydrogen and 24 percent helium by mass, with all heavier elements together accounting for only a small remainder. The core has changed as fusion converts hydrogen into helium, so even the phrase “the composition of the Sun” needs a location and an epoch attached to it.

The abundance hierarchy also became a history of matter. Hydrogen was produced overwhelmingly in the early universe. Stars later assembled many heavier nuclei and dispersed them through winds and explosions. SpaceDaily’s recent explanation of where the elements in the human body came from begins from that divide: most hydrogen predates the first stars, while the calcium, iron and other heavy elements required stellar processing.

Payne’s thesis helped reveal how small that processed fraction is. The elements most tangible to us are not the elements that dominate ordinary matter across the universe.

Priority is clearer than the morality play

There is a temptation to make the episode tidier than the record allows. One version turns Russell into a thief. Another treats the four-year delay as nothing more than healthy scientific caution. Neither contains enough of the truth.

Russell’s objection arose from genuine weaknesses in early abundance methods and from an accepted framework that had produced useful results. His later analysis added independent evidence. He cited Payne and acknowledged the close agreement.

At the same time, Payne was a young woman working within institutions that denied women ordinary academic standing. She had calculated the decisive result, and a senior man’s judgment led her to disown its reality in the very document that established her priority. His eventual endorsement carried more authority than her original evidence had been permitted to carry.

The scientific lesson is not that every improbable result should be defended against criticism. Most extreme results do fail. The better lesson is that evidence and expectation must remain separable, particularly when the person supplying the expectation has more power than the person supplying the evidence.

Payne’s tables survived the caveat. Russell’s 1929 paper confirmed the central abundance pattern and directed readers back to her. The record is clear enough: he helped astronomy accept the result, but she had found it first.