Its common knowledge that people used to think that the sun is a ball of fire or molten metal, but when did science start to prove otherwise?

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$
    – called2voyage
    Commented May 26, 2020 at 12:47

4 Answers 4


I think it's maybe not the case that there was a moment when the astronomy community conclusively rejected the ball-of-fire hypothesis; astronomers simply accumulated more and more evidence against it. If you want to put a rough date on it, you could put your finger somewhere in the middle of the 19th century, as by then, other ideas had taken hold.

Back in the classical period, Anaxagoras had proposed that the Sun was a heap of molten metal. I don't know whether this was widely accepted by his contemporaries. The idea of the Sun as a ball of metal or fire certainly persisted for some time, though perhaps largely for lack of any better ideas. We didn't even understand oxygen and combustion until the work of Lavoisier and others in the late 18th century, so detailed calculations were presumably out of the question for a millennium or two after Anaxagoras. I don't know when calculations of how long combustion could sustain the Sun were first done, but it appears to have been not more than several decades after the theory of combustion was developed.

Why? Well, we can say that by the middle of the 19th century, the predominant explanation for the Sun's luminosity was not the burning of coal but instead gravitational potential energy. By the 1860s, it was widely known that chemical reactions could only power the Sun for a few thousand years. We also now had a potentially viable alternative: a decade earlier, Hermann von Helmholtz had begun exploring the idea that gravitational contraction of some sort, by what we now call the Kelvin-Helmholtz mechanism, was the source of energy, with gravitational potential energy being transformed into heat$^{\dagger}$. Around the same time, Lord Kelvin suggested that meteors falling into the Sun provided the necessary energy, a similar mechanism to Helmholtz's. I believe astronomers continued with the contraction hypothesis through the turn of the century - I've seen an article written around 1900 to that effect.

However, during the early and mid- 1900s, quantum theory and nuclear physics were being developed, and the work of Eddington, Bethe and others would lay the groundwork for our current understanding of solar energy production. Previous models (including, finally, Kelvin-Helmholtz contraction) were now known to be insufficient because they allowed the Sun to shine for only thousands or millions of years, and geologists had established that Earth itself was much older than this. Fusion, on the other hand, allows the Sun to survive for billions of years - a timescale that matches up well with the age of the Earth. We also knew that hydrogen and helium were the dominant constituents of the Sun and other stars; while Wollaston and Fraunhofer had performed the first solar spectroscopy observations in the early 1800s, the true composition of the Sun was not accepted for more than a century, when Cecilia Payne made a detailed study of spectral lines.

$^{\dagger}$ While this does produce heat in various bodies, including T Tauri stars, it is not significant in most stars beyond that stage.

  • 4
    $\begingroup$ “it is not significant in most stars” – well it is; it's what's gets fusion going in the first place! It is just by itself not enough to keep the star from cooling down again quickly. $\endgroup$ Commented May 25, 2020 at 7:49
  • 2
    $\begingroup$ @leftaroundabout "Quickly" being relative, of course. It's still enough for hundreds of millions of years, and in the case of objects like white dwarves, it's enough to keep them shining for longer than the Sun is going to spend on the main sequence. That's part of why we didn't search very hard for a better explanation - there was a long... let's say fight... between geologists who've been finding more and more evidence that the Earth's age is on the order of billions of years, and physicists who thought that was preposterous. $\endgroup$
    – Luaan
    Commented May 26, 2020 at 11:25
  • 3
    $\begingroup$ @Luaan A white dwarf cools from a solar luminosity to a tenth of that in less than a few million years. They are "long-lived" at luminosities of $10^{-5}$ to $10^{-6}$ that of the Sun. $\endgroup$
    – ProfRob
    Commented May 26, 2020 at 14:33
  • 2
    $\begingroup$ Just to add to the history here, it actually took a long time before people decided that gravitational collapse alone couldn't sustain the Sun because they had no idea how old it was and similarly also didn't know how old the Earth was (presumably both would be approximately the same age). It took strong advances in geology (some by Darwin himself, in his attempt to prove evolution) before people started believing the Earth and thus the Sun was older than hundreds of millions of years and thus gravitational collapse could not support the entire energy output of the Sun. $\endgroup$
    – zephyr
    Commented May 26, 2020 at 18:16
  • 1
    $\begingroup$ @farhanhubble Anaxagoras was living circa 500 BCE, so I did indeed mean "millennium". $\endgroup$
    – HDE 226868
    Commented Jun 4, 2020 at 14:23

Scientists figured that the sun couldn't be a ball of coal during the industrial age, because given the mass of the Sun, all the coal would have burned out before humans appeared on Earth. But we didn't know what else could be going on.

In 1904, Rutherford suggested radioactive decay as a possible process that could account for the Sun's energy. But it is only after Einstein and the discovery of $E=mc^2$ that they knew that fusion could be going on inside. And indeed, that's what Eddington suggested in 1920.

Finally, in 1925, Cecilia Payne-Gaposchkin suggested that the Sun might be mostly hydrogen.

  • 8
    $\begingroup$ From what I remember, the prevailing theory was that gravitational contraction provided the energy for the Sun, which would only last a few tens of millions of years, despite geologists insisting Earth was hundred of millions or billions of years old. And from what I heard a lot of physicists worked out the details of fusion in stars. $\endgroup$ Commented May 24, 2020 at 16:15
  • 3
    $\begingroup$ Are there old books discussing the ball of coal theory and why it can't hold and what else could be happening? $\endgroup$
    – DKNguyen
    Commented May 24, 2020 at 23:10
  • 4
    $\begingroup$ This is not the right answer. Lord Kelvin spent many years in the 1860s calculating the sun's age based on gravitational potential energy. The answer he got (few tens of millions of years) did not concur with geologists estimates of billions (Lyell et al). This led to an enduring puzzle that was only solved by the discovery of nuclear energy in the 20th century. $\endgroup$ Commented May 25, 2020 at 7:32
  • 4
    $\begingroup$ @OscarBravo: The question wasn't about the discovery of fusion, but about the rejection of fire. This is therefore the correct answer, because the gravitational model does not involve fire. $\endgroup$
    – MSalters
    Commented May 25, 2020 at 8:48
  • 2
    $\begingroup$ @OscarBravo it's a dubious argument that other (essential incorrect) theories such as contraction meaningfully constitute knowledge of "what else could be going on". As you point out, the actual nature of energy and mechanism were not known until substantially later. The question is in fact rather specific about "fire" theories, and while there is another answer which gives a broader view of the evolution of understanding overall, this answer specifically addresses the actual question by providing the basis for rejecting the "fire" idea. $\endgroup$ Commented May 25, 2020 at 17:24

Scientists used telescopes aligned with a prism in the 1860's and found that sunlight is the same continuous bands of color as plasma light from an electric tube lamp, plus some non-ionized elements visible as narrow bands.

Bunsen and Kirchhoff introduced spectroscopy as a lab method in 1860.

A video with images of Bunsen's apparatus and experiments is here:

Here is a summary:

Because the electrons of the sun are not bound to the nucleus on specific orbits, it's light spectrum is a continuous rainbow of all the colors.

It had already been found that the sun had black lines in it's spectrum. i.e. Wollaston from 1802 and Fraunhofer in 1812.

Bunsen wanted to measure the color of elemental flames using "color filters", and Kirchoff suggested the he use a prism instead. They were surprised to find color bands that gave every element a different optical fingerprint. It caused great wonder in the scientific world of the time.

They found that the spectrum of the sun was the same as an ionised plasma, and they found the optical fingerprints of various gas phase elements also in the sun's using telescopes aligned with prism and measuring devices.


It was not proven at the time, but in 1863 a famous article was published on this question.


I don't know if fully quoting is allowed, so I'll quote the first paragraph.

“If the sun were composed of coal, it would last at the present rate only 5,000 years. The sun, in all probability, is not a burning, but an incandescent, body.

  • 3
    $\begingroup$ So did scientists know in the 1860s that coal was compressed vegetable matter? Because if so it brings up the question: did they really think that the sun had previously been a massive ball of forest? $\endgroup$ Commented May 25, 2020 at 12:37
  • 10
    $\begingroup$ @RobinWhittleton coal is a reasonable, widely familiar in 1860 proxy for carbon and chemical fuels in general. Any error there is insignificant compared to the many orders of magnitude discrepancy between durations that would possible and estimated age. $\endgroup$ Commented May 25, 2020 at 17:12
  • $\begingroup$ @RobinWhittleton: Mikhail Lomonosov and Caspar Neumann were aware of the "fossil fuel" origin of coal by the 1750's. Though, it's likely that the 1863 article was just using "coal" (which by then played an important role in industry) as a metonym for "combustible organic chemicals" without assuming a biological origin. $\endgroup$
    – dan04
    Commented Jan 4, 2023 at 1:18

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .