Astronomy › Section T · Galaxies and quasars › Topic T.1
The discovery of galaxies
A century ago, many astronomers thought the Milky Way was the whole universe. Measuring the distance to one fuzzy spiral in Andromeda showed that it is one galaxy among billions.
Smudges among the stars
Through a telescope, most of the night sky resolves into sharp points of starlight, but scattered among them are faint, fuzzy patches. Astronomers once called all of these nebulae, from the Latin for "clouds". In the 1770s and 1780s the French comet hunter Charles Messier listed about a hundred of them so that he would not mistake them for comets, and William and Caroline Herschel later catalogued thousands more. Some proved to be star clusters and some glowing gas, but many stayed a mystery. In 1845 the Earl of Rosse, using a 1.8-metre telescope in Ireland, saw that some of these nebulae, such as M51, have a spiral shape.
Island universes?
What were the spirals? In 1755 the philosopher Immanuel Kant proposed that the Milky Way is a flattened disc of stars and that some nebulae might be whole star systems like ours, so distant that their stars blur together. The idea later became known as "island universes". The rival view was that spirals were small, nearby objects inside a Milky Way that made up the entire universe, perhaps solar systems in the making. Both pictures fitted what telescopes showed, because nobody could measure how far away a nebula was. Without a distance, a small cloud nearby and a huge star system far away can look exactly alike.
The Great Debate
By 1920 the question had become pressing. On 26 April 1920, at a meeting of the National Academy of Sciences in Washington, DC, Harlow Shapley and Heber Curtis presented opposing views in what became known as the Great Debate. Shapley had used globular clusters to show that the Milky Way is enormous, about 300,000 light-years across by his estimate, with the Sun far from its centre. In a galaxy that large, he argued, the spirals fit comfortably inside. Curtis argued for a much smaller Milky Way and for spirals as separate galaxies far beyond it. Nobody won outright: the evidence needed to settle the matter did not yet exist.
The evidence on each side
Shapley pointed to Adriaan van Maanen's report that some spirals could be seen rotating over a few years, impossible if they were huge and remote. He also cited S Andromedae, a star that flared in the Andromeda nebula in 1885 so brightly that it seemed impossibly powerful at a great distance. Curtis replied that ordinary novae found in Andromeda looked far fainter than novae in the Milky Way. He explained why spirals avoid the band of the Milky Way: dust in our galaxy's disc hides them. Van Maanen's rotations later proved to be measurement errors, and S Andromedae was a supernova, far brighter than any nova.
Worked example: the nova argument
Follow Curtis's reasoning with round, illustrative numbers. A nova in the Milky Way at a known distance of 2,000 light-years peaks at apparent magnitude 4, and a nova of the same kind in a spiral nebula peaks at magnitude 14. How far away is the nebula? 1. Difference: \(14 - 4 = 10\) magnitudes. 2. As The brightness of stars showed, 5 magnitudes is a factor of 100 in brightness, so 10 magnitudes is \(100 \times 100 = 10^4\). 3. Brightness falls as \(1/d^2\), so the distance ratio is \(\sqrt{10^4} = 100\). 4. Distance: \(100 \times 2000 = 200{,}000\) light-years. 5. Interpret: the answer rests on both novae having the same peak luminosity, exactly the assumption S Andromedae seemed to break.
Hubble finds a Cepheid
The decisive tool was the 2.5-metre (100-inch) Hooker telescope on Mount Wilson in California, completed in 1917 and then the largest in the world. Edwin Hubble used it to photograph the Andromeda nebula, M31, and could pick out individual stars in its outer parts. On a plate taken in October 1923 he found a star that brightened and faded with a period of about 31 days: a Cepheid variable. As Variable stars as distance markers explained, a Cepheid's period reveals its luminosity through Henrietta Leavitt's period-luminosity relation, and comparing that luminosity with its apparent brightness gives its distance. Hubble had found a standard candle in a spiral.
Worked example: a Cepheid in Andromeda
Use the calibration from Variable stars as distance markers,
for a Cepheid with a period of 31.4 days and an average apparent magnitude of 19.0, ignoring dust. 1. Logarithm: \(\log_{10} 31.4 = 1.497\). 2. Absolute magnitude: \(M_V = -2.43(0.497) - 4.05 = -5.26\). 3. Distance modulus: \(m - M = 19.0 - (-5.26) = 24.26\). 4. Distance: \(d = 10^{(24.26 + 5)/5} = 10^{5.852}\), about \(7.11 \times 10^5\) pc or 2.32 million light-years. 5. Check: the modern distance to Andromeda is about 2.5 million light-years, so one star comes within about 7%, and the answer is many times larger than Shapley's 300,000-light-year galaxy.
The universe grows
Hubble's paper was presented to the American Astronomical Society on 1 January 1925. He placed Andromeda about 900,000 light-years away, beyond even Shapley's Milky Way, so the spirals had to be separate galaxies. His number was less than half the modern value: in 1952 Walter Baade showed that two kinds of Cepheids had been mixed up, and dust also dims the stars. But the conclusion stood. Each side of the Great Debate was partly right: Shapley about a huge Milky Way with the Sun far from its centre, Curtis about the spirals. Extragalactic astronomy had begun.
Galaxies everywhere
Today the word nebula is kept for clouds of gas and dust, and the spirals are called galaxies. Andromeda, about 2.5 million light-years away, is the nearest large spiral. Deep images show how crowded the universe is: the Hubble Ultra Deep Field, a patch of sky about a tenth of the full Moon's width, contains about 10,000 galaxies. Scaling such counts to the whole sky gives estimates of a few hundred billion to about two trillion galaxies in the observable universe, depending on how many faint dwarfs exist. As of 2026, the most distant spectroscopically confirmed galaxy, MoM-z14 (reported in 2025), is seen as it was about 280 million years after the Big Bang.
Common mistakes to avoid
First, thinking the Great Debate was won on the night; the question was settled by Hubble's Cepheid work of 1923 to 1925. Second, assuming a nebula's apparent size or shape reveals its distance; without a standard candle, a small nearby cloud and a giant remote galaxy can look alike. Third, crediting Hubble with discovering that Cepheids measure distance: the period-luminosity relation was Leavitt's. Fourth, quoting Hubble's 900,000 light-years as Andromeda's distance; the modern value is about 2.5 million. Fifth, thinking Shapley was wrong about everything; his large galaxy with an off-centre Sun was correct.
Key takeaways
Fuzzy nebulae could be nearby clouds or distant star systems, and only a distance could decide. Kant suggested in 1755 that some are other Milky Ways. In the Great Debate of 1920, Shapley argued for one huge galaxy containing the spirals and Curtis for spirals as separate galaxies. In 1923 Hubble found a Cepheid in Andromeda with the 2.5-metre Mount Wilson telescope, and its period-luminosity distance put the nebula far outside the Milky Way. His 1925 distance of about 900,000 light-years was too small; the modern value is about 2.5 million. The observable universe holds hundreds of billions of galaxies or more.
Flashcards
Flashcard
Any fuzzy patch of light that is not a sharp star
Test your understanding
Question 1 of 5
Before the 1920s, why could astronomers not decide whether spiral nebulae lay inside the Milky Way?
Question 2 of 5
What did Hubble find in the Andromeda nebula in 1923 that settled the question?
Question 3 of 5
A nova in a spiral nebula peaks 10 magnitudes fainter than an identical nova in the Milky Way 3,000 light-years away. How far away is the nebula?
Question 4 of 5
Hubble's 1925 distance to Andromeda, about 900,000 light-years, was too small. Why?
Question 5 of 5
Which statement about the Great Debate of 1920 is accurate?
