March, 1929 Hubble Expands the Universe

This article was originally written for publication one year when there wasn’t a current seasonal item suitable for a March article.

March, 1929

March is the anniversary of a historic publication that was a major turning point in Astronomy: Edwin Hubble’s landmark paper[5] showing that distant galaxies are receding from us, and that recession speed increases with distance. Although this was arguably the most important of Hubble’s works, he made many other important contributions in the years leading up to it, especially in one of the major topics of twentieth-century Astronomy: measurement. Developing methods to estimate distances to remote objects, and using the evolving measurement techniques to determine the size and large-scale structure of our universe was the major focus of early 20th century astronomy.

In this article we’ll review some of the events that lead up to our current model of the size of the universe and our place in it. The story provides an excellent example of how science progresses through a combination of independent research, development of theory, debate, and correction of error.

As is often the case, different groups of astronomers were pursuing, at approximately the same time, streams of research that, at first, seemed unrelated. Later, relationships between their results were discovered, and the combined research enabled new and profound conclusions to be reached.

The Spiral Nebulae Mystery

In the late 19th and early 20th century, astronomers were studying nebulae—the wispy patches of light they could see with their telescopes. The nature of these nebulae, and whether there was more than one kind and, if so, how they were grouped, was a subject of debate. It became clear that there was more than one kind of nebula in 1864 when astronomers began to take their spectra. Some of them had “line spectra” typical of glowing gas, but others had “continuous spectra”, typical of starlight[7]. Were there two or more kinds, one gaseous, and one somehow related to stars?

Roberts 1888 M31 image

The study became more interesting when, twenty years later, long-exposure astrophotography began to reveal details of the structure of the nebulae. Reacting to the spiral structure visible in the first detailed photograph of M31, the Andromeda nebula, Roberts  wrote, “No verbal description can add much to the information which the eye at a glance sees on the photograph. . . Here we (apparently) see a new solar system in the process of condensation from a nebula — the central sun is now seen in the midst of nebulous matter which in time will be either absorbed or further separated into rings.”[12] Lacking any way to measure the distance to the object, he assumed it was relatively nearby, and that he had captured a new solar system in the process of forming from material orbiting a young star.

Another puzzling datum was added as astronomers developed techniques to measure the radial velocity of objects (their speed toward or away from the Earth) by measuring the doppler shift of their spectral lines. When this technique was first applied to M31, the Andromeda nebula, the result was surprising: a measured radial velocity of 300 kilometres per second — the highest ever measured[14]. Again, not realizing the object’s great distance, the cause of this high velocity was not interpreted correctly, but it seemed clear that the spiral nebulae were some special new class of object.

Galactic Distance and Scale

Measuring Distance

In the 1914 Journal of RASC, J.C. Kapteyn summarized the problem of making further progress on modeling the universe as one of determining distances, writing, “From the moment that the distances become known we shall be able to make a model which will be a true representation of our stellar system.”[8] At that time the only accurate way to measure stellar distances was the parallax method, which notes the small movement of a star against the fixed background when observed from opposite sides of the Earth’s orbit around the sun. This method, unfortunately, was usable only for relatively close distances (about 500 parsecs, or 1600 light years, about 2% of the diameter of the Galaxy).

When Kapteyn wrote this, a critical new distance measurement tool had already been discovered but its importance had only recently been suggested, and it had not yet been calibrated.

Leavitt’s 1912 Cepheid P-L graph

In 1908 Henrietta Leavitt published a study of almost two-thousand variable stars. Her paper included the offhand remark, “it is worthy of notice that. . . the brighter variables have the longer periods.”[9] This observation was expanded in her next paper in which she studied 25 variable stars of a specific character, writing, “a remarkable relation between the brightness of these variables and the length of their periods will be noticed.”[10]

These Cepheid Variables are also easy to detect because of the characteristic shape of their light curve — quickly gaining in brightness then smoothly dimming over a longer period. They would provide the tool to accurately measure objects at significantly greater distances. By measuring a Cepheid’s period, an astronomer could determine its absolute magnitude. Then, by comparing this to its apparent magnitude, the distance could be calculated, since a star’s apparent brightness decreases as a function of distance. Cepheid Variables were to become one of the most important tools for measuring distances, and the tool that ended a long-running debate on the nature of the Galaxy.

The Great Debate

Around 1920, a debate was raging in the astronomical community regarding the size of the galaxy and our place in it. Related to this debate was the question of the nature of the spiral nebulae.

Did we live in a large galaxy that comprised the entire universe, encompassing everything we can see, including the spiral nebulae? Or were the spiral nebulae other galaxies like our own, implying a vast universe in which we occupied no special position? And whether there was one Galaxy or many, where were we located with respect to our Galaxy — in some special position such as the centre?

These questions were the subject of a debate held at the National Academy of Science in 1920[11] .

Harlow Shapley took the position that the universe consisted only of our Galaxy, which was very large — about 300,000 light-years in diameter. The spiral nebulae, while distant, were still part of our galaxy.[13] He argued the spirals could not be separate galaxies because, to appear so small, they would have to be very far away — so remote that the novae occasionally observed in them would have to be unimaginably bright [which they were — at that time astronomers did not know about supernovae, which are thousands of times brighter than novae].

Shapley’s arguments were based on determining the distances to a variety of remote objects, especially globular clusters. His distance measurements were based on a variety of techniques but depended most heavily on the use of the newly-discovered Cepheid Variables. Assuming the globulars were distributed evenly around the centre of the galaxy, and observing that there seem to be more of them in one area of our sky, he determined that we are not in the centre of the galaxy, and correctly estimated its direction. As we shall see, however, his distance estimates were quite inaccurate.

Heber Curtis argued for a smaller galaxy — about 30,000 light-years in diameter — that was one of a vast number of similar systems. The spiral nebulae, he said, were separate star systems similar to our own galaxy, and at great distances “from 500,000 to 10,000,000 light-years away”[2]. Curtis noted that the spiral nebulae had unique properties such as high radial velocities and star-like spectra, and he explained the fact that they appear primarily above and below the galactic poles by proposing that our own galaxy contains some kind of “obscuring material” in the disk. While correct on these important points, Curtis’ distance estimates were too low, and he incorrectly concluded we are near the centre of our galaxy.

Both astronomers were mislead by a not-yet-discovered aspect of Cepheid variables: that there are two populations of Cepheids, with different period-luminosity relationships. This made Shapley over-estimate distances by a factor of 1.5 times, and made Curtis find Cepheids an unreliable yardstick. (We now use a figure of approximately 100,000 light-years for the diameter of our galaxy.)

The event did not include announcing a “winner”. Curtis was generally thought to have the stronger case, but both participants made important discoveries, and were right (and wrong) on key subjects.

Solving the Spiral Mystery

The ultimate solution to the debate, and to the nature of the spiral nebulae, brings us back to Hubble.

Hubble noting Cepheid variables on image of M31

In 1923 Hubble was studying novae in M31, the Andromeda nebula. He thought he had identified 3 novae when further study made him realize that one of them was not a nova after all, but a Cepheid variable. Using Leavitt’s period-luminosity relationship, he was able to calculate the distance to this star and, thus, to the nebula in which it was contained. The result placed M31 at a great distance, far outside even Shapley’s large estimate for the size of our galaxy.

Hubble wrote to Shapley, “You will be interested to hear that I have found a Cepheid variable in the Andromeda Nebula. . . “, and Shapley is said to have remarked to a colleague, “Here is the letter that has destroyed my universe.”[1]

To his credit, Shapley responded in the manner of a true scientist. He accepted that the data rendered his previous hypothesis incorrect, accepted the new findings, and worked for many years to further improve our understanding of the now much-larger universe. Hubble improved his data on the distance of spiral nebulae in several papers published over the next few years. He calculated the distance to spiral NGC 6822 as 700,000 light-years and called it “the first object definitely assigned to a region outside the galactic system”[3]. He then published a distance of 860,000 light-years for spiral M33[4], and finally officially published his figures for M31, with a distance of 900,000 light-years[6].

Note that all of these distances are low, by a factor of 2 or more. The current measures for the distances to NGC 6822, M33, and M31, are 1,800,000, 3,000,000, and 2,500,000 light-years, respectively. Measurements in the 1920s were small because of the unknown dual population of Cepheid variables, and insufficient allowance for the dimming of distant starlight by interstellar dust. However, even those low distances were sufficient to establish the spiral nebulae as separate star systems, and to show the universe to be of immense size.


Hubble’s landmark paper, published in March, 1929, and which we commemorate with this March article, built on the new science of measuring distances to remote galaxies. He accurately calculated the distance to 7 remote galaxies, estimated distances to a dozen more, and calculated their radial velocity using their spectra.

Hubble’s graph of distance versus velocity. (Hubble 1929)

His surprising discovery was that most galaxies are rapidly speeding away from us and that their speed of recession was directly proportional to their distance. This result was so surprising that he felt a need to defend it, saying, “for such scanty material, so poorly distributed, the results are fairly definite”[5].

Hubble’s initial figure for the speed of recession as a function of distance was 500 km/s/Mpc. That is, for each million parsecs distant, an object’s speed of recession is 500 kilometres per second. Over the next eighty years, many studies were used to refine this Hubble Constant that describes the expansion of the universe, and the accepted value is converging to something around 65 km/s/Mpc.

Open or Closed?

Since Hubble’s surprising discovery, astronomers have studied the speed of expansion, and tried to estimate the total mass of the universe, in an attempt to answer what seemed like a simple question.

All the matter in the universe is expanding outward. However, the gravity from the mass of that matter must be pulling inward on it at the same time, so the expansion must be gradually slowing. If there was enough mass (more than a certain limit), gravity would eventually overcome expansion, and the universe would begin to shrink, collapsing back in on itself. Is there enough mass for the expansion to reverse, or insufficient, so it will go on forever?

The surprising answer, only recently discovered, was “neither”. Studies of supernovae at extreme distances (and thus of light originating deep in the past) show that the expansion rate was slower in the past than today. In other words, the expansion is not slowing, it is accelerating.

This is a shocking discovery.

Acceleration requires a driving force. What energy, never before detected, is powering the accelerating expansion of the universe? The nature of this Dark Energy is one of the great puzzles of modern Astronomy, as was the nature of the “Spiral Nebulae” a century ago.


We have used the March anniversary of Hubble’s important paper describing the expansion of the universe to trace the events leading up to that discovery, highlighting how the development of ways to measure astronomical distances was one of the major accomplishments of early 20th century Astronomy. The story also shows how the process of scientific research progresses, with parallel research and theories competing, challenging, assisting, validating, and continuously improving one another, always supported most strongly by the evidence provided by the best available observation and measurement techniques.

Image Credit

M31; Bob Olson, Ottawa RASC

The beautiful image of M31 that lead to this article from the RASC site’s home page, and that serves as a logo for this page, was taken by RASC Ottawa Centre member Bob Olson.


AIP, From Our Galaxy to Island Universes, Cosmology. (1923). [Source]
H. Curtis, The Scale of the Universe, Bulletin of the National Research Council. 2 (1921) 194.
E. Hubble, NGC 6822, a remote stellar system, Astrophysical Journal. 62 (1925) 409–433.
E. Hubble, A spiral nebula as a stellar system: Messier 33, Astrophysical Journal. 63 (1926) 236–274.
E. Hubble, A Relation between Distance and Radial Velocity among Extra-Galactic Nebulae, Proceedings of the National Academy of Science. 15 (1929) 168–173.
E. Hubble, A spiral nebula as a stellar system, Messier 31, Astrophysical Journal. 69 (1929) 103–158.
W. Huggins, W.A. Miller, On the Spectra of Some of the Nebulae, Philosophical Transactions Series. 154 (1864) 437–444.
J.C. Kapteyn, On the Structure of the Universe, Journal of the Royal Astronomical Society of Canada. 8 (1914) 145-.
H. Leavitt, 1777 variables in the Magellanic Clouds, Annals of Harvard College Observatory. 60 (1908) 87–108.
H. Leavitt, E.C. Pickering, Periods of 25 Variable Stars in the Small Magellanic Cloud, Harvard College Observatory Circular. 173 (1912) 1–3.
NRC, NRC Transcripts of the “Great Debate,” Bulletin of the National Research Council. 2 (1921) 171–217.
I. Roberts, Photographs of the nebulae M 31, h 44, and h 51 Andromedae, and M 27 Vulpeculae, Monthly Notices of the Royal Astronomical Society. 49 (1888) 65-.
H. Shapley, The Scale of the Universe, Bulletin of the National Research Council. 2 (1921) 171.
V.M. Slipher, The radial velocity of the Andromeda Nebula, Lowell Observatory Bulletin. 2 (1913) 56–57.

Leave a comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.