Astronomers just found the smallest galaxy ever

When it comes to the galaxies in the Universe, most of us think about the Milky Way and galaxies similar to it. After all, it is our galactic home, containing hundreds of billions of stars and spanning more than 100,000 light-years across. It’s an interesting fact that galaxies comparable in size to the Milky Way (as well as larger ones) hold the majority of stars present within the Universe today, but that they only represent about ~1% of all galaxies, overall. The majority of galaxies present in the Universe are:

• small,
• low in mass,
• contain very few numbers of stars, overall,
• but are dominated largely by dark matter.

Most of the galaxies in the Universe, because they’re small and contain very few stars within them, are also exceedingly difficult to detect: they’re also ultra-faint galaxies. It takes wide, deep surveys to reveal them at all, and even once they’re imaged, the stars within them need to be measured individually to determine that they’re all at the same distance, and that they’re all gravitationally bound to one another.

Earlier in the 21st century, galaxies like Segue 1 were discovered: with only about 1000 stars inside of it, just a few hundred times as bright as the Sun is, but with hundreds of thousands of solar masses worth of dark matter inside. These represented the smallest, faintest, lowest-mass galaxies known for nearly two decades. Until now, when the galaxy Ursa Major III/UNIONS 1 was discovered and measured. It has a total stellar mass of just 16 solar masses inside: the faintest, smallest, lowest-mass galaxy ever discovered. Here’s what its discovery implies for our understanding of the Universe.

A wide-field view of dwarf galaxy Wolf-Lundmark-Melotte (WLM), alongside with the region that JWST imaged using its NIRCam instrument (inset). The power of JWST to reveal individual stars, even the faint, low-luminosity ones, in galaxies like this one located ~3 million light-years away is poised to set us on a better path toward understanding the star-formation history in our Universe across cosmic time.

To tell the story of the lowest-mass galaxies, we have to understand where galaxies come from overall. The story begins long before any stars first form: back during the epoch of cosmic inflation. During inflation, space isn’t filled with matter or radiation, but rather a form of field energy — the energy of the quantum field that drives inflation — which itself must be inherently quantum in nature. This field energy behaves as though it’s a form of energy intrinsic to space itself, like vacuum energy or a cosmological constant, and causes the Universe to expand relentlessly: doubling in size in all three dimensions with each fraction-of-a-second that elapses, and then doubling again and again and again when that same fraction-of-a-second elapses once more.

Although this stretches the Universe flat and imbues it with the same properties everywhere, inflation does something else, in addition: the quantum field behind it fluctuates. These fluctuations translate into slight differences in energy density: differences at approximately the 1-part-in-30,000 level or so. As the Universe inflates, earlier fluctuations get stretched to larger and larger cosmic scales, while later, newer fluctuations get stretched only to smaller scales. Eventually, inflation ends, the hot Big Bang begins, and the Universe becomes “seeded” with an initial spectrum of overdensities and underdensities: the seeds of cosmic structure.

The fluctuations in the cosmic microwave background, as measured by COBE (on large scales), WMAP (on intermediate scales), and Planck (on small scales), are all consistent with not only arising from a (slightly tilted, but almost-perfectly) scale-invariant set of quantum fluctuations, but of being so low in magnitude that they could not possibly have arisen from an arbitrarily hot, dense state. The horizontal line represents the initial spectrum of fluctuations (from inflation), while the wiggly one represents how gravity and radiation/matter interactions have shaped the expanding Universe in the early stages.

Back before neutral atoms ever form, these fluctuations begin to experience a combination of forces and impulses acting on them.

• The Universe expands, diluting them.
• Gravitation, propagating at the speed of light, pulls matter and radiation into the overdense regions.
• On very small cosmic scales, these fluctuations grow too fast, and radiation streams out of them, causing them to shrink again.
• On larger cosmic scales, these fluctuations won’t begin to grow until the Universe is old enough for a gravitational signal to propagate across these density fluctuations.
• And on the largest scales, or scales significantly greater than the cosmic horizon, they won’t grow at all.

This creates a non-uniform spectrum for density imperfections in the Universe by the time neutral atoms form. On the smallest of cosmic scales, structure is suppressed by these acoustic oscillations: where imperfections grow, shrink, grow again, shrink again, etc., leading to only a low number of very small-scale regions that will eventually gravitationally collapse to form stars. On larger scales, greater numbers of more massive galaxies will form, but only up to a limit. Beyond that scale, fewer numbers will form yet again.

This snippet from a structure-formation simulation, with the expansion of the Universe scaled out, represents billions of years of gravitational growth in a dark matter-rich Universe. Over time, overdense clumps of matter grow richer and more massive, growing into galaxies, groups, and clusters of galaxies, while the less dense regions than average preferentially give up their matter to the denser surrounding areas. The “void” regions between the bound structures continue to expand, but the structures themselves do not.

Only at late times, where larger cosmic structures form — groups of galaxies, clusters of galaxies, the grand cosmic web — and when many smaller galaxies have mutually gravitated and merged together, can we recover a Universe that looks like what we see nearby: 13.8 billion years after the Big Bang. Even though there were large numbers of lower-mass galaxies that were created, initially, the fact that they can fall into larger galaxies should destroy many of them. Only a few small galaxies should persist nearby: largely in the halos of bigger galaxies, like our own, as well as in the space found in between large galaxies and in the deep depths of intergalactic space.

That poses a problem for observers: these are intrinsically faint objects, and they’re also likely to be quite far away. At the distances these galaxies are likely to be located at — tens of thousands of light-years, at minimum — they’ll likely be indistinguishable from stars in the halo of our own Milky Way. They’re faint for two reasons:

• they have very small numbers of stars within them,
• and the stars that are present likely formed all at once, long ago, so that the brightest, most massive ones among them have died, leaving only the faintest, lowest-mass ones behind.

That presents a challenge, but there’s a way around it with good enough observations.

Many nearby galaxies, including all the galaxies of the local group (mostly clustered at the extreme left), display a relationship between their mass and velocity dispersion that indicates the presence of dark matter. NGC 1052-DF2 is the first known galaxy that appears to be made of normal matter alone, and was later joined by DF4 in 2019. Galaxies like Segue 1, however, are particularly dark matter-rich; there are a wide diversity of properties, and the dark matter-free galaxies are only poorly understood.

There are a few things we can measure about stars that aren’t too far away. They include:

• what their distances are (including in 3D space),
• what their star colors and intrinsic brightnesses are,
• and what their metallicities are, or their abundances of heavy elements.

With a little bit of effort (i.e., by taking spectroscopic measurements), we can go a little bit further and measure what their velocities are: how quickly they’re moving, at least along our line-of-sight. If we can acquire all of these measurements for a set of stars in a particular region of space, we should be able to identify whether any groupings of stars are actually part of their own gravitationally bound structure or substructure.

Whereas stars typically exhibit large velocity dispersions — where they move at ~10s of kilometers-per-second relative to one another through space — stars that are bound together will all be moving at roughly the same speed, plus whatever “internal velocities” they have. If they were all born at once, they’ll exhibit similar color-magnitude properties, and will all have almost identical metallicities to one another. Finally, if the overall object itself that they’re a part of is gravitationally bound, we can compare their distance separations with their relative speeds, and infer how much “overall mass” there must be to hold such a structure together. It’s with measurements like these that we’ve been able to identify the smallest known galaxies to date.

Only approximately 1000 stars are present in the entirety of dwarf galaxies Segue 1 and Segue 3, the latter of which has a gravitational mass of an impressive 600,000 Suns. The stars making up the dwarf satellite Segue 1 are circled here. As we discover smaller, fainter galaxies with fewer numbers of stars, we begin to recognize just how common these small galaxies are as well as how elevated their dark matter-to-normal matter ratios can be; there may be as many as 100 for every galaxy similar to the Milky Way, with dark matter outmassing normal matter by factors of many hundreds or even more.

But there’s a catch. Sometimes, there are classes of objects that look like they’re very small galaxies, but they may not actually be. Here are some ways that the Universe can confuse us.

• Sometimes, instead of a small galaxy, we’ll find ourselves looking at a different ancient source of stars: a globular cluster. These ancient remnants were once very large star clusters, often forming stars all at once, but contain no dark matter and often have many of their stars stripped away. Segue 3, once thought to be an ultra-faint galaxy, is now viewed more likely as being a core of an ancient globular cluster.
• At other times, we’ll see stars that did form all at once and are distributed in a large area across the sky, but that’s because what we’re witnessing is an open star cluster in the process of dissociating, and turning into a moving group of stars. These objects are usually younger and higher in metallicity, and include the Hyades open star cluster: the closest star cluster to Earth.
• Or, in rare cases, we can wind up looking at the remnant core of a galaxy that was primarily absorbed and swallowed by a larger galaxy long ago. The dwarf galaxy Messier 32 (also known as NGC 221), orbiting the Andromeda galaxy, is thought to be one such example of an ancient, larger galaxy’s stripped, remnant core.

Whenever we identify a new small, low-mass, galaxy-like object that appears to have its stars distributed in an extended fashion, we have to take care that we aren’t confusing ourselves in one of these fashions.

This wide-angle view shows several satellite galaxies, circled in yellow, around our large neighbor Andromeda. While many of these are indeed satellite galaxies, some of them may be remnants of once-even-larger galaxies, such as Messier 32 (NGC 221), highlighted very close to Andromeda’s plane.

All of which brings us to the newest, latest discovery: galaxy (candidate) Ursa Major III/UNIONS 1. There was a large, deep, wide-field survey conducted at ultraviolet wavelengths called UNIONS, which stands for the Ultraviolet Near Infrared Optical Northern Survey. At a distance of around 33,000 light-years away, a small collection of stars was found all together: with a half-light radius (i.e., a spherical region where half of the light from all the stars within it is contained) of about 10 light-years. It appeared to be consistent with:

• having only about 16 solar masses worth of stars within it, total,
• is extremely low in metallicity, with a heavy element abundance of just ~0.6% of what’s found in the Sun,
• and is consistent with being composed exclusively of extremely old stars: stars more than 11 billion years old.

With these properties, it was clear that this couldn’t be a dissociating open cluster, and was very unlikely to be a stripped core of an ancient galaxy. Either this is a small, faint, low-mass galaxy all unto itself, or it’s the last remaining vestige of an ancient globular cluster: without any dark matter but not having been completely destroyed or dissociated just yet.

Within this dense star field, at left, a few candidate members of an individual structure, either a globular cluster remnant or a very low-mass dwarf galaxy, can be pulled out by sorting the stellar properties of the objects inside of it.

The next key step would be to conduct spectroscopic follow-ups on the individual stellar members that are part of this object, to determine if their properties are consistent (or not) with this being a galaxy all unto itself. This would require not just the type of data available through the UNIONS survey, but would need additional data acquired with a flagship-class ground-based telescope. Using the Keck II telescope and the DEIMOS spectrograph, exactly those observations were undertaken in 2023. After observing a total of 59 stars in that field, they found a significant difference between the properties of the stars associated with the suspected cluster (or galaxy) and all of the other stars in the field.

In the graph below, you can see, on the left, where these stars are located in the sky, and then at center and at right you can see a big difference between the stars that follow the isochrone curve (for stars formed all at once) versus stars that don’t. This teaches us that there are, indeed, two populations here:

• the stars that are members of this grouping, shown in the center panel in blue,
• and the stars that are part of the background of halo stars within the Milky Way, shown in the right panel.

The stars in the vicinity of the identified object Ursa Major III/UNIONS 1 (at left) can be sorted into stars that are identifiable members (or candidate members) by color/magnitude considerations, at center, and distinguish them from Milky Way halo stars in their vicinity, at right.

This is strong evidence that the stars really do form a grouping, but it isn’t quite unambiguous that they’d be members of a galaxy unto themselves, rather than some other type of cluster.

However, the researchers were also able to measure how quickly these objects were moving relative to one another: a key property that requires spectroscopic measurements. What they found was extremely compelling, and is illustrated in the figure below. Whereas most of the stars in the field were not shown to be members of this cluster, whatever its nature happens to be, a few (shown in the top-left panel, below, with blue and yellow markers) objects appear to be moving with the same heliocentric velocity, or the same direction and speed with respect to the Sun.

This figure from the study of the object Ursa Major III/UNIONS 1 identifies a large number of stars within the field of this object and measures their heliocentric velocities. The small cluster of points in the lower-right panel is key for identifying individual stellar members.

In the lower-right panel, you can see the profound difference between the non-member stars, shown with black circles and one with a red x, versus the member stars, all clustered together with the same heliocentric velocity. Altogether, there are only 11 stars that they were able to confirm are indeed spectacularly consistent with being member stars of this cluster, but that when you take them together and analyze them, you find an extremely small velocity dispersion within them: of about 3.7 km/s. For comparison, the stars in the next-smallest galaxies, like Segue 1, have a velocity dispersion that’s nearly 10 times as large.

The last clue, only hinted at in the study, is that we can attempt to reconstruct what the orbit of this object is, and attempt to see whether it’s more likely that this was an accreted galaxy or whether it was either a disk or halo globular cluster. The data is limited, but what is present favors the accretion scenario, leading the authors to conclude that this is most likely the smallest, faintest galaxy ever discovered, rather than a tiny remnant of a once-larger globular cluster.

Shown in this graph are all the known satellites of the Milky Way: dwarf galaxies, globular clusters, and a third population that is ambiguous. The newest object in this set, Ursa Major III/UNIONS 1, is highly ambiguous, forming an outlier from all three populations.

However, we cannot be certain based on this data, alone. What we would really love to do is to have even better data from the stars in the vicinity of this object, going to even fainter magnitudes. Although 11 member stars were identified, there are likely many more (roughly 60, overall, are estimated); they’re just extremely faint.

• Where are they located in 3D space?
• What are their velocities relative to the other stars in this cluster?
• Are there binary stars in this system, and does that influence the velocity dispersion?
• And finally, will this reveal the lowest-mass dark matter halo ever found, or will there be none at all, indicating this object’s nature as originating from an unusual globular cluster?

With dedicated follow-up observations, we should be able to answer these questions. However, it’s also important to note that there are many more “candidate faint galaxies” sure to come from new surveys like the NSF’s Vera Rubin observatory and ESA’s Euclid space telescope, and they will help reveal where the lines are drawn between globular clusters, which contain no dark matter, and accreted galaxies, which are dark matter-rich. With such a small number of stars and such a small velocity dispersion, it’s impossible to conclude that Ursa Major III/UNIONS 1 is definitely a galaxy. However, it might be. If it is, our understanding of the small-scale Universe, and of its implications for dark matter, may be about to take an astronomically great leap forward.