Through a telescope here on Earth, it looks like any other star. But an ultra-compact dwarf might be made up of two or three billion stars, millions of light-years from the Milky Way and grouped together so closely that they appear as a single point of light.

For hundreds of years, no one could tell that these “stars” weren’t single objects. Then, around the end of the twentieth century, scientists built powerful spectrographs that let them conduct surveys of all the frequencies of light coming from some of the universe’s brightest, most dense clusters of galaxies. Among these galaxies, they discovered the groups of stars that they now call ultra-compact dwarfs. The “dwarf” part is shorthand for “dwarf galaxy,” but astronomers have been debating for a decade about whether ultra-compact dwarfs are really galaxies or just super-dense clusters of stars. But UNC astronomers Mark Norris and Sheila Kannappan say they have the answer now.

The trouble with ultra-compact dwarfs, Norris says, was that they didn’t fit into the way astronomers had understood galaxies and groups of stars for several hundred years. There were supposed to be two things: galaxies and globular clusters. Galaxies were massive systems of stars and other matter many thousands of light-years wide. Globular clusters were balls of stars maybe a hundred light-years wide, orbiting way out from the main mass of a galaxy. They sprang up around galaxies billions of years ago, when the galaxies were young.

“A globular cluster is about a million times the mass of the sun,” Norris says. “And galaxies are at least ten or a hundred times more massive than globular clusters. There was this clear size gap between the two.” Ultra-compact dwarfs messed up that picture. They weren’t anywhere near as big as galaxies, but they were more massive and took up much more space than the biggest globular clusters. So what exactly were they?

Some people thought ultra-compact dwarfs were just globular clusters that were unusually large, maybe because they had merged with other clusters when they were young. Others suspected that the dwarfs were remnants of galaxies; they’d been larger once, but then they were drawn too close to other galaxies and had most of their stars ripped away by gravity. “It’s like the galaxy was a dandelion puff someone blew on,” Kannappan says. “All you’re left with is the seed head in the middle.”

Astronomers had been leaning toward what seemed like the simpler explanation, Kannappan says: that ultra-compact dwarfs were giant globular clusters. But she wasn’t so sure. Part of the reason why the dwarfs seemed like globular clusters was that they were the right age: eight or ten billion years old, about the same as their neighboring galaxies. The problem was, the galaxy clusters where astronomers have been studying ultra-compact dwarfs stopped growing a long time ago—everything in those regions is eight or ten billion years old.

So Norris started searching outside the denser areas of the universe. He went through archives of Hubble Space Telescope data, looking—literally eyeballing each telescope image—for ultra-compact dwarfs around galaxies that weren’t near lots of other galaxies. If these dwarfs were about the same age as their host galaxies, they might be giant globular clusters. But if an ultra-compact dwarf were a totally different age from its neighbor galaxy, that would be good evidence that it had been born as its own galaxy and was pulled into another’s orbit later on.

Norris found some candidates that no one had identified as ultra-compact dwarfs before. They look like stars when seen through a telescope on Earth, but in images from Hubble they appear with a very un-star-like fuzzy halo. He and Kannappan used data from the SOAR telescope in Chile, controlled from UNC, to confirm that the fuzzy stars near two large galaxies were really groups of stars in the size range between globular clusters and galaxies.

Click to read photo caption. Mark Norris

By chance, it turned out that another research team had high-quality spectroscopy data on one of these dwarfs, taken when the group was studying a large galaxy nearby. The spectrum of light coming from this ultra-compact dwarf shows that it’s about 3.4 billion years old—billions of years younger than the galaxy, NGC 4546, that it hangs out next to. This dwarf, Norris says, is the first clear example of an ultra-compact dwarf that’s a former galaxy. Not only is it too young to be a globular cluster, it’s also far more massive than clusters normally are around a galaxy the size of NGC 4546. And the ultra-compact dwarf is near a stream of gas extending into the host galaxy but rotating in the same direction as the dwarf—a remnant of the matter that got torn away when the smaller galaxy came too close to the larger one.

Around the other galaxy, Norris and Kannappan found ultra-compact dwarfs that have the right age, size, density, luminosity, and other traits to be giant globular clusters. The next step is to look deep into both kinds of dwarfs to find the differences their analysis suggests must be there. “Even though there are two ways of making what looks to us like almost the same object, there’s a big difference between them,” Kannappan says. “One type of ultra-compact dwarf is, at heart, a galaxy. It’s had a lot its outer stars torn off, but most of its dark matter won’t have been torn off. A globular cluster won’t have that.”

Dark matter is still mysterious to scientists, but they know well enough how to tell that it’s there. “The way it was first discovered was that the orbital motions of stars and gas in galaxies were moving too fast to hold together given the amount of mass you could see,” Kannappan says. “This would be the exact same thing. We’ll look to see whether the internal motions of these ultra-compact dwarfs are too fast to hold together given the matter we can see.”

Until not long ago, trying to detect the orbital motions in an object so small, especially one so close to a big, bright galaxy, was hard to do with anything less than one of the world’s biggest telescopes. But they’ll be able to do it with the smaller SOAR Telescope, thanks to its powerful spectrograph built by Carolina astrophysicist Chris Clemens. (See Endeavors, Fall 2004, “Star Power.”)

They’re also going back to the Hubble archives to identify more ultra-compact dwarfs and sort out how many were once galaxies and how many are globular clusters. Knowing what ultra-compact dwarfs are will help astronomers learn the histories of galaxies, Kannappan says. When there are a lot of globular clusters around a galaxy, it’s evidence that the galaxy was once two large galaxies that merged, compressing clouds of gas as they collided and triggering star formation.

The fact that some ultra-compact dwarfs are globular clusters while others are former galaxies should stop some of the back-and-forthing between scientists.

Says Norris: “It’s nice to be able to tell them, ‘You can stop arguing, because both of you are right.’”

Mark Norris is a postdoc and Sheila Kannappan is an assistant professor, both in the Department of Physics and Astronomy in the College of Arts and Sciences. Their paper, “The Ubiquity and Dual Nature of Ultra-Compact Dwarfs,” has been accepted for publication by Monthly Notices of the Royal Astronomical Society and is available on Kannappan’s website at