Hubble Space Telescope image showing the globular cluster stellar stream. (Credit: Hubble Space Telescope and Holm et al. (2026))
In a Nutshell
- Astronomers have found the first known extragalactic globular-cluster stellar stream, located about 115 million light-years away.
- Using the stream’s shape, researchers placed the first stream-based constraint on the dark matter halo surrounding its host galaxy.
- Future telescopes could find many more such streams and help scientists measure dark matter mass and density profiles in distant galaxies.
For decades, astronomers have watched stars peel away from star clusters inside the Milky Way, forming long, faint ribbons called stellar streams. These structures carry clues about the invisible dark matter that shapes galaxies. Astronomers expected these streams to exist around other galaxies, but until now they had only been observed in the Milky Way.
A team of astronomers has found what appears to be the first known extragalactic globular-cluster stellar stream, located about 115 million light-years away. Reported in the journal Nature, the ribbon of stars sits inside an ultra-diffuse galaxy called UGC 9050-Dw1 and was spotted using the Hubble Space Telescope. It is so thin and so remote that its detection highlights just how faint a structure modern telescopes can now pick out. The researchers named it Oyashio, after a cold Pacific ocean current.
Oyashio matters scientifically because it reveals what we know about dark matter. By analyzing the stream’s shape, scientists estimated the mass of the dark matter surrounding UGC 9050-Dw1, a type of galaxy that has puzzled researchers for years because its dark matter content has been nearly impossible to measure directly. This is the first stream-based constraint on the dark matter halo of an ultra-diffuse galaxy.
Why Ultra-Diffuse Galaxies Hide Their Dark Matter
UGC 9050-Dw1 belongs to a class known as ultra-diffuse galaxies. Despite spanning a physical extent comparable to the Milky Way, these galaxies contain far fewer stars and glow so faintly that they were only widely recognized as a distinct category in recent decades. Their dim appearance makes traditional measuring techniques challenging. Astronomers struggle to track how fast stars move inside them or map how their gravity behaves, which are the usual methods for figuring out how much dark matter a galaxy holds.
Stellar streams offer a workaround. When a host galaxy’s gravity slowly pulls stars away from an orbiting star cluster, a densely packed ball of stars, those stars spread into a long, thin arc that traces the gravitational environment around them. Because these clusters are relatively small and tightly bound, the streams they produce are narrow and sensitive to subtle variations in the surrounding dark matter. A string dragged through a river bends and stretches in ways that reveal the current beneath. Stellar streams work on the same principle, just on a galactic scale.
How Astronomers Found the Oyashio Stellar Stream
Researchers identified Oyashio first by eye in Hubble Space Telescope images taken in September 2022, then confirmed it independently using data from the Canada-France-Hawaii Telescope. Seeing the same feature in two entirely different datasets, processed separately, ruled out the possibility of a camera glitch or processing error.
Oyashio measured about 6,500 light-years long, with a width of roughly 72 parsecs, an extraordinarily thin ribbon given its length. (A parsec is a unit of distance used by astronomers, equal to about 3.26 light-years.) That narrow width was major evidence. Known dwarf-galaxy streams are much wider; the Orphan–Chenab stream, for example, is wider than 200 parsecs. Oyashio’s slender profile, taken together with its color and overall geometry, pointed strongly to a globular cluster as the source, not a small galaxy being torn apart.
To confirm the connection between the stream and a nearby compact star cluster visible in the image, the team compared their colors. Both registered nearly identical colors, which is exactly what scientists expect when a stream and its parent cluster formed from the same stellar population. Both also matched the color profile of other star clusters already cataloged inside UGC 9050-Dw1.
Researchers also systematically tested alternative explanations. A tidal shell from a radial collision would produce a curved arc with a differently offset center of curvature, and Oyashio’s geometry does not match that pattern. A background galaxy bent into an arc by gravity was considered, but no other telltale signs of such bending were found nearby. A dust cloud could create brightness fluctuations, but no color variation consistent with dust was found on either side of the feature. Computer simulations showed that a dwarf-like stream would need to be about five times more massive in stars to be visible, and it would still appear wider than Oyashio. None of the alternatives held up, though the authors note that chance alignment or projection cannot be entirely ruled out with the current data.
Using a Stellar Stream as a Dark Matter Scale
With the stream’s identity as well-established as current data allow, researchers used a computational tool called X-Stream, which fits mathematical models to the observed shape of a stream to infer properties of the surrounding dark matter. By testing possible stream models, the program identified which combinations could reproduce Oyashio’s observed shape.
Those models place an upper limit on the original cluster’s mass of less than 2.5 million times the mass of the Sun, at 95% confidence. For scale, matching Oyashio’s brightness in the models required a cluster roughly 20 times as massive as Palomar 5, a well-known globular cluster in the Milky Way with its own stream.
Modeling also points to a massive dark matter halo surrounding UGC 9050-Dw1, though these results are model-dependent and carry significant uncertainty. With those caveats in mind, the credible range of total halo masses the models favor points toward values comparable to estimates for the Large Magellanic Cloud, a satellite galaxy of the Milky Way. This modeled mass also aligned well with earlier estimates made by simply counting the galaxy’s star clusters.
Simulations further suggested the halo may be slightly less cored than some other ultra-diffuse galaxies studied. The internal shape of a dark matter halo carries clues about what dark matter actually is, so even this subtle detail matters to physicists trying to pin down the nature of the universe’s most mysterious substance.
What Comes Next for Dark Matter Research
Deeper observations with Hubble or the James Webb Space Telescope could strengthen the case for Oyashio’s origin. Spectroscopic follow-up with facilities such as the Keck Observatory could compare the stream with its presumed parent cluster, looking for similarities and differences in their stellar populations.
Looking further ahead, upcoming space telescopes including Euclid and the Nancy Grace Roman Space Telescope are expected to dramatically expand the search. If globular cluster streams exist in ultra-diffuse galaxies at anything like the rate predicted by galaxy models, surveys with these instruments could turn up many more.
At about 115 million light-years away, Oyashio shows that the invisible scaffolding of dark matter can be mapped with a ribbon of stars thin enough to have gone unnoticed until now. Scientists are only beginning to read what those stars leave behind.
Paper Notes
Limitations
Several important caveats accompany this finding. Only one arm of the stream is visible in the data; the other arm is thought to wrap behind the brighter central portion of the host galaxy, making it undetectable. Because only integrated starlight, rather than individual stars, can be measured at this distance, color comparisons between the stream and its parent cluster carry more uncertainty than would equivalent measurements within the Milky Way. Background subtraction is complicated by the irregular shape of UGC 9050-Dw1 itself, and the choice of background region affects the color measurements. The statistical significance reported covers only the signal strength of the detected feature, not the full probability that the feature is a genuine stellar stream rather than a chance alignment of unresolved stars. Chance alignment or chance projection cannot be entirely ruled out. Stream modeling results depend on assumed escape conditions and the specific model parameters used, and the halo’s outer density structure remains unconstrained. The authors note that deeper observations would be needed to determine Oyashio’s origin with greater certainty.
Funding and Disclosures
Julie Kiel Holm and Sarah Pearson disclose support from Villum Fonden (grant number VIL53081). Sarah Pearson acknowledges support from the European Union (ERC, BeyondSTREAMS, 101115754). David J. Sand acknowledges support from NSF grants AST-2205863 and 2508746. Tjitske Starkenburg acknowledges support from NSF grant AST-2510183 and NASA grants 22-ROMAN22-0055 and 22-ROMAN22-0013. Computing resources were provided by the Tycho supercomputer hosted at the SCIENCE HPC Centre at the University of Copenhagen. The authors declare no competing interests.
Publication Details
Authors: Julie Kiel Holm, Sarah Pearson, Jacob Nibauer, David J. Sand, Adrian M. Price-Whelan, Tjitske Starkenburg, David Hendel, and Catherine Fielder
Affiliations include: DARK, Niels Bohr Institute, University of Copenhagen; DTU Space, Technical University of Denmark; Department of Astrophysical Sciences, Princeton University; Steward Observatory, University of Arizona; Center for Computational Astrophysics, Flatiron Institute; Center for Interdisciplinary Exploration and Research in Astrophysics (CIERA) and Department of Physics and Astronomy, Northwestern University; NSF-Simons AI Institute for the Sky (SkAI Institute)
Journal: Nature (2026)
Paper Title: “Evidence for the first globular cluster stellar stream beyond the Milky Way”
DOI: 10.1038/s41586-026-10878-w
Published: August 12, 2026







