The gaps astronomers used to trace dark matter are made by the Milky Way itself

Look closely enough at any of the long, thin rivers of stars that trail through our galaxy and you will find the stream has a gap — a missing section, a kink, a point where the stars bunch up then spread apart for no obvious reason. Those imperfections were the closest thing astronomers had to a photograph of dark matter: invisible clumps of the substance, theory predicted, would gravitationally punch through a stellar stream and leave exactly that kind of mark.

New simulations have done something uncomfortable to that logic. A team from the University of Washington ran 15,000 virtual stellar streams through four Milky Way-sized galaxies that contained no dark matter subhalos at all — and still found that nearly every stream developed the same kinks, gaps, and spurs previously held as dark matter’s calling card. Out of 15,000 streams, only 70 remained entirely featureless. The Milky Way itself, without any help from dark matter, was producing the signals.

“In our simulations, the host galaxies alone caused the same kinds of irregularities that we observe in real stellar streams,” said Arpit Arora, a postdoctoral scholar at the UW Department of Astronomy and the study’s lead author.

Recommended ReadingDark matter gave galaxies their spin 13 billion years ago — now confirmed at 7-sigma

How they ran the simulations

The team built four virtual galaxies matching the Milky Way in mass and structural distribution, then seeded each with stellar streams — long arcs of stars that trail behind globular clusters or dwarf galaxies as they orbit the galactic center. Streams were introduced across the full range of orbital shapes and distances observed in real Milky Way surveys.

The simulations ran forward five billion years without inserting any dark matter clumps. What the galaxies did have was their own uneven gravitational field: stars distributed unevenly in the disk, the gravitational pull of a central bar, denser concentrations in the halo. Those uneven tugs accumulate. A stream that started as a clean arc developed wiggles, then kinks, then gaps — and in some cases was torn apart entirely before the simulation ended.

Streams orbiting close to the galactic center showed the most distortion. Streams on wide, distant orbits fared somewhat better. But clean streams were the exception in every run: roughly one in 215 remained featureless.

The evidence that was supposed to be solid

Stellar streams should, in an idealized smooth galaxy, remain largely intact as they orbit. When gaps appeared in well-observed streams like Pal 5 or GD-1 — both thin trails of stars in the Milky Way’s outer halo — the natural explanation was that a dense dark matter subhalo had passed through and gravitationally punched a hole. The leading cosmological model, Lambda-CDM, predicts those subhalos exist in large numbers; they are invisible to telescopes but should leave physical marks on anything passing near them.

Stream gaps looked like those marks. They had the right sizes, the right frequency, the right spatial distribution. The case for using them as dark matter evidence had been building for two decades, across dozens of papers and two generations of sky surveys.

The UW simulations demonstrate that the Milky Way itself — with no dark matter subhalos at all — replicates the same gap pattern with high fidelity. The method was not wrong in principle. The background was far noisier than anyone had modeled.

What this does not settle

The simulations do not rule out dark matter subhalos. Lambda-CDM still predicts they are there; the question is whether stellar stream gaps can reliably detect them against the background the host galaxy is already producing.

The study used four model galaxies that match key observed properties of the Milky Way but cannot reproduce every detail of the actual gravitational environment. The full influence of the Large Magellanic Cloud — which has been pulling on the Milky Way with sufficient force to distort its halo — was not included. Encounters between streams and giant molecular clouds were also not modeled.

The researchers explicitly note that their simulations set a floor, not a ceiling. The host galaxy produces a certain level of stream distortion on its own; any dark matter signal must be larger than that floor to be detectable. Whether the gap in GD-1 or the spur in Pal 5 rises above that floor will require models that account for both effects simultaneously — which this study does not yet provide.

Common questions about stellar streams and dark matter

What is a stellar stream?

A stellar stream is a long, thin ribbon of stars stripped from a globular cluster or dwarf galaxy by tidal forces as the cluster orbits the galaxy. The Milky Way’s halo contains dozens of identified streams, each tracing an arc that follows the source cluster’s orbital path.

Why were stream gaps thought to prove dark matter?

Dark matter subhalos — clumps of dark matter smaller than a galaxy but still massive enough to affect gravity locally — were predicted by Lambda-CDM cosmology to pass through stellar streams and punch gaps into them. Observed gaps matched the predicted size and spacing, making them the most direct indirect probe of dark matter clumping in the Milky Way.

What did the UW simulations change?

The simulations showed that the Milky Way’s own uneven gravitational field produces the same kinds of gaps, kinks, and clumps in stellar streams, even with no dark matter subhalos present. Observed gaps can no longer be assumed to indicate dark matter without first accounting for the host galaxy’s contribution.

Can the two causes still be separated?

Yes, in principle. The UW work establishes a detailed baseline for what a galaxy alone does to its streams. Future analyses can subtract that baseline from real observations and look for residual signals that only dark matter would add. “Now that we can predict what the host galaxy does on its own, we can start isolating the part that dark matter is responsible for,” Arora said.

Which streams are the best targets going forward?

Streams orbiting at greater distances from the galactic center showed less distortion from the host galaxy in the simulations, leaving more room for a clean dark matter signal. Wide, distant streams are likely to be the most productive targets for future dark matter searches using this method.

The search continues, recalibrated

Arora and the UW team are extending the framework to specific named streams in the Milky Way — mapping which observed gaps can be explained by galactic structure alone and which remain anomalous. A companion paper applying the method to GD-1 and Pal 5 is in preparation, and wider application to Rubin Observatory’s upcoming stream catalog is the next step after that.

The Vera Rubin Observatory, coming online in late 2026 and early 2027, will identify hundreds of new stellar streams in the Milky Way and nearby galaxies. Those streams will need to be run through the new galactic-noise model before any gap can be claimed as evidence for dark matter — a more demanding standard, but one that surviving candidates will be far better positioned to meet.

Reference: Arora et al., “No Stream Left Unscathed: The Imprint of a Host Galaxy,” The Astrophysical Journal, 2026. DOI: 10.3847/1538-4357/ae89af

Tags: astronomy, Dark Matter, Milky Way, Arpit Arora, stellar streams, University of Washington

More Like This

Webb caught a galaxy several times the Milky Way’s mass — and it isn’t spinning

JWST found ancient galaxies hiding four times their measured mass in small stars

James Webb mapped 164,000 galaxies to chart the cosmic web back a billion years

The Milky Way’s black hole has been blowing a wind for 20,000 years

DESI wasn’t looking to break the standard model. Its galaxy maps did it anyway

The heaviest black holes are built by merging smaller ones, 153 mergers show