A dwarf galaxy in the NGC 1052 group is being reported as a third candidate in the region that appears to contain little or no dark matter, based on new spectroscopic analysis of data from the W. M. Keck Observatory’s Keck Cosmic Web Imager (KCWI) described in an arXiv preprint. The galaxy, referred to as NGC 1052-DF9 (DF9) in that work, sits along a linear trail of faint galaxies that researchers believe formed during a violent high-speed collision billions of years ago. If confirmed through peer review, the finding would strengthen a provocative theory: that rare cosmic smash-ups can strip dark matter from the ordinary matter that makes up stars and gas, producing galaxies that defy standard expectations.
What is verified so far
The case for DF9 as a dark-matter-deficient galaxy rests on absorption-line spectroscopy performed with the Keck Cosmic Web Imager, or KCWI. That instrument measured a stellar velocity dispersion of 6.4 km/s, with uncertainties of +4.0 and -4.3 km/s. Velocity dispersion is a proxy for how much mass a galaxy contains: the faster its stars move, the more gravitational pull, and therefore more total mass, the system must have. A reading as low as 6.4 km/s implies that DF9’s dynamical mass is far smaller than what astronomers would expect if a typical dark-matter halo were present.
DF9 follows a pattern set by two earlier discoveries in the same neighborhood. The first, NGC 1052-DF2, drew international attention when researchers used its unusual globular clusters as kinematic tracers to estimate its total mass. Those compact objects orbited so slowly that the galaxy appeared to have almost no dark matter at all. Shortly after, a second deficient galaxy, NGC 1052-DF4, was identified in the same group using similar techniques. The chronology of discoveries, DF2 then DF4 then DF9, has turned what once looked like a statistical fluke into a pattern demanding explanation.
That explanation took shape in a peer-reviewed paper proposing that DF2, DF4, and several fainter galaxies form a linear trail consistent with a bullet-dwarf collision. In this scenario, two gas-rich dwarf galaxies slammed into each other at high speed. Because dark matter interacts only through gravity, it passed through the collision largely undisturbed, while the gas and normal matter piled up and fragmented into new, dark-matter-poor galaxies strung along the collision axis. The concept borrows its logic from the famous Bullet Cluster, where colliding galaxy clusters showed a similar separation of dark and ordinary matter, but scales it down to dwarf galaxies.
Before the dark-matter question could even be asked, DF9 first had to be confirmed as a real member of the NGC 1052 group. Earlier work using KCWI velocity data corrected a prior misclassification and established that DF9 belongs to the group rather than being a foreground or background interloper. A separate kinematic study validated the trail itself as a physically connected structure by showing that the galaxies share a coherent position–radial-velocity relationship, not just a chance alignment on the sky. Together, those results underpin the idea that DF9 is part of the same dynamical story as DF2 and DF4.
The broader observational picture also supports the notion that something unusual happened in this corner of the universe. The galaxies in the trail are ultra-diffuse: they have the sizes of normal dwarfs or even larger systems, but with far fewer stars spread over a much wider area. Their stellar populations appear old and metal-poor, suggesting that most of their stars formed long ago and that they have not experienced much recent star formation. The alignment of these ghostly systems along a single axis, combined with their shared motion through space, is difficult to explain with standard models of group evolution that do not invoke a past collision.
What remains uncertain
The most important caveat is that the claim of DF9’s dark-matter deficiency currently exists only as a preprint submitted to The Astrophysical Journal Letters. It has not yet passed formal peer review. While the earlier DF9 identification and stellar-population analysis did receive peer-reviewed publication in a journal study, that paper addressed group membership and stellar properties, not the dark-matter question directly. The gap between those two stages of evidence matters: the velocity dispersion measurement of 6.4 km/s carries large error bars relative to the value itself, meaning the true dispersion could plausibly range from roughly 2 km/s to over 10 km/s. At the higher end of that range, the case for missing dark matter weakens considerably.
There is also a broader methodological question that has shadowed this line of research since the DF2 announcement. Measuring velocity dispersions in ultra-diffuse galaxies is technically demanding. These systems are faint, their stars are spread thin, and extracting reliable kinematics from integrated starlight requires long exposures and careful treatment of instrumental effects. Some independent teams have previously argued that distance uncertainties and small sample sizes in globular-cluster-based measurements could inflate the apparent dark-matter deficit. The shift from globular cluster tracers in DF2 to integrated-light absorption spectroscopy in DF9 represents a methodological evolution, but it introduces its own systematic challenges that reviewers will scrutinize.
The bullet-dwarf collision model, while elegant, also lacks direct observational confirmation of the collision event itself. The linear trail is suggestive, and the kinematic coherence strengthens the case, but no one has yet identified the dark-matter-rich remnants that should exist at the ends of the trail if the theory is correct. Without those counterparts, the model remains a best-fit hypothesis rather than a proven mechanism. Alternative explanations, such as tidal stripping by the massive elliptical galaxy NGC 1052 or unusual initial conditions in the group’s formation, have not been completely ruled out.
Another open question concerns how representative these galaxies are. If DF2, DF4, and DF9 truly lack dark matter, they could pose a challenge to certain modified-gravity theories that attempt to eliminate dark matter altogether, because those frameworks generally predict tighter links between visible matter and gravitational effects. On the other hand, if the apparent deficit turns out to be an artifact of measurement uncertainties or assumptions about distance, the episode would underline how easily selection effects can shape discoveries at the limits of observability. For now, the sample size is too small to draw sweeping conclusions about cosmology.
How to read the evidence
Readers evaluating this story should distinguish between three tiers of evidence. The strongest layer consists of the peer-reviewed papers: the Nature study establishing the trail geometry and collision framework, the MNRAS paper confirming DF9’s group membership, and the earlier ApJL papers on DF2 and DF4. These have survived independent scrutiny and form the foundation of the narrative that something unusual is happening in the NGC 1052 group.
The second tier is the new preprint on DF9’s velocity dispersion. Preprints hosted on arXiv allow astronomers to share results quickly with the community, but they have not yet been formally refereed. In practice, many influential astrophysics results appear first as preprints and are later revised in response to feedback before journal publication. However, until that process plays out, the numbers in the DF9 analysis should be treated as provisional, especially given the relatively low signal-to-noise ratio and the importance of subtle systematic effects.
The third tier consists of the broader theoretical interpretations and media narratives built on top of those technical results. Claims that DF9 “proves” dark matter can be separated from normal matter at galactic scales, or that it “rules out” alternative gravity theories, go beyond what the data currently support. A more cautious reading is that DF9, taken together with DF2 and DF4, strengthens the case that rare dynamical events in galaxy groups can produce systems with unusually low inferred dark-matter content, and that understanding those outliers will sharpen tests of both dark-matter and modified-gravity models.
For non-specialists trying to assess credibility, it helps to look at how the community responds over time. Independent teams may attempt to re-measure DF9’s velocity dispersion with different instruments, or to model the system using alternative assumptions about its distance and orientation. If those efforts converge on similar values, confidence in the dark-matter-deficient interpretation will grow. If they diverge significantly, the debate will shift to which methods and assumptions are most robust.
It is also worth remembering that the infrastructure behind these debates matters. Platforms like arXiv, which is supported by a network of institutional and individual backers, make it possible for researchers worldwide to access cutting-edge results without paywalls. That openness accelerates the cycle of critique, replication, and refinement that ultimately determines which claims endure.
For now, DF9 occupies an intriguing but provisional place in the cosmic inventory. The galaxy appears to join DF2 and DF4 as part of a trail of ultra-diffuse systems that may have formed in the aftermath of a high-speed collision, and preliminary measurements suggest it contains far less dark matter than standard models would predict. Yet the key measurement is still under review, the theoretical picture is incomplete, and alternative explanations remain on the table. As additional observations come in and the peer-review process runs its course, DF9 will either solidify its status as a genuine dark-matter-deficient galaxy or recede into the background as a cautionary tale about pushing data to their limits. Either outcome will teach astronomers something valuable about how galaxies, and the invisible matter that shapes them, come to be.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
