The galaxy that is in the neighbourhood and out of it
On the sky, the dwarf galaxy Shapiro DG-III sits 289 kiloparsecs from the Pinwheel Galaxy. Its distance is uncertain across 1.9 megaparsecs, six and a half times that gap, so whether it is a neighbour flips inside its own error bar.
In 2025 a high-school student in New York went looking in the dark around a famous galaxy. Julian Shapiro searched public archives from the Subaru Telescope and the Canada-France-Hawaii Telescope for faint dwarf galaxies in a ring just outside the halo of M101, the Pinwheel Galaxy, about 22 million light-years away. He found three. They are now catalogued as Shapiro DG-I, DG-II and DG-III, and the paper describing them was published in The Astrophysical Journal this month with him as its only author. He is now a first-year astrophysics student at UC Berkeley.
The three are interesting for what they are not doing. None of them is making stars. NASA’s GALEX ultraviolet survey, which lights up wherever young stars are forming, sees nothing at their positions. Shapiro added eight and a half hours of his own hydrogen-alpha exposures of DG-I from a telescope in the Canary Islands, and saw nothing there either. In astronomy that kind of nothing is a measurement: the images were deep enough that new stars would have shown up.
Quiet, and alone
Dwarf galaxies usually stop forming stars for a familiar reason. A big neighbour’s hot halo strips away their cold gas. So quenched dwarfs are normally found close to large galaxies, and dwarfs out on their own normally keep forming stars. Large surveys put quenched, isolated dwarfs at well under a tenth of a percent of the population. Shapiro’s galaxies look like exactly that rarity.
There is a known way to make one. A backsplash galaxy falls deep through a large galaxy’s halo, loses its gas in the plunge, and coasts back out past the halo’s edge, still on its orbit. Simulations of the standard cosmological model predict such galaxies should be common. Confirmed examples near us are not. Three candidates beside one Milky-Way-sized spiral would be a real test.
The number a photograph cannot hold
Here is the catch, and it is the part Shapiro spent most of his effort on. A telescope image records angles. DG-III sits 2.47 degrees from M101 on the sky. At M101’s distance that converts to 289 kiloparsecs, a little beyond the 250-kiloparsec working edge of its halo. That is the separation the picture shows, and it is known very well.
What the picture cannot show is depth. Shapiro estimated DG-III’s distance from surface-brightness fluctuations, which read how grainy a galaxy’s light looks: fewer, brighter grains mean closer. The answer is 8.0 megaparsecs, but anything from 7.1 to 9.0 is consistent with the data. That window is 1.9 megaparsecs wide, six and a half times the separation on the sky.
Put the two together and the verdict changes inside the error bar. At 7.1 megaparsecs DG-III sits about half a megaparsec from M101 in three dimensions, comfortably within the roughly 1.5 megaparsecs where simulations expect backsplash galaxies. At the central 8.0 it is 1.33 megaparsecs away, still inside. At 9.0 it is 2.3 megaparsecs away, and outside. The same galaxy, in the same photograph, is a neighbour or not depending on where in its own error bar it really sits.
How sharp the ruler must be
Run it backwards and you get a telescope proposal. For the whole error bar to land inside the backsplash zone, assuming the central value holds, DG-III can be no farther than 8.17 megaparsecs. That allows an upper error of 0.17 megaparsecs, about 2 percent, against today’s 12.5 percent. In the astronomer’s own units, the distance modulus has to be good to about 0.05 magnitudes rather than 0.3. That is roughly six times sharper, and it is why the paper closes by asking for Hubble or JWST images that resolve individual stars.
A weaker claim, a stronger paper
The paper’s own history makes the same point. The version Shapiro first posted, in December 2025, called the galaxies likely backsplash and gave DG-I a distance from about fifty resolved stars. The version accepted after peer review calls them candidates, replaces that estimate with surface-brightness fluctuations, and reports an error bar for DG-I more than twice as wide. Shapiro thanks the reviewer in the acknowledgments.
The discovery did not shrink. Three quiet dwarfs that nobody had catalogued are still there. What changed is how sure anyone can be about where they have been. Shapiro has said he hopes the work shows that “young astronomers are capable of uncovering these findings.” The images he used are public, and so are the tools.
Try it
Place the dwarf
Enter a dwarf’s angle from M101 and its distance range. The calculator turns the angle into kiloparsecs, finds the dwarf’s 3D distance from M101 at the near, central and far ends, and checks each against the 1.5 Mpc backsplash zone.
DG-I 2.81 · DG-II 2.77 · DG-III 2.47
DG-I 4.2 · DG-II 6.2 · DG-III 8.0
DG-I 3.0 · DG-II 3.9 · DG-III 7.1
DG-I 6.6 · DG-II 9.9 · DG-III 9.0
3D distance ≈ √(sky separation² + (distance − 6.7)²)
| Quantity | Value |
|---|---|
| Sky separation at M101’s 6.7 Mpc | 289 kpc |
| 3D distance from M101, near end | 0.49 Mpc · inside |
| 3D distance from M101, central | 1.33 Mpc · inside |
| 3D distance from M101, far end | 2.32 Mpc · outside |
| Distances that stay inside the zone | 5.23 to 8.17 Mpc |
| Precision needed to prove it, if the central value holds | ±2.1% (0.046 mag) |
The verdict flips inside the error bar. The photograph cannot settle it; only a sharper distance can.
Model: a right-triangle shortcut, with M101 at 6.7 Mpc and the 1.5 Mpc zone edge taken from simulations. The exact geometry differs by under 0.01 Mpc for these galaxies.
Classroom packet · 22 pages
Cosmic Distance & Dwarf Galaxies · Astronomy Case Study
The full classroom version of this analysis, built as a 90-minute block for grades 10 through 12 and early college, with a two-period split and a 50-minute short path.
- Students turn sky angles into kiloparsecs and run the distance-modulus equation forward and backward.
- Part 2B is the calculation above: the 1.9-megaparsec window, plotted to scale, and the verdict that flips inside it.
- Part 3 designs the measurement backwards, the way a real telescope proposal would.
- Part 4 compares both versions of the real paper and asks students to write a headline as strong as the evidence allows, and no stronger.
Left: the classroom packet on Teachers Pay Teachers. Right: a deck of cards covering all 110 Messier objects, including M101, the Pinwheel Galaxy whose neighbourhood this article explores — an Amazon affiliate link. As an Amazon Associate this site earns from qualifying purchases, at no additional cost to you.
Sources
J. Shapiro, “Discovery of Isolated, Quenched Candidate Backsplash Dwarf Galaxies Near M101,” The Astrophysical Journal (2026), doi:10.3847/1538-4357/ae99cb. Submitted and accepted versions: arXiv:2512.14808. A. Rony, “A UC Berkeley undergraduate’s research leads to an astronomical first,” Berkeley News, 6 October 2026. C. C. Petersen, “A Student Finds Three Galaxies and Possibly Solves a Mystery,” Universe Today, 9 October 2026.