Euclid DR1 reveals 3D shapes for 40.9 million galaxies, confirming decline in prolate fraction among low-mass galaxies
One of the first major science papers from the European Space Agency’s Euclid mission is turning the telescope’s first public dataset into a population-scale result: using Euclid DR1 imaging, researchers inferred the 3D shape distribution of 40.9 million star-forming galaxies — the largest analysis of its kind so far — and found that an earlier hint about how low-mass galaxies change over time still holds up in the much bigger sample.
That matters because galaxies are seen on the sky as flat, 2D projections. Their true 3D structure has to be reconstructed statistically rather than read directly from a single image. Earlier work with the Hubble Space Telescope and the James Webb Space Telescope suggested that lower-mass galaxies were more often prolate — elongated, or cigar-like — in the distant past and less often so closer to today. But those studies relied on much smaller samples.
The new paper, “Euclid: Data Release 1 (DR1) — Bayesian 3D shape inference of star-forming galaxy populations,” was posted on arXiv as arXiv:2611.16109v3 and last revised Nov. 18. Led by M. Dunn, M. Huertas-Company and J. H. Knapen on behalf of the Euclid Consortium, it uses Euclid DR1 optical imaging and a Bayesian population-level inference framework to estimate intrinsic shapes from observed 2D galaxy images. As the abstract puts it, “This analysis is performed for 40 864 304 star-forming galaxies with stellar masses M★ = 10⁹–10¹⁰·⁵ M☉ at redshifts 0 < z < 1.2.”
Within that sample, the study includes more than 9 million galaxies at redshifts below 0.5, a nearby-universe regime the paper says had previously been limited by sample size in this kind of shape work. Using observed Sérsic axis ratios and effective radii, the team inferred population-level 3D ellipticity, triaxiality and size. It reports high mean ellipticities — about 0.600 to 0.821 across bins — which the authors say points to shape distributions dominated by flattened geometries.
The headline scientific result is that the larger Euclid sample reproduces the previously reported decline in the prolate fraction toward the present day for galaxies at the low-mass end of the study, specifically those with stellar masses of 10^9 to 10^9.5 times the mass of the sun. “We recover key evolutionary trends identified in previous smaller-sample studies,” the abstract says. The authors argue that because the same pattern appears in independent datasets with different depths and resolutions, it is unlikely to be just a small-sample artifact.
The scale jump is dramatic. According to the paper, per-bin sample sizes exceed those in early JWST shape studies by factors of 440 to 52,800 in overlapping mass-redshift ranges, and exceed legacy Hubble studies by factors of 440 to 17,600. That does not mean Euclid is delivering definitive 3D classifications for individual galaxies. It means the mission is large enough to make a statistical statement about galaxy populations with far stronger counting power than earlier surveys.
That counting power comes from Euclid DR1-Foundation, the mission’s first public data tranche, released Nov. 12. The release covers about 1,900 square degrees and, according to Euclid Consortium press materials, includes more than 900 million galaxies and more than 160 million spectra. “In just over one year, Euclid catalogued more than 900 million galaxies across the released area,” the consortium said in a release. This shape study is one of the first major DR1-Foundation science papers to show what that dataset can do beyond the release announcement itself.
The authors also stress the limits of the result. The work is a population-level inference based on single-band optical imaging and single-Sérsic morphological fits, not a direct 3D measurement of each galaxy. They say improved methods will be needed in later Euclid releases, as the mission moves beyond this first foundation dataset.