Wide JWST Survey Finds Fewer Galaxies at z≈10, Bringing Counts Closer to Spectroscopy

A newly posted arXiv preprint reports that a wide James Webb Space Telescope survey spanning 36 separate sightlines found fewer galaxies around redshift 10 — a measure of cosmic distance and look-back time — than many earlier JWST photometric studies. The result brings those abundance estimates closer to spectroscopic measurements, which are generally considered more secure, and suggests some of the early tension with pre-JWST galaxy-formation models may be less severe than headline-grabbing counts first implied.

The study matters in part because of how it was designed. The preprint, posted April 20 and led by Kimi C. Kreilgaard, uses the second data release of BEACON, a JWST NIRCam pure-parallel imaging survey. It covers about 400 square arcminutes across 36 independent sightlines, a setup intended to reduce cosmic variance — the risk that one or two unusually rich patches of sky can skew conclusions about the early universe. The paper, titled “BEACON: JWST NIRCam Pure-parallel Imaging Survey. III. Constraints on the UV LF and the Clustering of z~7-14 Galaxies,” is listed as “to be submitted to ApJ” and has not yet been peer reviewed.

The authors identify 164 galaxy candidates at redshifts between 7 and 12, including 150 F090W-dropouts and 14 F115W-dropouts, with no robust F150W-dropouts. They also say 11 survey pointings overlap with public JWST spectroscopy and report 100% purity in those overlaps, meaning every candidate in that subset was supported by spectroscopic information. That is the authors’ assessment of the overlap sample, not a blanket validation of the full catalog.

In the key test of how common early galaxies were, the team finds that at about redshift 7.5, BEACON’s ultraviolet luminosity function — a count of galaxies by brightness — agrees with earlier work at the bright end but gives lower number densities for galaxies with ultraviolet magnitudes between minus 21 and minus 19 than some previous photometric studies. At about redshift 10, the new measurements are lower than most photometric JWST results but consistent with spectroscopic constraints. The paper says its luminosity functions at both redshift 7.5 and redshift 10 are consistent with pre-JWST models. At redshifts above 13, however, the survey’s limits do not rule out a possible excess of very early galaxies, but they do not establish one either.

That does not mean JWST’s earlier discoveries were mistaken. Webb has already pushed confirmed galaxy detections to roughly redshift 14. But many of the strongest claims about unexpectedly large numbers of bright early galaxies came from a small number of fields, leaving open questions about contamination and whether those regions were unusually crowded. BEACON’s broader footprint offers a different kind of check: a wider and cleaner sample that points to more moderate abundance estimates at some redshifts while still showing that very bright galaxies existed surprisingly early.

The same preprint also finds that the brightest galaxies were not spread evenly across the sky. The authors report significant clustering for galaxies brighter than ultraviolet magnitude minus 20.5 at redshifts between 7 and 10. Fields containing such bright galaxies were about three times more likely to be overdense than the survey as a whole, suggesting these objects tended to sit in especially crowded environments.

Comparing those observations with semi-numerical simulations, the authors estimate that galaxies brighter than minus 20.5 lived in halos — the large clumps of dark matter thought to host galaxies — that were about 0.9 dex less massive at redshift 11 than at redshift 7. In plain terms, the paper argues that similarly bright galaxies may have formed stars more efficiently at earlier times than expected. It also says the clustering appears stronger than predicted by pre-JWST relations between galaxy brightness and halo mass, implying either more massive halos than older models assumed, multiple galaxies sharing the same halo, or both. For now, though, the findings remain those of an arXiv preprint awaiting peer review.