arxiv: 2604.16666 · v1 · submitted 2026-04-17 · 🌌 astro-ph.GA

Recognition: unknown

A search for the first galaxies across >0.6 deg² of JWST imaging: new evidence for a rapid decline in star-formation activity at z>12

D. J. McLeod (1) , J. S. Dunlop (1) , R. J. McLure (1) , C. T. Donnan (2) , R. Begley (10) , S. Antonogiannaki (1) , D. Magee (3) , G. D. Illingworth (3)

show 36 more authors

P. Arrabal Haro (4 5) C. Bondestam (1) A. C. Carnall (1) F. Cullen (1) M. Dickinson (2) R. S. Ellis (6) B. L. Frye (7) H. Golawska (1) N. A. Grogin (8) I. J. B. Holst (1) P. S. Kamieneski (9) H. -H. Leung (1) F. -Y. Liu T. M. Stanton (1) E. R. Tittley (1) ((1) Institute for Astronomy University of Edinburgh (2) NSF's National Optical-Infrared Astronomy Research Laboratory (3) Department of Astronomy Astrophysics UCO/Lick Observatory (4) Center for Space Sciences Technology UMBC (5) Astrophysics Science Division NASA Goddard Space Flight Center (6) Department of Physics Astronomy University College London (7) Department of Astronomy/Steward Observatory University of Arizona (8) Space Telescope Science Institute (9) Department of Physics & Astronomy Chalmers University of Technology (10) Armagh Observatory Planetarium)

Authors on Pith no claims yet

Pith reviewed 2026-05-10 07:21 UTC · model grok-4.3

classification 🌌 astro-ph.GA

keywords high-redshift galaxiesUV luminosity functionstar-formation rate densityJWST NIRCamcosmic dawngalaxy formationredshift evolutionfirst galaxies

0 comments

The pith

JWST wide-area search finds star-formation rate density declining four times faster at z>12

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper reports results from a search for galaxies at redshifts 12.5 to 18.5 across more than 0.6 square degrees of JWST NIRCam imaging that includes over 150 independent sight-lines. The authors find evidence for a rapid drop in the ultraviolet luminosity function and the inferred star-formation rate density between redshift 11 and 13.5, with the decline becoming even steeper at higher redshifts. They measure that the drop in star-formation rate density at z>12 is about four times steeper than the trend at lower redshifts, placing the first galaxies around redshift 15. The results fit both models that invoke special adjustments to explain bright galaxies at z~12 and a simpler model driven by dark-matter halo growth combined with younger stellar ages.

Core claim

We present a new determination of the evolving galaxy UV luminosity function over 12.5<z<18.5 based on a wide-area search of >0.6 deg² of JWST NIRCam imaging. We find evidence for an accelerated decline in the UV LF and hence inferred star-formation rate density over the ≃100 Myr cosmic time interval between z=11 and z=13.5, with an even more rapid descent at earlier times. Our new measurement of ρ_SFR at z≃15.5 lies significantly below an extrapolation of the log-linear ρ_SFR(z) relation, favoring instead a piece-wise log-linear description in which the decline in ρ_SFR(z) at z>12 is ≃4 times steeper than at z<12. These results are consistent with a simple galaxy evolution model in which no

What carries the argument

The evolving galaxy UV luminosity function measured from photometrically selected candidates in wide-area JWST NIRCam imaging, used to derive the star-formation rate density ρ_SFR as a function of redshift.

Load-bearing premise

Photometric selection and redshift estimation correctly identify a complete sample of z>12 galaxies with negligible contamination from lower-redshift sources, and the absence of candidates at z>14.5 reflects a true lack rather than detection limits.

What would settle it

Spectroscopic confirmation of a substantial number of galaxies at z>14.5 or clear evidence that many of the z>12 candidates are actually lower-redshift contaminants.

Figures

Figures reproduced from arXiv: 2604.16666 by (10) Armagh Observatory, (2) NSF's National Optical-Infrared Astronomy Research Laboratory, (3) Department of Astronomy, (4) Center for Space Sciences, 5), (5) Astrophysics Science Division, (6) Department of Physics, (7) Department of Astronomy/Steward Observatory, (8) Space Telescope Science Institute, (9) Department of Physics & Astronomy, A. C. Carnall (1), Astronomy, Astrophysics, B. L. Frye (7), C. Bondestam (1), Chalmers University of Technology, C. T. Donnan (2), D. J. McLeod (1), D. Magee (3), E. R. Tittley (1) ((1) Institute for Astronomy, F. Cullen (1), F. -Y. Liu, G. D. Illingworth (3), H. Golawska (1), H. -H. Leung (1), I. J. B. Holst (1), J. S. Dunlop (1), M. Dickinson (2), N. A. Grogin (8), NASA Goddard Space Flight Center, P. Arrabal Haro (4, Planetarium), P. S. Kamieneski (9), R. Begley (10), R. J. McLure (1), R. S. Ellis (6), S. Antonogiannaki (1), Technology, T. M. Stanton (1), UCO/Lick Observatory, UMBC, University College London, University of Arizona, University of Edinburgh.

**Figure 1.** Figure 1: The cumulative raw survey area searched as a function of the global 8𝜎 limiting magnitude in F277W. The limiting magnitude is calculated in a 0.20′′−diameter aperture and corrected to total assuming a point-source correction. Note that > 70% of the area is as deep as 28.0 mag, providing a powerful combination of wide-area and depth with which to search for the first galaxies. advantage of providing mostly … view at source ↗

**Figure 2.** Figure 2: Depth maps corresponding to 5𝜎 limiting magnitudes (0.20′′−diameter apertures and point-source corrected to total) for a small subset of the fields studied. This subset spans ≃ 10% of the area and ≃ 7% of the independent sightlines, illustrating the dynamic range in sensitivity and area that this study has probed. The fields are scaled to reflect their relative sizes. 3.2 SED fitting Next, we ran the SED f… view at source ↗

**Figure 3.** Figure 3: Example SED plots, including an inset 𝜒 2 distribution, for a subset of our high-redshift candidates. The photometry presented is in 0.20′′–diameter apertures. For each source we show both the primary photometric redshift solution in blue and the (disfavoured) secondary low-redshift solution in red. For non-detections we indicate 1𝜎 limiting magnitudes with downward arrow symbols [PITH_FULL_IMAGE:figures/… view at source ↗

**Figure 4.** Figure 4: Example NIRCam postage stamps (from left to right: F090W, F115W, F150W, F200W, F277W, F356W, F410M, F444W where available) for the same subset of our high-redshift candidates as in [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗

**Figure 5.** Figure 5: Our new determination of the galaxy UV luminosity function at 12.5 < 𝑧 < 14.5. In the left-hand panel, we include, for comparison, the 𝑧 ≃ 11 UV LF determination from McLeod et al. (2024), as well as a number of other observational studies at similar redshifts. In the right-hand panel, we compare our results with the predictions of several recent theoretical models of galaxy evolution. The full references … view at source ↗

**Figure 6.** Figure 6: The UV LF at 𝑧 = 13.5 with three alternative functional forms over-plotted. With the present data, the Schechter and double power-law fits are indistinguishable. Based on the limits at 𝑀1500 = −21.7, we require another ≳ 1 dex of survey volume to make progress. 𝑧 = 13.5 LF determination prevents us from identifying the functional form with the present data. The upper limit at 𝑀1500 = −21.7 is unable to di… view at source ↗

**Figure 7.** Figure 7: Our determination of the galaxy UV luminosity function at 14.5 < 𝑧 < 18.5. In the left-hand panel, we include a comparison with both the 𝑧 ≃ 11 UV LF determination from McLeod et al. (2024), and the new 𝑧 = 13.5 results from the current study. We see a factor > 100× decrease in the number density of galaxies between 𝑧 ≃ 11 and our new 𝑧 ≃ 15.5 bin. In the right-hand panel, we include a compendium of litera… view at source ↗

**Figure 8.** Figure 8: The evolution of the cosmic star-formation rate density with redshift, as measured by this study in the context of other observations at 𝑧 > 9. We include the smoothly declining, log-linear log10 𝜌UV (z) relationship from Donnan et al. (2024) which closely matches existing observations up to 𝑧 ≃ 12, as well as the rapidly declining model from Harikane et al. (2022). Our 𝑧 = 13.5 results lie slightly below … view at source ↗

**Figure 9.** Figure 9: The evolving cosmic star-formation rate density as measured in this study and previously by Donnan et al. (2024) in the context of theoretical models (left-hand panel) and simulations (right-hand panel). Our measurements lie above some of the models (e.g., Yung et al. (2019); Mason et al. (2023) and simulations such as Millenium TNG (Kannan et al. 2023) and FiBY (e.g., Johnson et al. 2013; Paardekooper et … view at source ↗

**Figure 10.** Figure 10: This plot demonstrates that the evolving halo mass function model with progressively younger stellar ages from Donnan et al. (2025b) is wellmatched to the observations from McLeod et al. (2024) and Donnan et al. (2024) as well as with the new measurements from the current study. Similarly, a simple piecewise power-law, fitted to the data points, provides a good description of the trend in log10 𝜌UV (z) a… view at source ↗

read the original abstract

We present a new determination of the evolving galaxy UV luminosity function (LF) over the extreme redshift range $12.5<z<18.5$, based on a wide-area search of $>$0.6 deg$^2$ of JWST NIRCam imaging containing $>150$ independent sight-lines. We find evidence for an accelerated decline in the UV LF, and hence inferred star-formation rate density ($\rho_{\rm SFR}$), over the $\simeq100\rm{Myr}$ cosmic time interval between $z=11$ and $z=13.5$. Moreover, based on a notable lack of galaxy candidates at $z>14.5$, we find evidence for an even more rapid descent in star-formation activity towards earlier times, with our new measurement of $\rho_{\rm SFR}$ at $z\simeq15.5$ lying significantly below an extrapolation of the log-linear $\rho_{\rm SFR}(\rm z)$ relation inferred from early JWST LF studies. Instead, we find that the evolution in $\rho_{\rm SFR}(\rm z)$ at these very early times is better described by a piece-wise log-linear relation, in which the decline in $\rho_{\rm SFR} (\rm z)$ at $z>12$ is $\simeq4$ times steeper than at redshifts $z < 12$. Our observational results are consistent with a number of theoretical models of galaxy evolution which have incorporated a range of treatments in an attempt to explain the prevalence of UV-bright galaxies at least out to $z \simeq 12$ (e.g., increased star-formation efficiency, stochastic star-formation histories, an evolving stellar initial mass function and/or a shift towards attenuation-free stellar populations). However, our results are also entirely consistent with a relatively simple galaxy evolution model with no such adjustments, in which the rapid evolution of the dark-matter halo mass function at early times is for a while partially masked by progressively younger stellar ages, with the inferred epoch of first galaxy formation lying at $z\simeq15$.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Wide-area JWST data show a steeper drop in UV LF and SFR density above z=12 than log-linear fits, with first galaxies around z=15, but the high-z selection needs close checking.

read the letter

This paper reports a new UV luminosity function measurement from JWST NIRCam imaging over more than 0.6 square degrees and 150 independent sightlines, covering 12.5 < z < 18.5. The central result is that the inferred star-formation rate density declines about four times faster above z=12 than below, with the z~15.5 point falling below earlier extrapolations; they favor a piece-wise log-linear form and place the onset of galaxy formation near z=15. The large area is a clear plus compared with prior smaller-field JWST studies, cutting down cosmic variance and giving firmer counts at these redshifts. They also lay out comparisons to several classes of models, including ones with boosted efficiency or stochastic histories and a baseline model where halo mass function evolution is partly offset by younger stellar populations, showing both can fit the data. The methods appear standard for Lyman-break and SED-based selection, with the usual completeness and contamination steps implied. The main soft spot is exactly where the stress-test flagged: the absence of candidates above z=14.5 is what drives the steepest slope and the z~15 inference. At these redshifts the Lyman break sits between NIRCam filters whose depths and color terms differ across the fields, so any residual incompleteness or low-level contamination from interlopers would scale directly into the reported decline. The abstract gives no numbers on those rates, but the full paper presumably supplies them; if the simulations and tests hold up, the result strengthens. This is useful for anyone running early-universe simulations or reionization calculations who needs updated high-z LF constraints. It is worth sending to peer review because the dataset is substantial and the claim is timely, even if referees will press on the selection details and volume calculations.

Referee Report

3 major / 3 minor

Summary. The manuscript reports a wide-area search for galaxies at 12.5 < z < 18.5 in >0.6 deg² of JWST NIRCam imaging across >150 sight-lines. It derives the UV luminosity function and infers an accelerated decline in star-formation rate density (ρ_SFR) between z=11 and z=13.5, with an even steeper drop at z>14.5 based on the absence of candidates, implying first galaxy formation at z≃15. Results are compared to theoretical models, including both adjusted and simple galaxy-evolution scenarios.

Significance. If the photometric selection and completeness corrections are robust, the work would provide valuable constraints on early galaxy formation by showing that ρ_SFR(z) declines ~4 times more steeply at z>12 than at lower redshifts and lies below log-linear extrapolations, while remaining consistent with simple models in which young stellar ages partially mask the rapid evolution of the halo mass function.

major comments (3)

[§3 (Sample Selection)] §3 (Sample Selection): The headline claim of a factor-of-4 steeper decline in ρ_SFR at z>12 and the z≃15.5 measurement rests on the absence of z>14.5 candidates reflecting a true density drop. The manuscript must supply quantitative completeness simulations for mock z>14.5 galaxies down to the survey limit, explicitly accounting for varying NIRCam filter depths and color terms across the 150+ independent sight-lines, plus contamination rates from lower-z interlopers.
[§4 (UV Luminosity Function)] §4 (UV Luminosity Function): The upper limits on the LF at z>14.5 and the derived ρ_SFR(z=15.5) require an explicit statement of the effective comoving volume after all cuts, together with the full error budget (Poisson + cosmic variance). Without this, the statistical significance of the deviation from log-linear extrapolation cannot be assessed.
[§5 (ρ_SFR Evolution)] §5 (ρ_SFR Evolution): The statement that the decline at z>12 is ≃4 times steeper than at z<12 must be accompanied by the explicit piecewise log-linear fit, its uncertainties, and the precise z-bins used for the comparison; the current description leaves the numerical factor under-specified.

minor comments (3)

[Abstract] Abstract: Include a brief quantitative summary of the number of candidates per redshift bin and the survey 5σ depth to allow immediate assessment of the claimed decline.
[Figures] Figures 3–5: Ensure LF plots display both Poisson and cosmic-variance errors and clearly distinguish upper limits at z>14.5 from detections.
[Throughout] Notation: Define ρ_SFR consistently and avoid potential confusion with other density quantities used in the literature.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their careful and constructive review of our manuscript. We have addressed each major comment point by point below. In cases where additional detail or clarification was needed, we have revised the manuscript accordingly to improve transparency and robustness.

read point-by-point responses

Referee: §3 (Sample Selection): The headline claim of a factor-of-4 steeper decline in ρ_SFR at z>12 and the z≃15.5 measurement rests on the absence of z>14.5 candidates reflecting a true density drop. The manuscript must supply quantitative completeness simulations for mock z>14.5 galaxies down to the survey limit, explicitly accounting for varying NIRCam filter depths and color terms across the 150+ independent sight-lines, plus contamination rates from lower-z interlopers.

Authors: We agree that quantitative completeness and contamination assessments are necessary to support the interpretation of the non-detection at z>14.5. The original manuscript described the photometric selection and noted the absence of candidates, but we acknowledge the need for more explicit simulations. In the revised §3 we have added a dedicated subsection presenting completeness simulations in which mock z>14.5 galaxies with realistic SEDs and colors are inserted into the actual NIRCam images from each of the >150 sight-lines. These mocks account for the varying filter depths and transmission curves across fields. We report the recovery fraction as a function of magnitude and discuss the resulting completeness corrections. Contamination from lower-redshift interlopers is quantified by applying the same selection to both simulated low-z catalogs and to the data, with the results shown in a new figure. These additions confirm that our selection remains robust. revision: yes
Referee: §4 (UV Luminosity Function): The upper limits on the LF at z>14.5 and the derived ρ_SFR(z=15.5) require an explicit statement of the effective comoving volume after all cuts, together with the full error budget (Poisson + cosmic variance). Without this, the statistical significance of the deviation from log-linear extrapolation cannot be assessed.

Authors: We thank the referee for highlighting this omission. In the revised §4 we now explicitly state the effective comoving volume for the z>14.5 bin, computed from the surveyed area of >0.6 deg² multiplied by the comoving depth of the redshift interval and corrected for the average completeness after all selection cuts. The full error budget is detailed in the text and figure captions, combining Poisson statistics on the number counts (zero detections for the upper limits) with an estimate of cosmic variance obtained from the analytic formalism appropriate to our survey geometry and redshift range. These updates permit a quantitative assessment of the significance of the ρ_SFR(z=15.5) measurement relative to log-linear extrapolations. revision: yes
Referee: §5 (ρ_SFR Evolution): The statement that the decline at z>12 is ≃4 times steeper than at z<12 must be accompanied by the explicit piecewise log-linear fit, its uncertainties, and the precise z-bins used for the comparison; the current description leaves the numerical factor under-specified.

Authors: We agree that the piecewise log-linear description should be stated more explicitly. In the revised §5 we now provide the explicit functional form of the piecewise log-linear fit to log(ρ_SFR) versus z, including the best-fit slopes and their uncertainties for each segment. We define the redshift bins used for the comparison (low-redshift segment spanning z=8 to z=12 using literature values, high-redshift segment spanning z=12 to z=16 incorporating our new measurements at z=13.5 and the upper limit at z=15.5). The ratio of the two slopes is quantified together with its uncertainty, fully specifying the factor-of-4 claim and allowing readers to evaluate its robustness. revision: yes

Circularity Check

0 steps flagged

No circularity: results follow from direct counts and standard LF computation

full rationale

The paper derives the UV LF and ρ_SFR(z) directly from photometric candidate counts in >0.6 deg² of JWST NIRCam data across >150 sight-lines. The claimed accelerated decline at z>12 and the z≃15.5 measurement follow from the observed number densities (including the absence of z>14.5 candidates) inserted into standard Schechter-function fits and comoving-volume calculations; no equation or self-citation reduces these quantities to parameters defined by the same fit. External model comparisons are presented only as consistency checks, not as load-bearing inputs. The derivation chain is therefore self-contained against the imaging data.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Only the abstract is available, so the ledger is necessarily incomplete. The analysis implicitly relies on standard domain assumptions for high-redshift surveys rather than new free parameters or invented entities.

axioms (2)

domain assumption Photometric redshifts and color selections place galaxies at z>12 with negligible low-redshift contamination
Required to interpret the lack of candidates at z>14.5 as a real drop in number density
domain assumption Completeness corrections and luminosity-function fitting accurately convert observed counts into ρ_SFR
Necessary to claim a quantitative factor-of-four change in slope

pith-pipeline@v0.9.0 · 6019 in / 1528 out tokens · 59455 ms · 2026-05-10T07:21:24.005058+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

2 extracted references · 1 canonical work pages

[1]

IVOA Working Draft 15 October 2012 , year = 2012, month = oct, pages =

Acebron A., et al., 2018, ApJ, 858, 42 Acebron A., et al., 2019, ApJ, 874, 132 Acebron A., et al., 2020, ApJ, 898, 6 Adams N. J., et al., 2024, ApJ, 965, 169 Adams N. J., et al., 2025, arXiv e-prints, p. arXiv:2502.10282 ArnoutsS.,CristianiS.,MoscardiniL.,MatarreseS.,LucchinF.,FontanaA., Giallongo E., 1999, MNRAS, 310, 540 Arrabal Haro P., et al., 2023, N...

work page doi:10.5479/ads/bib/2012ivoa.rept.1015r 2018
[2]

7 BEA-20:58−42:47 - - 27.3 27.5 27.7 28.0 28.1 - 27.6 BEA-23:03−62:50 - 27.3 27.3 28.5 28.6 29.0 29.0 27.9 28.2 BEA-23:16−59:09 - - 27.4 27.7 27.8 28.3 28.3 - 27.9 BEA-23:25−12:03 - 27.9 28.2 28.5 28.6 28.8 28.9 28.2 28.5 BEA-23:25−12:15 - - 26.7 27.0 27.2 27.7 27.8 - 27.5 POP-07:40+31:04 - - 26.9 27.2 27.4 27.4 27.5 - 27.1 POP-09:12+52:54 - - 28.1 28.3 2...

2026