Pith. sign in

REVIEW 4 major objections 4 minor 33 references

The paper argues that the blue and red ultraviolet colors of the first galaxies are set by how completely clustered supernovae vent their own dust out of natal clouds and out of the galaxy's gas disk.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review

2026-08-01 12:37 UTC pith:UZZDOYGD

load-bearing objection Mechanically plausible two-stage venting model, but the central Fig. 1 validation is circular and needs an independent ε⋆ before it tests the mechanism. the 4 major comments →

arxiv 2607.19471 v1 pith:UZZDOYGD submitted 2026-07-21 astro-ph.GA astro-ph.CO

The first dusty galaxies across 6 lesssim z lesssim 14: Blue monsters, red monsters, and the bimodality of dust content in early galaxies

Sergio Mart\'inez-Gonz\'alez , Casiana Mu\~noz-Tu\~n\'on , Santiago Jim\'enez This is my paper
classification astro-ph.GA astro-ph.CO
keywords high-redshift galaxiesUV-continuum slopedust extinctionsupernova feedbacksuperbubble breakoutgalaxy formationblue monstersred monsters
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

JWST and ALMA have found galaxies at redshifts 6–14 that split into two camps: extremely blue systems with almost no dust (β_UV ≲ −2.4) and red, dusty, massive systems (β_UV ≳ −1). This paper tries to show that the split is neither a difference in dust production nor an evolutionary clearing sequence, but a difference in how much freshly made supernova dust stays inside the gas layer that screens ultraviolet light. Dust is made and shock-processed in clustered supernovae; some is blown out of the parent cloud, some is carried out of the thin high-redshift disk, and only the dust that fails both escapes contributes to reddening. Combining this two-stage mechanical venting with absorption opacities from clustered-supernova simulations, the model reproduces blue monsters at low stellar mass and high star-formation efficiency and red monsters at higher mass and stronger radiative losses. If the paper is right, the dust content of the earliest galaxies is controlled by supernova momentum and disk geometry rather than by radiation pressure alone.

Core claim

Blue monsters and red monsters at z≳10 are, in this model, the same dust-formation engine seen through different escape efficiencies. Clustered supernovae condense dust and process it through multiple shocks, leaving a shock-processed dust-to-stellar ratio log ξ_d,0 ≈ −2.67. A fraction f_mech of that dust is removed by mechanical blowout from the natal cloud; of what remains, a further fraction f_break is carried out of the galaxy's gas disk by individual or coalesced superbubbles. The retained ratio is ξ_d = ξ_d,0 (1 − f_mech f_break), and this retained column, projected over the UV-emitting area, sets the UV slope through a shock-processed grain opacity with κ_abs(1600 Å) = 1.45×10^4 cm² g

What carries the argument

The load-bearing mechanism is the two-stage dust-removal chain: cloud-scale mechanical blowout followed by disk breakout. The driver is the mechanical luminosity of a stellar cluster over its supernova phase, L_SC = 6.3×10^35 η_SN M_SC (E_SN/10^51 erg)(t_SN/5×10^7 yr)^−1 erg s⁻¹, with η_SN ≈ 1/95.5 M_sun⁻¹ for a Kroupa IMF and E_SN reduced from 10^51 to 5×10^49 or 10^49 erg by radiative losses. Breakout from the disk occurs when the superbubble reaches the vertical scale height H_gas before stalling, with thresholds L_SC/(πH_gas²) ≥ 10^−4 erg cm⁻² s⁻¹ and M⋆,SN/(πR_e² t_SN) ≥ 0.1 M_sun yr⁻¹ kpc⁻². Coalescence of overlapping bubbles (interaction distance d_int = 2 min(R_sb, H_gas)) boosts the

Load-bearing premise

The load-bearing premise is that two relations measured at lower redshift—star formation rate rising steeply as (1+z)^2.30 and galaxy size shrinking as (1+z)^−0.75—continue unchanged out to z=14, because these extrapolations set the disk scale heights, gas densities, and supernova surface densities that decide whether dust stays or escapes.

What would settle it

Find a spectroscopically confirmed z≈12 galaxy with stellar mass near 10^9.4 M_sun, compact half-light radius ≈0.4 kpc, and gas fraction ≈0.8 that nevertheless shows β_UV ≈ −2.5; the model's mass-retention branch would fail, since this combination of inputs should retain enough dust to be red. Equally, a 10^8.5 M_sun galaxy with β_UV ≈ −1 would indicate dust retention is not set by the mass/geometry scaling. A statistical version would map β_UV versus stellar mass for z>10 galaxies and compare the slope to the predicted transition near 10^8.5–10^9 M_sun.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X LinkedIn Reddit HN

If this is right

  • At z≳10, blue galaxies need not be intrinsically dust-free; their low dust-to-stellar ratios (log ξ_d ≲ −4) follow from venting the dust out of the UV-attenuating layer, so dust production and retention decouple.
  • A confirmed red monster at z=11.45 is reproduced without a separate clearing stage; it is a higher-mass system whose disks retain more dust, and the model predicts many more such systems at 10^9–10^9.5 M_sun.
  • Toward z≈6–7, the same framework naturally produces dusty, ALMA-detected galaxies: larger gas scale heights and lower mechanical-power surface densities suppress breakout, and grain growth in compressed supershells can add to the retained dust.
  • The model predicts that the blue/red boundary in β_UV should move with stellar mass, size, gas fraction, and the fraction of star formation occurring in bound clusters, so combined JWST and ALMA samples of z>10 galaxies can test the mass trend.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The model implies that the dust removed from blue monsters is not destroyed but transferred to the circumgalactic medium; detecting diffuse dust or metal absorption around blue galaxies at z≈10 would be a direct, paper-implicit test.
  • Because the escape fraction depends on the clumpiness of the gas, the same galaxy seen along different sightlines could scatter between blue and red values; current unresolved slopes may hide this variance, and resolved imaging of nearby analogs could calibrate it.
  • A sharper test would isolate f_coll, the fraction of clusters whose superbubbles merge before breakout; if JWST morphological data can constrain how clustered star formation is, the model's retention curve becomes a quantitative prediction rather than an input.
  • The redshift extrapolations used for size and main sequence are probably too smooth; if the main sequence flattens at z>9, the blue-monster region would shrink, which future spectroscopic samples of z=12–14 galaxies will soon decide.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. This Letter proposes that the blue and red monster galaxies at 6≲z≲14 are different outcomes of supernova dust retention in clustered star-forming systems. The model follows dust through two mechanical stages — cloud-scale blowout and disk breakout of superbubbles — computing a retained dust-to-stellar ratio via Eqs. (1)–(7) and converting it into a UV-continuum slope using post-shock dust opacities from the authors’ 3D clustered-supernova simulations. Redshift-dependent sSFR and size scalings set the disk structure, and the model is compared to observed blue monsters and EGS-z11-R0 in Fig. 1. The central claim is that mechanical venting, not radiation-pressure clearing alone, sets the dust content and UV slope of early galaxies.

Significance. If established, the proposed mechanical-venting mechanism would be a genuine alternative to the attenuation-free radiation-pressure paradigm and would connect a wide body of simulation work on clustered supernovae and dust processing to JWST/ALMA observations. The UV opacity is a concrete prediction from the evolved post-shock grain-size distribution, and the two-stage breakout formalism is explicit and falsifiable in principle. However, the quantitative empirical support currently rests on a circular comparison in Fig. 1, and several key fractions are not derived in the manuscript. With independent ε⋆ constraints or forward predictions, this could become a significant contribution; as written, the evidence does not decisively favor mechanical venting over other parameterized models.

major comments (4)
  1. [§3 / Fig. 1] The comparison to observations is circular as presented. The caption states that markers are ε⋆ inferred from observed (z, βUV) values using the same model that predicts βUV(z, ε⋆). Because βUV is a single-valued function of ε⋆ at fixed z and M⋆, any observed βUV can be assigned an ε⋆ that places it exactly on the model surface; the visual coincidence therefore carries no independent evidential weight. The paper does not list the inferred ε⋆ values or compare them to independent measurements or limits. Please provide an independent calibration of ε⋆ (e.g., from SED fitting, local cluster-scale constraints, or theoretical distributions), or show a forward prediction with an assumed ε⋆ distribution and compare predicted βUV without inversion.
  2. [§2.2, Eq. (6), and §3] The disk-breakout fractions f_ind_break and f_coll_break are introduced in Eq. (6) and quoted in §3 (0.32–0.57 for individual superbubbles, 0.72–1 for coalesced complexes), but their functional forms and the definition of f_coll are not given. These quantities control the retained dust and hence βUV. Similarly, f_mech is taken from Paper I without enough information for a reader to reproduce the population averages over cluster masses and cloud-core radii. Please provide explicit formulas (or a reproducible code/software release) for f_break, f_coll, and the averaging procedure.
  3. [§2.3, Eqs. (8)–(9)] The predicted boundary between blue and red monsters depends on M_star,SN, Σ_gas, and H_gas, which are set by extrapolating the sSFR relation (calibrated over 3≤z≤9) to z=14 and the size relation Re(z)∝(1+z)^{-0.75} (calibrated at z<3) to z=14. The manuscript acknowledges these extrapolations but does not test how sensitive the bimodality is to the assumed slopes. A flatter sSFR evolution or a different size trend would shift the breakout thresholds in Eq. (3) and could move EGS-z11-R0 and the blue-monster region. Please add a sensitivity test, e.g., varying the (1+z) exponents within their quoted uncertainties, or state which observational constraints break the degeneracy.
  4. [§2.2 and §3, panel c] The red-monster branch is obtained by reducing ESN from 5×10^49 erg to 1×10^49 erg, i.e., from 95% to 99% radiative loss, without a derivation tying the loss fraction to stellar mass, cluster properties, or environment. As a result, the agreement with EGS-z11-R0 in panel c is partly an input rather than a clean prediction of the two-stage venting picture. Please provide a physical scaling for ESN or a sensitivity study showing that the red-monster regime appears for a plausible range of radiative-loss fractions.
minor comments (4)
  1. [§2.2, Eq. (5)] The denominator '2π r_s^2' in the central overlap parameter η0 is unclear. If it is meant to be the disk area over which clusters are distributed, please define it explicitly; otherwise the expression is difficult to interpret.
  2. [§2.4, Eq. (11)] Since κ_abs(λ) is given as a power law with bUV=-0.48, it would be helpful to state the resulting algebraic relation between βUV, β_int, bUV, and κ0 Σ_d,UV. This would make the mapping from retained dust to βUV more transparent.
  3. [Abstract and §3] The term 'red monster' is used with slightly different thresholds: the abstract uses β_UV≳−1, while §3 defines a 'red-monster regime' reaching β_UV≃−1.5 to −0.5. Please harmonize the terminology.
  4. [Table 1 caption] The scaling note for M⋆,gal=10^9.4 M⊙ says to scale H_gas by 0.31 while Σ_gas scales by 3.2; a sentence explaining the assumed velocity dispersion (fixed or varying) would help readers understand the implied vertical support.

Circularity Check

2 steps flagged

Fig. 1 validation is an inversion, not a prediction: ε⋆ is inferred from the observed βUV using the same model that returns βUV, so blue/red-monster agreement is built in; self-citation to Paper I provides the only quantitative prior check.

specific steps
  1. fitted input called prediction [Section 3 (Results), Fig. 1 caption]
    "Markers show ε⋆ inferred from observed (z, βUV) values (Ziparo et al. 2023; Bunker et al. 2023; Finkelstein et al. 2022; Carniani et al. 2024; Rodighiero et al. 2026); colored rings show their βUV uncertainties."

    The model outputs βUV as a function of ε⋆ through Eqs. (6), (7), and (11): ξd depends on fmech(ε⋆) and fbreak, and the attenuated continuum is Lλ,obs ∝ exp[-κabs(λ)Σd,UV]. Placing every observed galaxy at the ε⋆ that solves βUV_model(z, ε⋆) = βUV_obs ensures the markers lie on the model surface. Any observed βUV can be mapped to some ε⋆, so the close match of blue and red monsters to the colored contours is not an independent confirmation of mechanical venting; it is the inverse of Eq. (11). No externally constrained cloud-scale star-formation efficiency is given, so the figure cannot discriminate mechanical venting from radiation-pressure clearing.

  2. self citation load bearing [Section 2.1, text after Eq. (1)]
    "The cloud-scale blowout model yielded dust-to-stellar mass ratios consistent with the log ξd ∼ −4 values inferred for spectroscopically confirmed blue monsters (Fig. 2 of Paper I)."

    The only quantitative consistency check for the cloud-scale stage is a reference to the authors' own Paper I (Martínez-González et al. 2026, same three authors). This paper also takes fmech from the Paper I scalings and ξd,0 and κabs from the same group's 2022 simulations. The central claim is therefore supported by a self-referential chain rather than by an external, independent benchmark; the current Fig. 1 then reuses that same model to infer ε⋆ from βUV. This compounds the circularity: the mechanism is not tested against a prediction from outside the model.

full rationale

The paper builds a two-stage mechanical-venting model with many physically motivated ingredients (Roy et al. 2013 disk-breakout thresholds, Simmonds et al. 2025 sSFR, van der Wel et al. 2014 sizes, Draine optical constants). These inputs are not themselves part of the target result, so the derivation is not circular in its formal equations. The serious problem is the empirical validation. Fig. 1 compares observed galaxies with model contours after converting each observed βUV into a cloud-scale star-formation efficiency ε⋆ using the same model (Eqs. 6-11). Because every observed βUV can be mapped to some ε⋆, the figure demonstrates flexibility, not falsifiability: it would also be possible to force a radiation-pressure model through the same points. The blue-monster consistency is further attributed to Fig. 2 of Paper I, a same-author result, and the red-monster branch is helped by selecting ESN=1e49 erg in panel c; no independent estimate of ε⋆ is reported. The central claim that mechanical venting, rather than radiation-pressure clearing alone, sets the dust content of z≈6-14 galaxies therefore rests on a model whose key comparison is an inversion of its own output. This is a partial circularity: the mechanism is still a legitimate hypothesis with independent sub-components, but the paper's headline observational support is built in rather than predicted. Score 6 reflects one-by-construction validation plus a load-bearing self-citation chain.

Axiom & Free-Parameter Ledger

5 free parameters · 8 axioms · 0 invented entities

The central claim rests on prior self-cited models for dust production, blowout, and opacity, plus extrapolated empirical scaling relations. No new entities are introduced. The parameters ESN and ε⋆ carry much of the blue/red separation, and the extrapolations of sSFR and size relations are load-bearing.

free parameters (5)
  • Effective supernova energy ESN for large-scale driving = 5×10^49 erg (fiducial, 95% loss); 1×10^49 erg (panel c, 99% loss)
    Chosen from radiative-loss assumptions; the 99%-loss variant is needed to produce the red-monster regime (Sect. 2.2, Fig. 1c).
  • Cloud-scale star formation efficiency ε⋆ = 0.01–1.0, with observed markers inferred from βUV
    A free model axis; observed markers are placed using βUV-inferred ε⋆ (Fig. 1 caption), making part of the apparent agreement circular.
  • Pre-breakout gas fraction fgas = 0.80
    Chosen to represent the larger gas reservoir before venting (Sect. 2.3); sets Σgas and Hgas.
  • Gas velocity dispersion σgas = 75 km/s
    Adopted for vertical support from the observed 53–100 km/s range (Carniani+24); affects Hgas and breakout (Sect. 2.3).
  • Initial shock-processed dust-to-stellar ratio log ξd,0 = -2.67
    Taken from the authors' own 3D clustered-SN simulations (Martínez-González+22); normalizes all retention equations (Eqs. 1, 6).
axioms (8)
  • domain assumption Disk breakout occurs when L_SC/(πH_gas^2)≥1e-4 erg cm^-2 s^-1 and M_SN/(πR_e^2 t_SN)≥0.1 Msun yr^-1 kpc^-2 (Roy et al. 2013).
    Adopted as the breakout criterion in Sect. 2.2, Eq. (3); central to computing f_break.
  • domain assumption Cloud-scale mechanical blowout fraction f_mech from Paper I is valid.
    Eq. (1) uses f_mech from Martínez-González+2026 without re-derivation; the present paper's results inherit that model.
  • domain assumption UV absorption opacities from the authors' 3D clustered-SN simulations are applicable.
    κ_abs(1600Å)=1.45e4 cm^2 g^-1 and b_UV=−0.48 come from Martínez-González+2022 (Sect. 2.4), a self-cited prior result.
  • domain assumption The high-redshift star-forming main sequence (Eq. 8) can be extrapolated from z=9 to z=14.
    sSFR∝(1+z)^2.30 is calibrated over 3≤z≤9 (Simmonds+25) and extrapolated to z=14 in Sect. 2.3.
  • domain assumption The late-type size relation Re(z)∝(1+z)^−0.75 (Eq. 9) can be extrapolated from z<3 to z=14.
    Used to set R_e and hence Σgas and Hgas; calibrated at z<3 and extrapolated in Sect. 2.3.
  • domain assumption Intrinsic dust-free UV slope β_int = −2.62.
    Adopted from Cullen+24 as the blue limit (Sect. 2.4).
  • domain assumption High-redshift compact galaxies are cluster-dominated, with cluster fraction Γ≈1.
    Invoked in the Introduction from Vanzella+26 and Menon & Smith 26; the whole cluster-based framework depends on it.
  • domain assumption Equal graphite-silicate dust mixture with Draine 1985 optical constants.
    Sets the opacity normalization and spectral slope in Sect. 2.4; not independently constrained here.

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of The first dusty galaxies across $6 \lesssim z \lesssim 14$: Blue monsters, red monsters, and the bimodality of dust content in early galaxies." pith.science (2026). https://pith.science/paper/UZZDOYGD

@misc{pith2026260719471,
  author       = {Pith},
  title        = {Pith review of: The first dusty galaxies across $6 \lesssim z \lesssim 14$: Blue monsters, red monsters, and the bimodality of dust content in early galaxies},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UZZDOYGD}},
  note         = {Machine review of arXiv:2607.19471}
}
Share X LinkedIn Reddit HN
read the original abstract

JWST has revealed galaxies at $z>10$ with extremely blue UV-continuum slopes approaching $\beta_{\rm UV}\lesssim-2.4$, and low inferred dust contents. At similar redshifts, JWST and ALMA reveal dusty massive systems reaching $\beta_{\rm UV}\gtrsim-1$. We aim to explain blue monsters and red monsters at $z\gtrsim10$ as different outcomes of supernova dust retention in clustered star-forming systems and connect this bimodality to dusty massive galaxies at $z\simeq6$--7. We model dust removal in a cluster-dominated regime through cloud-scale mechanical blowout followed by breakout from stratified high-redshift disks. The shock-processed dust mass retained in the UV-attenuating layer is converted into a UV-continuum slope using absorption opacities derived from 3D hydrodynamical simulations of clustered supernovae in porous molecular clouds. In compact, gas-rich galaxies, efficient cloud blowout and disk breakout reduce the retained dust fraction to blue-monster levels, yielding $\beta_{\rm UV}\lesssim-2.4$. Larger stellar masses increase the retained dust column and produce red-monster-like systems with $\beta_{\rm UV}\gtrsim-1.5$, reaching $\beta_{\rm UV}\simeq-0.5$ when radiative losses weaken large-scale driving. The transition depends on whether clustered-supernova ejecta cross both the natal cloud and the galactic gas layer. Supernova dust production, shock processing, radiative losses, and mechanical venting can jointly explain UV-bright dust-poor galaxies and red dusty massive systems across $6\lesssim z\lesssim14$.

Figures

Figures reproduced from arXiv: 2607.19471 by Casiana Mu\~noz-Tu\~n\'on, Santiago Jim\'enez, Sergio Mart\'inez-Gonz\'alez.

Figure 1
Figure 1. Figure 1: UV-continuum slope in (z, ε⋆) for fgas = 0.80, σgas = 75 km s−1 , tSN = 30 Myr, and Re(zref = 2.25) = 1.165 kpc. Here ε⋆, σgas, tSN, and Re denote the cloud-scale star-formation efficiency, gas velocity dispersion, supernova-phase duration, and effective radius. Colors and dashed contours give βUV. Panel a uses M⋆,gal = 108.5 M⊙; panels b and c use 109.4 M⊙, with 99 per cent radiative loss in panel c. Blue… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

33 extracted references · 3 linked inside Pith

  1. [1]

    L., et al

    Arrabal Haro, P., Dickinson, M., Finkelstein, S. L., et al. 2023, ApJ, 951, L22

    2023

  2. [2]

    Bakx, T. J. L. C., Sommovigo, L., Tamura, Y ., et al. 2025, MNRAS, 544, 1502

    2025

  3. [3]

    J., Saxena, A., Cameron, A

    Bunker, A. J., Saxena, A., Cameron, A. J., et al. 2023, A&A, 677, A88

    2023

  4. [4]

    K., et al

    Burgarella, D., Buat, V ., Inoue, A. K., et al. 2026, arXiv e-prints, arXiv:2605.09829

    Pith/arXiv arXiv 2026

  5. [5]

    L., & Storchi-Bergmann, T

    Calzetti, D., Kinney, A. L., & Storchi-Bergmann, T. 1994, ApJ, 429, 582

    1994

  6. [6]

    2024, Nature, 633, 318

    Carniani, S., Hainline, K., D’Eugenio, F., et al. 2024, Nature, 633, 318

    2024

  7. [7]

    2013, MNRAS, 432, 2061

    Chevallard, J., Charlot, S., Wandelt, B., & Wild, V . 2013, MNRAS, 432, 2061

    2013

  8. [8]

    J., McLure, R

    Cullen, F., McLeod, D. J., McLure, R. J., et al. 2024, MNRAS, 531, 997

    2024

  9. [9]

    Draine, B. T. 1985, ApJS, 57, 587

    1985

  10. [10]

    2025, A&A, 694, A286

    Ferrara, A., Pallottini, A., & Sommovigo, L. 2025, A&A, 694, A286

    2025

  11. [11]

    2026, arXiv e-prints, arXiv:2605.22914

    Ferrara, A., Rodighiero, G., Carniani, S., et al. 2026, arXiv e-prints, arXiv:2605.22914

    Pith/arXiv arXiv 2026

  12. [12]

    2022, MNRAS, 512, 58

    Ferrara, A., Sommovigo, L., Dayal, P., et al. 2022, MNRAS, 512, 58

    2022

  13. [13]

    L., Bagley, M

    Finkelstein, S. L., Bagley, M. B., Arrabal Haro, P., et al. 2022, ApJ, 940, L55

    2022

  14. [14]

    Inami, H., Algera, H. S. B., Schouws, S., et al. 2022, MNRAS, 515, 3126 Jiménez, S., Silich, S., Mayya, Y . D., & Zaragoza-Cardiel, J. 2024, ApJ, 960, 81 Jiménez, S., Tenorio-Tagle, G., & Silich, S. 2019, MNRAS, 488, 978 Jiménez, S., Tenorio-Tagle, G., & Silich, S. 2021, MNRAS, 505, 4669

    2022

  15. [15]

    2001, MNRAS, 322, 231 Martínez-González, S., Jiménez, S., & Muñoz-Tuñón, C

    Kroupa, P. 2001, MNRAS, 322, 231 Martínez-González, S., Jiménez, S., & Muñoz-Tuñón, C. 2026, A&A, 705, L10 Martínez-González, S., Silich, S., & Tenorio-Tagle, G. 2021, MNRAS, 507, 1175 Martínez-González, S., Wünsch, R., Palouš, J., et al. 2018, ApJ, 866, 40 Martínez-González, S., Wünsch, R., Silich, S., et al. 2019, ApJ, 887, 198 Martínez-González, S., Wü...

    2001

  16. [16]

    Menon, S. H. & Smith, A. 2026, arXiv e-prints, arXiv:2605.13982

    Pith/arXiv arXiv 2026

  17. [17]

    2024, ApJ, 963, 9

    Morishita, T., Stiavelli, M., Chary, R.-R., et al. 2024, ApJ, 963, 9

    2024

  18. [18]

    2025, arXiv e-prints, arXiv:2509.18266

    Narayanan, D., Torrey, P., Stark, D., et al. 2025, arXiv e-prints, arXiv:2509.18266

    arXiv 2025

  19. [19]

    J., Adams, N

    Ormerod, K., Conselice, C. J., Adams, N. J., et al. 2024, MNRAS, 527, 6110 Planck Collaboration, Aghanim, N., Akrami, Y ., et al. 2020, A&A, 641, A6

    2024

  20. [20]

    2026, arXiv e-prints, arXiv:2603.15841

    Rodighiero, G., Ferrara, A., Catone, M., et al. 2026, arXiv e-prints, arXiv:2603.15841

    arXiv 2026

  21. [21]

    B., Sharma, P., & Shchekinov, Y

    Roy, A., Nath, B. B., Sharma, P., & Shchekinov, Y . 2013, MNRAS, 434, 3572

    2013

  22. [22]

    J., Ormerod, K., et al

    Schouws, S., Bouwens, R. J., Ormerod, K., et al. 2025, ApJ, 988, 19

    2025

  23. [23]

    2024, MNRAS, 534, 1433

    Shen, X., V ogelsberger, M., Borrow, J., et al. 2024, MNRAS, 534, 1433

    2024

  24. [24]

    Silich, S., Tenorio-Tagle, G., Muñoz-Tuñón, C., & Cairos, L. M. 2002, AJ, 123, 2438

    2002

  25. [25]

    2025, MNRAS, 544, 4551

    Simmonds, C., Tacchella, S., McClymont, W., et al. 2025, MNRAS, 544, 4551

    2025

  26. [26]

    K., Somerville, R

    Sommovigo, L., Cochrane, R. K., Somerville, R. S., et al. 2025, ApJ, 990, 114

    2025

  27. [27]

    2022, MNRAS, 513, 3122

    Sommovigo, L., Ferrara, A., Pallottini, A., et al. 2022, MNRAS, 513, 3122

    2022

  28. [28]

    H., et al

    Sommovigo, L., Lancaster, L., Menon, S. H., et al. 2026, ApJ, 1003, 170 Szécsi, D., Agrawal, P., Wünsch, R., & Langer, N. 2022, A&A, 658, A125

    2026

  29. [29]

    & Muñoz-Tuñón, C

    Tenorio-Tagle, G. & Muñoz-Tuñón, C. 1997, ApJ, 478, 134

    1997

  30. [30]

    & Muñoz-Tuñón, C

    Tenorio-Tagle, G. & Muñoz-Tuñón, C. 1998, MNRAS, 293, 299 van der Wel, A., Franx, M., van Dokkum, P. G., et al. 2014, ApJ, 788, 28

    1998

  31. [31]

    2026, A&A, 705, A171

    Vanzella, E., Messa, M., Adamo, A., et al. 2026, A&A, 705, A171

    2026

  32. [32]

    A., Bakx, T., Mitsuhashi, I., et al

    Zavala, J. A., Bakx, T., Mitsuhashi, I., et al. 2024, ApJ, 977, L9

    2024

  33. [33]

    2023, MNRAS, 520, 2445 Article number, page 4 of 4

    Ziparo, F., Ferrara, A., Sommovigo, L., & Kohandel, M. 2023, MNRAS, 520, 2445 Article number, page 4 of 4

    2023

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.