{"id":"d9c6e9ac-f008-492b-9c5a-ba146a40814d","arxiv_id":"1909.01360","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Stellar radiation feedback reduces the number of satellites above 10^7 solar masses inside a z=6 quasar host halo by 40-60% in radiation-hydrodynamic simulations.","lead":"Simulations of a massive galaxy at redshift 6 show that radiation from young stars makes its future satellite galaxies puffier and easier for tidal forces to tear apart. The result predicts that JWST may find fewer small companion galaxies around bright quasars than expected, and that this deficit could signal strong early feedback rather than a lighter halo.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reported 40–60% satellite reduction is not supported by the paper's own cumulative counts: 102±8 vs 73±5 is a 28% drop at Mstar>1e7, and 26±2 vs 16±2 is a 38% drop at Mstar>1e8.","rationale":"The reader's weakest assumption focused on halo-sample representativeness and subgrid convergence. That is a legitimate external-validity concern. However, the more immediate problem is internal: the quantitative headline is contradicted by the paper's own numbers. The abstract says the Mstar > 1e7 satellite count drops by up to 60%, while the quoted means imply 28% at that threshold and 38% at 1e8. Since the central claim is specifically about this amplitude, the paper should present a consistent calculation before the number is used as an observational prediction. I still regard the differential mechanism as plausible and well motivated: the two simulations differ only in radiative feedback, particle tracking shows more complete tidal disruption of the selected z=8 progenitors in SN+RT, and the outside-Rvir counts are similar. These elements support the direction of the effect. The conditional verdict remains appropriate, but the condition should include correcting or explicitly re-scoping the 40–60% figure, not only extending the halo sample and resolution. No code or data release is available to resolve the discrepancy from the text alone, which strengthens the need for the proposed numerical check.","tokens_in":17144,"tokens_out":7457,"duration_ms":73715,"concrete_test":"Reconstruct the cumulative N(>Mstar) curves at z=6 from the snapshot data (or digitize Fig. 2) and compute (SN−SN+RT)/SN with Poisson errors at Mstar = 1e7 and 1e8, both at fixed z=6 and time-averaged over 6.1<z<6. Also compute the reduction per differential mass bin. If no threshold or bin reaches 50%, revise the abstract's 'up to 60%' and Section 3's '40–60%' to the actual values; if 60% is confined to a narrow high-mass bin, that scope must be stated explicitly instead of being attached to Mstar > 1e7.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3 and the abstract make the central quantitative claim that stellar radiation reduces the number of satellites with Mstar > 1e7 Msun within Rvir by 'up to 60%' / '40–60%'. The same section reports the cumulative mean counts from Fig. 2: 102±8 (SN) vs 73±5 (SN+RT) for Mstar > 1e7, i.e. a reduction of 29/102 ≈ 28%, and 26±2 vs 16±2 for Mstar > 1e8, i.e. 10/26 ≈ 38%. The text derives the 40–60% figure from 'an excess of ≈30 massive clumps' (which is 29%) and '≈10' (38%). Only the earlier method-(i) halo-based count (6 vs 2 systems with Mstar > 1e8 inside Rvir) gives a ~67% drop, but that is not the stellar-clump census on which the headline is based. So the paper's own reported statistics do not support the 40–60% amplitude at the thresholds stated in the abstract. This is an internal arithmetic inconsistency, independent of the acknowledged single-halo, resolution, and subgrid uncertainties. The qualitative mechanism may still be robust, but the headline number is currently unverified by the paper's own data.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents two cosmological zoom-in radiation-hydrodynamic simulations of a rare z=6 halo with Mvir≈2.4×10^12 Msun, differing only in whether stellar radiative feedback is included (SN vs SN+RT). The authors report that stellar radiation makes satellite progenitors less tightly bound before supernovae, so that they are more easily tidally disrupted once they enter the massive halo. Consequently, the number of satellites and stellar clumps with Mstar>10^7 Msun inside the virial radius is claimed to drop by 40–60%, and the central galaxy becomes more extended and diffuse (Reff increases by a factor ≲3). The paper argues that high-redshift quasar hosts may therefore show anomalously few luminous satellites, a prediction testable with JWST.","tokens_in":17372,"tokens_out":3045,"duration_ms":30488,"significance":"If the central claim holds, the paper identifies a genuinely important mechanism: the timing of stellar radiative feedback relative to supernovae can control satellite survival in massive high-z haloes, with direct observational consequences for quasar environments. The experimental design is a strength: two simulations that differ only in the inclusion of stellar radiative feedback, with galaxies identified by two independent methods (halo-based and stellar-particle-based), and the qualitative difference in satellite counts is visible in both. The paper also provides a plausible causal chain (lower binding energy at z≈7–8, shorter tidal disruption timescale) supported by particle-ID matching between the runs. The prediction of a diffuse stellar envelope and reduced satellite counts around z>6 quasars is falsifiable with upcoming JWST data. The main weaknesses are that the quantitative amplitude rests on a single halo realization and on numerical choices that are not tested for convergence, and that the reported 40–60% figure does not follow from the paper's own cumulative counts.","major_comments":[{"comment":"The headline quantitative claim is not supported by the paper's own numbers. The caption to Fig. 2 reports mean cumulative counts within Rvir of 102±8 (SN) vs 73±5 (SN+RT) for Mstar>10^7 Msun, which is a reduction of 29/102≈28%, not 40–60%. For Mstar>10^8 Msun the counts are 26±2 vs 16±2, a 38% reduction. The text's own description, 'an excess of ≈30 massive clumps' and '≈10', corresponds to these same 28% and 38% reductions. The abstract states 'drops by up to 60%' and the text states '≈40%−60% reduction', for which the only basis appears to be the method-(i) halo-based count of 6 vs 2 massive satellites, a different galaxy census from the cumulative counts quoted in the same section. This is an internal arithmetic inconsistency in a load-bearing claim, and the headline number must be corrected or explicitly re-derived.","section":"Section 3, Fig. 2 and Abstract"},{"comment":"The quantitative amplitude of the satellite suppression rests on a single halo realization: the second most massive halo in a 500 h^-1 Mpc box at z=6. The paper acknowledges the need for larger samples in Section 4, but does not provide any estimate of cosmic variance from even one additional halo, nor any resolution convergence test. Since the minimum cell size is 40 pc (Section 2) and the result depends on resolving the internal structure of 10^10 Msun subhaloes (which are only marginally resolved in a 40 pc cell), the 40–60% amplitude could be substantially altered at higher resolution. The direction of the effect is likely robust, but the amplitude central to the paper's abstract is not demonstrated to be converged or representative.","section":"Section 2 and Section 4, 'second most massive halo' and 'larger number of massive z=6 haloes'"},{"comment":"The simulations deliberately exclude AGN feedback, yet the paper frames the target halo as a likely quasar host. While this choice is justified for isolating stellar radiative feedback, it leaves open the possibility that AGN radiation or outflows, if switched on at earlier times, could either amplify or erase the reported suppression. The discussion in Section 4 acknowledges this only qualitatively ('it remains possible for the process outlined here to be amplified'). The paper should state more clearly that the prediction for real quasar hosts is conditional on the unknown AGN onset time and strength.","section":"Section 2, 'We therefore exclude radiation and mechanical feedback from AGN'"}],"minor_comments":[{"comment":"The abstract contains a duplicated word: 'the strength of stellar radiative feedback and and can be anomalously low'.","section":"Abstract"},{"comment":"The sentence 'the total stellar mass in SN+RT is at most times lower by20−40%' is awkwardly worded; 'at most times' should be 'at most times' or 'typically lower by 20–40%'.","section":"Section 3, fourth paragraph"},{"comment":"The caption states 'the number of galaxies within the halo drops by up to 40%' while the text and abstract quote '40%−60%' and 'up to 60%'. These should be made consistent.","section":"Section 3, Fig. 2 caption"},{"comment":"The table header 'T able 1' contains a spacing typo.","section":"Table 1"},{"comment":"The redshift range in Fig. 2 is described as 'between 6.1<z<6', which reverses the inequality direction; it likely means 6<z<6.1.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper's core mechanism is interesting and the differential simulation setup is clean, but the published abstract and text overstate the measured effect by roughly a factor of two. Since the quantitative claim is the main headline, the authors should be asked to recompute and restate the reduction percentages using the actual counts from their Fig. 2, and to clarify which selection (halo-based or clump-based) underlies the abstract. The single-halo and no-resolution-convergence limitations are acknowledged in the text but should be reflected in the abstract's strength of the claim (e.g., 'up to ~40%' rather than 'up to 60%')."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. The mechanism is worth taking seriously, but the headline number is not supported by the paper's own cumulative counts. The abstract says the number of satellites with M* > 1e7 within Rvir drops by up to 60%, and Section 3 repeats 40–60%. But the numbers reported there are 102±8 (SN) vs 73±5 (SN+RT), a 28% drop, and for M*>1e8, 26±2 vs 16±2, a 38% drop. The 40–60% figure appears to come from a different analysis (the 6 vs 2 halo-based count, which is 67%) and from the average over snapshots, but the stated means and the figure caption don't add up. This is an arithmetic inconsistency, not a subtle modeling issue, and it should be fixed before anyone quotes the amplitude.\n\nWhat the paper does well: it's a clean differential simulation. Two runs identical except for stellar radiative feedback. The physical story is plausible and supported by particle tracking: radiation acts before supernovae, makes satellite progenitors less bound, and they are tidally shredded sooner. The authors are honest about the main caveats—single halo, no resolution convergence, AGN feedback excluded—and they frame the result as a mechanism rather than a firm prediction. The extension to quasar environments and the predicted diffuse stellar envelope is a useful observational handle.\n\nSoft spots, in proportion: the arithmetic issue is load-bearing because the abstract and the Section 3 summary make the amplitude the takeaway. Beyond that, the single-halo realization really does limit the quantitative claim; cosmic variance could easily be comparable to the effect. The reduced speed of light (0.03c) and dust opacities are standard choices but are not convergence-tested. AGN exclusion is appropriate for isolating stellar feedback, but real quasar hosts have AGN, so the observational extrapolation is speculative. None of this kills the qualitative result—the direction of the effect and the physical mechanism are solid enough for a referee to take seriously.\n\nWho this is for: anyone working on high-redshift galaxy formation, satellite demographics, or interpreting JWST counts around z>6 quasars. It deserves a serious referee, but the authors need to correct the percentage and either add a convergence test or explicitly soften the quantitative claim.","headline":"Clean differential simulation with a plausible mechanism, but the headline 40–60% satellite reduction is not backed by the paper's own mean counts.","tokens_in":17966,"tokens_out":2725,"would_cite":true,"duration_ms":24922,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Stellar radiation can hide up to 60% of the satellites around $z\\approx6$ quasars, simulations suggest.","keywords":["galaxy formation","satellite galaxies","high-redshift quasars","stellar radiative feedback","tidal disruption","radiation-hydrodynamic simulations","cosmological zoom-in simulations","z=6 massive haloes"],"falsifier":"Deep JWST imaging of a statistically meaningful sample of $z\\approx6$ quasar hosts would settle it: if the number of satellites with $M_\\star > 10^7\\,M_\\odot$ within roughly one virial radius is not tens of percent lower than simulations without stellar radiation predict, or if the predicted diffuse stellar envelope is absent, the central claim is falsified.","tokens_in":16861,"feed_emoji":"🌌","tokens_out":10340,"duration_ms":94855,"temperature":0.7,"pith_summary":"The paper aims to establish that ordinary starlight, not just supernovae, can dramatically reduce the number of satellite galaxies around the most massive galaxies at redshift $z\\approx6$, including the likely hosts of bright quasars. In two otherwise identical radiation-hydrodynamic simulations of a $\\sim 2\\times10^{12}\\,M_\\odot$ halo, turning on stellar radiative feedback lowers the mean number of $M_\\star > 10^7\\,M_\\odot$ systems inside the virial radius from about 102 to 73, a 40--60% drop, while leaving counts outside the virial radius almost unchanged. The proposed cause is that radiation acts within about a million years of star formation, before the 10-million-year supernova delay, puffing up young dwarf galaxies and making them less tightly bound; the host's tidal field then shreds them sooner. A sympathetic reader should care because deep JWST observations of $z>6$ quasars may find surprisingly few faint satellites, and that deficit could be mistaken for evidence of a lighter dark-matter halo rather than recognized as a fingerprint of early stellar feedback.","feed_headline":"Early starlight hides up to 60% of quasar satellites","feed_subtitle":"Simulations show radiation loosens dwarf galaxies before the host quasar's tides tear them apart.","key_machinery":"The load-bearing object is a matched pair of cosmological zoom-in radiation-hydrodynamic simulations of the same $\\approx 2.4\\times10^{12}\\,M_\\odot$ halo at $z=6$, identical except that one tracks stellar radiative feedback (ionizing and non-ionizing radiation, dust radiation pressure, and photo-heating) and the other does not. The mechanism is a timing-and-binding chain: radiation from young stars acts essentially immediately, while supernovae are delayed by 10 Myr, so radiation curbs the collapse of gas in low-mass progenitors and leaves them less tightly bound; when these dwarfs fall into the host's deep potential well, where peak stellar circular velocities exceed $700\\,\\mathrm{km\\,s^{-1}}$, tidal forces disrupt them on a shorter timescale than in the no-radiation run, converting many of them into diffuse stellar material rather than surviving cores.","core_discovery":"On the paper's own terms, the central discovery is that stellar radiative feedback sets the satellite population of a massive high-redshift galaxy by acting before supernovae. Including photo-ionization, photo-heating, and radiation pressure produces galaxies that are slightly less massive and distinctly less concentrated: at early times the mean peak stellar circular velocity of galaxies in haloes above $10^{10}\\,M_\\odot$ drops from about 65 to $35\\,\\mathrm{km\\,s^{-1}}$. The same tidal forces that strip dark matter from infalling dwarfs then destroy their stellar cores as well, so that many systems that survive as compact clumps of mass $10^7$--$10^9\\,M_\\odot$ in a supernova-only run are completely dispersed in the radiative run. The result is a 40--60% reduction in satellites with $M_\\star > 10^7\\,M_\\odot$ within the virial radius, a nearly unchanged population outside it, and a more diffuse, smoother central galaxy with an effective radius larger by a factor $\\lesssim 3$.","pith_inferences":["The paper leaves implicit that the same loosen-then-shred mechanism should be testable as a function of host mass: if the host halo is less massive, the suppression should weaken, so satellite counts could become a quantitative probe of both halo mass and early feedback strength.","I would extend the argument to quasar radiation: since the paper deliberately turns AGN feedback off, adding quasar radiation could amplify or alter the suppression, a direct and testable next step rather than a result of the present runs.","A further inference is that present-day massive galaxies may retain the signature of this process in their diffuse stellar haloes, built partly from dwarf progenitors whose binding energy was lowered by reionization-era starlight before they were tidally shredded."],"forward_implications":["JWST-class imaging of $z>6$ quasar fields should find noticeably fewer $M_\\star > 10^7\\,M_\\odot$ satellites inside the virial radius than supernova-only galaxy formation models predict.","The missing satellites should reappear as diffuse stellar light, tidal streams, and an extended stellar envelope out to roughly 10 kpc around the quasar host.","Surviving satellites in these extreme environments should be preferentially compact, irregular, low-mass cores, since extended systems are shredded first.","A low satellite count around a quasar would not by itself imply a light host halo; the same halo mass yields different counts depending on when early feedback acted.","The suppression should be weaker around less massive or lower-redshift galaxies, where tidal fields are weaker."],"supporting_citations":[{"why":"It motivates the choice of a $\\sim 2\\times10^{12}\\,M_\\odot$ halo as a $z\\approx6$ quasar-host candidate by showing that efficient supermassive black hole growth requires such massive haloes.","marker":"Costa et al. 2014"},{"why":"It provides the parent cosmological volume from which the target halo is drawn as the second most massive halo at $z=6$.","marker":"Costa et al. 2018"},{"why":"It sets the radiative-transfer prescription, five-bin spectra, dust opacities, and the reduced speed of light used in the run with stellar radiation.","marker":"Rosdahl et al. 2015"},{"why":"It supplies the Schmidt-law star-formation prescription with variable efficiency used in both runs.","marker":"Kimm et al. 2017"},{"why":"It supplies the supernova feedback model, depositing thermal energy or momentum, applied identically in both runs.","marker":"Kimm et al. 2015"},{"why":"It describes the radiative transfer module that allows stellar radiation to be followed self-consistently.","marker":"Rosdahl et al. 2013"},{"why":"It provides the homogeneous, time-evolving ultraviolet background adopted for external ionizing flux.","marker":"Faucher-Giguère et al. 2009"},{"why":"It provides the base adaptive-mesh-refinement hydrodynamics and N-body solver on which the simulations run.","marker":"Teyssier 2002"}],"fun_headline_variants":["Early starlight makes dwarf satellites fragile, cutting their numbers by 60%","Radiation loosens dwarf galaxies, letting quasar tides shred them—satellites drop 60%","Pre-supernova starlight weakens dwarfs, up to 60% fewer satellites in quasar halos","Stellar glow strips dwarf cores before tidal death, slicing satellite counts by 60%","Early radiative feedback loosens dwarfs, cutting quasar satellite numbers by 60%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative result rests on treating a single simulated halo, with AGN feedback deliberately switched off, as representative of real $z\\approx6$ quasar environments; if that halo is atypical, or if AGN radiation or subgrid choices change how tightly bound dwarf progenitors are, the 40--60% amplitude could change even if the direction of the effect is robust.","fun_headline_variants_meta":{"raw":{"variants":["Early starlight makes dwarf satellites fragile, cutting their numbers by 60%","Radiation loosens dwarf galaxies, letting quasar tides shred them—satellites drop 60%","Pre-supernova starlight weakens dwarfs, up to 60% fewer satellites in quasar halos","Stellar glow strips dwarf cores before tidal death, slicing satellite counts by 60%","Early radiative feedback loosens dwarfs, cutting quasar satellite numbers by 60%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000678,"raw_usage":{"total_tokens":3106,"prompt_tokens":990,"completion_tokens":2116,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":606,"completion_tokens_details":{"reasoning_tokens":1995}},"tokens_in":606,"tokens_out":2116,"duration_ms":15102,"temperature":1.0,"reasoning_tokens":1995,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:20:08.508657+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Deep JWST imaging of a statistically meaningful sample of $z\\approx6$ quasar hosts would settle it: if the number of satellites with $M_\\star > 10^7\\,M_\\odot$ within roughly one virial radius is not tens of percent lower than simulations without stellar radiation predict, or if the predicted diffuse stellar envelope is absent, the central claim is falsified.","supporting_citations":[],"review_version":1}