{"id":"b4cb74eb-0a48-4a61-9ebd-1a94f4af37ae","arxiv_id":"2411.16876","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Supernova feedback strength, not quasar black hole feedback, is the main regulator of when and whether Milky Way-like galaxies form stellar bars.","lead":"A team resimulated six Milky Way-like galaxies with different supernova and black hole feedback strengths to test what makes galactic bars form. They found stronger supernova winds delay or even prevent bars, while black hole feedback has little effect on bar formation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The causal attribution to SN feedback strength is not yet separated from subgrid stochasticity: with no repeated realizations, the one-of-six loss of the bar in the TNG50-like control and the acknowledged scatter in the stability quantities (Sec.","rationale":"The paper is transparent and appropriately cautious; Sec. 5.4 already states that stochasticity limits the physical conclusions and that five strong-bar galaxies preclude general conclusions. That is why the reader's CONDITIONAL verdict is appropriate. My stress-test isolates the narrowest load-bearing condition for the causal claim in the abstract: the feedback-model contrast must exceed the intrinsic stochastic scatter. The suite has no seed repeats, and the TNG50-like control itself fails to reproduce one of the six parent bars, so the signal-to-noise ratio of the bar-presence and bar-formation-time comparisons is unknown. The proposed control runs measure exactly this. Selection on parent bars makes the suppression finding conservative rather than circular; the unrealistic NW model is used mainly to illustrate criterion failure and is acknowledged as such. I therefore do not change the verdict: it should remain CONDITIONAL until seed-variation control runs are provided.","tokens_in":29193,"tokens_out":5359,"duration_ms":50921,"concrete_test":"Re-run the six halos in at least the TNG50-like and SW models with three or more independent random seeds for the Arepo stochastic subgrid draws (star formation and wind injection directions), keeping gravity and the initial conditions fixed, and compare bar presence, A2,max, and zbf across seeds. If any SW halo forms a bar in any seed, or the TNG50-like bar count fluctuates by more than one galaxy, the claimed feedback-driven delay/suppression is not separated from stochasticity. A cheaper first step: re-seed only galaxy 543376 in the TNG50-like model and check whether it sometimes retains a bar.","verdict_should_be":"UNCHANGED","load_bearing_attack":"To support the abstract's claim that SN energy delays bar formation and that the strongest feedback forms stable discs, the across-model differences in bar outcome must be caused by the varied feedback parameters (ew, epsilon_f,high), not by chaotic divergence of identical initial conditions. That condition is untested: the 42 runs contain one realization per (halo, model), with no seed variations. The authors' Sec. 5.4 concedes that subgrid stochasticity can change the radial velocity dispersion and the disc extent, exactly the quantities entering QT, epsilon_ELN, and lambda_MMW (Figs. 4, 5, 8), and that one of the six TNG50-like controls (galaxy 543376, the weakest parent bar with A2,max = 0.38) fails to keep its bar. Since binary bar/no-bar outcomes change across adjacent wind models (MW: 1/6 barred; TNG50-like: 5/6; WW: 6/6) and BH-model bar-formation times vary 'dramatically' (Sec. 3.2), a single stochastic realization cannot distinguish a systematic feedback effect from run-to-run noise. The reported trend may be real, but the load-bearing causal claim is currently underdetermined by the published runs.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper resimulates six TNG50 Milky Way-mass barred disc galaxies with the Arepo code and seven variants of the TNG feedback model, varying the galactic wind energy parameter e_w and the quasar-mode BH feedback efficiency ε_f,high. The authors find that stronger supernova feedback delays disc assembly and suppresses bar formation, with the strongest wind model producing stable discs, while the no-wind model fails to form a bar because of a massive compact bulge. Quasar BH feedback variations do not change whether a bar forms but do change bar strength and length. The paper evaluates the Toomre, ELN, and MMW instability criteria at bar formation and shows they correctly predict bar presence/absence except in the no-wind and weak-wind cases, which they attribute to the missing effect of a central bulge.","tokens_in":29374,"tokens_out":10417,"duration_ms":91919,"significance":"The study is a clean controlled parameter experiment on a well-characterised simulation sample and provides a useful test of analytic bar-instability criteria under controlled variations of subgrid physics. The explicit discussion of stochasticity in Section 5.4 is honest, and the suite of 42 zoom-in runs is a valuable resource. If the causal attribution is confirmed with additional realizations, the conclusion that SN feedback strength is the dominant regulator of bar formation in MW-mass systems would constrain subgrid models. However, as presented, the central causal claim is not yet separated from run-to-run noise, which limits the strength of the conclusions that can be drawn.","major_comments":[{"comment":"The central claim that supernova feedback strength causally delays and can prevent bar formation is underdetermined by the single realization per (halo, model). The paper's own Section 5.4 states that subgrid stochasticity can affect the radial velocity dispersion and disc extent, which are exactly the quantities entering Q_T, ε_ELN, and λ_MMW (Figs. 4, 5, 8), and that one of six control galaxies loses its bar in the TNG50-like model. With one run per model, the binary bar/no-bar outcomes across adjacent wind models (WW 6/6, TNG50-like 5/6, MW 1/6, SW 0/6, NW 0/6) are not statistically distinguishable from noise: for example, the Clopper-Pearson 95% intervals for 1/6 and 5/6 overlap. Please either provide multiple stochastic realizations (different random seeds) for at least the models near the bar/no-bar boundary, or temper the abstract and Section 6 to say the effect is seen 'in our realizations' and justify quantitatively why the trend exceeds stochastic scatter.","section":"Sec. 5.4; Abstract"},{"comment":"The statement in the Abstract that quasar-mode BH feedback 'does not affect bar formation' is not consistent with the text in Section 3.2 that 'the bar is present in all simulations, but the bar formation time varies dramatically' across the BH models. If 'formation' is meant as the presence of a bar, the claim should be reworded to 'does not affect whether a bar forms'; if it includes formation time, the statement is internally contradictory. Given the stochastic scatter conceded in Section 5.4, this dramatic variation also needs to be assessed against run-to-run noise before a null effect on formation is claimed.","section":"Sec. 3.2; Abstract"}],"minor_comments":[{"comment":"The mass-loading relation in Eq. (5) appears dimensionally incorrect: η_w = 2 v_w^2 e_w (1−τ_w) should read η_w = 2 e_w (1−τ_w)/v_w^2, consistent with the text describing available wind energy divided by wind kinetic energy.","section":"Eq. (5)"},{"comment":"In the sentence beginning 'For the galaxies in the models that form a bar', the expression '(λ_MMW/λ_crit) ∼< 12' is presumably a typo for '(λ_MMW/λ_crit) < 1'; as written, it is not meaningful for the stability threshold.","section":"Sec. 4.2"},{"comment":"The acronyms 'WWM' and 'ELT' appear where 'MMW' and 'ELN' are intended (Section 5.3 and the summary bullet list); please correct.","section":"Sec. 5.3 and Sec. 6"},{"comment":"The sentence 'however, the radial velocity dispersion and the extent of the disc could be' is incomplete; it should end with 'affected'.","section":"Sec. 5.4"},{"comment":"The Abstract should state explicitly that the sample is preselected to be barred in TNG50, so the conclusions concern galaxies that form bars under the TNG50 model; the limitation is only stated in Section 5.4.","section":"Sec. 2.1; Abstract"}],"recommendation":"major_revision","confidential_remarks":"The paper is fundamentally a carefully executed parameter study, and the authors are transparent about the stochasticity limitation in Section 5.4. My main concern is that the abstract overstates the causal claim relative to the evidence in the paper. If the authors can provide even a small number of seed variations for the borderline models, or explicitly reframe the conclusions as trends in single realizations, the paper would be acceptable. I recommend major revision rather than rejection because the underlying simulations and analysis are careful and the limitation is acknowledged."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid, transparent parameter study that mostly confirms earlier work. Its genuinely new piece — three classical instability criteria failing in opposite directions for extreme feedback — is worth taking seriously. The central causal claim about supernova wind strength is plausible but not fully separated from stochasticity, because there is one realization per halo and model.\n\nWhat is new: six TNG50 barred Milky Way-like galaxies, each resimulated with seven feedback variants (five wind strengths and two black-hole quasar-mode efficiencies), 42 runs. The qualitative results agree with Zana et al. 2019 and Irodotou et al. 2022, but this is a cleaner controlled comparison: same initial conditions, same code, same analysis pipeline. The criteria test is the real contribution. In the no-wind model a massive compact bulge forms, no bar appears, yet ELN and MMW say the disc should be unstable. In the weak-wind model the Toomre criterion says stable while a bar does form. That is a concrete demonstration, tied to Saha & Elmegreen and Kataria & Das.\n\nSoft spots: the stress-test concern lands. Section 5.4 concedes that subgrid stochasticity affects radial velocity dispersion and disc extent, exactly the quantities behind QT, epsilon_ELN, and lambda_MMW, and one of the six TNG50-like control galaxies loses its bar. With one seed per model, the binary bar/no-bar differences across adjacent wind strengths could be partly chaotic. The monotonic trend across five wind strengths makes me think the feedback effect is real, but the abstract's wording (\"energy released by the supernovae causes a delay...\") goes a step beyond what the runs can support. A seed-variation test on a couple of haloes would fix this, or the authors should soften the causal language. The sample is selected on parent TNG50 bars, so it says nothing about the unbarred population. No data or initial-condition code are released, which limits independent replication, though the methods are standard. The no-wind model is unrealistic, and the authors say so.\n\nCitation pattern is fair; prior work is credited. This is not a major discovery, but it is a useful, honest reference. Send it to a serious referee; the referee should push on stochasticity and on the strength of the abstract's causal claim.","headline":"Transparent feedback-variation study whose criteria-failure result is worth citing, but whose causal claim outruns a one-seed-per-model design.","tokens_in":29997,"tokens_out":3804,"would_cite":true,"duration_ms":37171,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Supernova feedback, not black-hole feedback, sets whether a Milky Way-like disc grows a bar.","keywords":["galactic bars","supernova feedback","AGN feedback","disc stability","zoom-in simulations","cosmological hydrodynamics","Milky Way analogues","bar instability criteria"],"falsifier":"A decisive test would be to re-run the same six haloes with several independent random seeds for both the TNG50-like and strong-wind models: if any strong-wind realisation forms a stable bar while the matching TNG50-like run does not, the claimed suppression would not hold up. An observational falsifier would be to find Milky Way analogues with extended, low-velocity-dispersion discs that still host strong bars at the same rate as compact discs, contrary to the stability picture.","tokens_in":28918,"feed_emoji":"🌌","tokens_out":12096,"duration_ms":105506,"temperature":0.7,"pith_summary":"Galactic bars are the elongated stellar structures seen in most disc galaxies, including the Milky Way. The paper asks which galaxy-formation physics decides whether such a bar forms, and it argues that the answer is largely the strength of supernova feedback. In a suite of 42 zoom-in cosmological resimulations of six Milky Way-sized galaxies selected because the parent TNG50 simulation gave each a strong bar, the authors vary only the energy injected per supernova and the efficiency of quasar-mode black hole feedback. They report that supernova feedback alone changes whether and when bars appear: stronger winds delay or suppress bars, while the black-hole variations only alter the bar's strength and length. The same runs show that the classic disc-instability criteria used to predict bars fail exactly in the extreme cases where a massive compact bulge is present.","feed_headline":"Supernova feedback decides when galaxies grow bars","feed_subtitle":"Rerunning six barred Milky Way-like galaxies with different feedback strengths shows black holes only reshape the bar.","key_machinery":"The central objects are three analytic disc-instability diagnostics — the Toomre parameter $Q_T=\\kappa\\sigma_R/(3.36G\\Sigma)$, the ELN ratio $\\epsilon_{\\rm ELN}=v_{\\rm max}/(G M_{\\rm disc}/r_{\\rm disc})^{1/2}$, and the MMW spin criterion $\\lambda_{\\rm MMW}$ — together with a controlled zoom-in resimulation design in which only the supernova wind energy parameter $e_w$ or the quasar-mode feedback efficiency $\\epsilon_{f,\\rm high}$ is changed. The criteria are used to judge whether a simulated disc should be bar-unstable, and the feedback variations isolate which channel changes that judgement. Bars are identified by Fourier decomposition of the face-on stellar surface density, with strength $A_2$ and length set by the constant-phase radius.","core_discovery":"The paper's central claim is that, for Milky Way-like discs, bar formation is regulated by supernova feedback strength, while quasar-mode black hole feedback plays at most a secondary role. Each of six haloes is re-simulated in seven model variants covering five wind strengths (no wind, weaker, TNG50-like, medium, strong) and two black hole feedback variants, and the outcome is that only the supernova wind strength shifts bar formation time or suppresses the bar outright. Stronger winds delay disc assembly and produce less massive, more extended discs with lower radial velocity dispersion, which resist bar instability; the strongest winds stop bar formation entirely. At the opposite extreme, no wind leads to early runaway star formation and a massive compact bulge that also prevents a bar. The paper further argues that the Toomre, ELN, and MMW criteria correctly mark the bar-forming cases at the time of bar formation, but in the no-wind model ELN and MMW incorrectly predict instability while Toomre predicts stability for the weak-wind model that does form a bar, and the authors attribute these failures to the criteria not accounting for the central bulge.","pith_inferences":["The authors do not state this, but the result implies that surveys measuring bar fractions as a function of disc size and stellar mass could act as an observational probe of feedback strength in real galaxies.","The failure pattern of the three criteria suggests a simple repair: adding a bulge-concentration term to ELN and MMW should recover the no-wind case, and this could be tested directly on the same simulations.","Because stochasticity removed one of six bars even in the TNG50-like run, a stronger test of the feedback effect would come from re-running the same haloes with multiple random seeds; the authors flag this limitation, and seed-averaged comparisons would settle whether the feedback effect is real.","If the supernova-dominated picture is correct, high-redshift barred galaxies detected by JWST may preferentially be systems with weak feedback or low bulge concentration, a pattern that future high-redshift bar surveys could check."],"forward_implications":["Bar formation time in Milky Way-like galaxies should be earlier for weaker supernova-driven winds and suppressed entirely for the strongest winds, so observed bar fractions can encode the average supernova feedback strength.","Quasar-mode black hole feedback can be varied without changing whether a bar forms; its main effect is on bar strength and length.","The ELN and MMW criteria are not by themselves reliable predictors of bar formation when a massive compact bulge is present; a bulge-concentration condition is needed.","Disc-dominated morphology is robust across all feedback models, so a massive disc alone does not guarantee that a bar will form.","The statistics of barred galaxies can serve as a constraint on subgrid galaxy-formation models, since different feedback strengths give different bar fractions and formation redshifts."],"supporting_citations":[{"why":"Supplies the Toomre $Q_T$ stability criterion used as one of the three bar-instability diagnostics.","marker":"Toomre (1964)"},{"why":"Supplies the ELN criterion that the paper tests against the simulated discs.","marker":"Efstathiou, Lake & Negroponte (1982)"},{"why":"Supplies the MMW spin-based criterion used as the third instability diagnostic.","marker":"Mo, Mao, & White (1998)"},{"why":"Provides the compact-bulge limit ($B/D>0.3$) that the paper invokes to explain the no-wind suppression of bars.","marker":"Kataria & Das (2018)"},{"why":"Earlier evidence that feedback processes shape bar formation time and properties, motivating the feedback variations.","marker":"Zana et al. (2019)"},{"why":"Prior Auriga result that AGN feedback changes bar properties but not bar formation, which the paper compares with its BH models.","marker":"Irodotou et al. (2022)"},{"why":"Supplies the TNG50 bar catalogue from which the six strongly barred parent galaxies are selected.","marker":"Rosas-Guevara et al. (2022)"},{"why":"Describes the TNG galactic wind model whose energy parameter $e_w$ is varied in this suite.","marker":"Pillepich et al. (2018b)"},{"why":"Documents the stochastic butterfly effect of feedback that the paper cites when discussing run-to-run scatter.","marker":"Genel et al. (2019)"}],"fun_headline_variants":["Supernova feedback sets the timing of galactic bars","Black holes reshape bars, but supernovae decide birth","Stronger stellar winds delay or kill galaxy bars","Galactic bar formation hinges on supernova energy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the six selected haloes, chosen because the parent simulation made each one strongly barred, are a representative sample, so that differences between feedback models trace the changed physics rather than random stochastic scatter in the zoom-in re-simulations.","fun_headline_variants_meta":{"raw":{"variants":["Supernova feedback sets the timing of galactic bars","Black holes reshape bars, but supernovae decide birth","Stronger stellar winds delay or kill galaxy bars","Galactic bar formation hinges on supernova energy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000194,"raw_usage":{"total_tokens":1417,"prompt_tokens":1070,"completion_tokens":347,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":686,"completion_tokens_details":{"reasoning_tokens":286}},"tokens_in":686,"tokens_out":347,"duration_ms":3929,"temperature":1.0,"reasoning_tokens":286,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:47:22.842634+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be to re-run the same six haloes with several independent random seeds for both the TNG50-like and strong-wind models: if any strong-wind realisation forms a stable bar while the matching TNG50-like run does not, the claimed suppression would not hold up. An observational falsifier would be to find Milky Way analogues with extended, low-velocity-dispersion discs that still host strong bars at the same rate as compact discs, contrary to the stability picture.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Toomre $Q_T$ stability criterion used as one of the three bar-instability diagnostics."},{"cited_title":"K., Das M., 2018, MNRAS, 475, 1653","cited_arxiv_id":null,"evidence_quote":"Provides the compact-bulge limit ($B/D>0.3$) that the paper invokes to explain the no-wind suppression of bars."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier evidence that feedback processes shape bar formation time and properties, motivating the feedback variations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior Auriga result that AGN feedback changes bar properties but not bar formation, which the paper compares with its BH models."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the TNG50 bar catalogue from which the six strongly barred parent galaxies are selected."},{"cited_title":"L., Springel V., Hernquist L., Nelson D., Pillepich A., Weinberger R., et al., 2019, ApJ, 871, 21","cited_arxiv_id":null,"evidence_quote":"Documents the stochastic butterfly effect of feedback that the paper cites when discussing run-to-run scatter."}],"review_version":1}