{"id":"776a9ea2-35ae-4a4d-b195-9e2d50d8cdb3","arxiv_id":"1909.01230","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"A large sample of star-forming elliptical galaxies shows a significant tail of low gas-phase metallicity, indicating externally accreted gas in at least ~37% of such systems.","lead":"This letter measures gas-phase metallicities of 567 star-forming elliptical galaxies in SDSS and Galaxy Zoo and finds that about 7.4% are far more metal-poor than spirals of the same mass. The authors argue this low-metallicity gas was accreted from outside the galaxy and estimate that external accretion fuels residual star formation in at least 37% of these systems.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The '>37%' estimate hinges on a single toy-model visibility timescale with no error bars; plausible variations in ChemEvol inputs could shift the inferred accretion fraction by a factor of several.","rationale":"The paper's core observational detection is statistically strong: the excess of low-metallicity ellipticals over spirals (7.4% vs 1.7%) has bootstrap p < 1e-6, and the gas-to-stellar metallicity anti-correlation for outliers supports an external origin. However, the central quantitative claim of >37% accreted systems is not a direct measurement; it is the observed 7.4% multiplied by the ratio t_total/t_visible = 5, where t_visible comes from a single-zone toy model with hand-picked inputs and no uncertainty propagation. The paper itself labels the SFH 'somewhat arbitrary' and acknowledges that changing model inputs changes the derived timescale, yet no error bars or sensitivity analysis are given. This makes the headline number unsatisfactorily fragile: a plausible factor-of-two change in t_visible would move the estimate below the >37% threshold or far above it, so the specific numerical claim is not yet supported. The dilution alternative (stellar mass loss mixed with circumgalactic material) is also explicitly unresolved, further muddling the 'external accretion' interpretation. These concerns are real but do not invalidate the main detection or the direction of the effect; they warrant a CONDITIONAL verdict, which the reader already assigned. Hence the verdict is left unchanged.","tokens_in":10130,"tokens_out":4723,"duration_ms":42965,"concrete_test":"Run a sensitivity grid with ChemEvol, varying the initial gas mass (0.5–5×10^8 Msun), initial metallicity (0.02–0.3 Zsun), SFH shape (Gaussian sigma 100–1000 Myr and exponential timescales 0.1–1 Gyr), and with/without galactic winds. For each run, compute t_visible as the time for the model to move from >2σ below the Tremonti MZR to within the 0.26 dex scatter. Propagate the resulting range of t_visible through Equation 1. If the lower envelope of f_true falls below the claimed >37%, the headline number is not robust; if it remains >37% for all plausible inputs, the visibility-timescale concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative headline, 'at least 37%', is obtained from Equation 1: f_true = (t_total/t_visible) f_visible, using f_visible = 7.4% (Section 3.1), t_total = 2 Gyr (assumed spiral depletion time), and t_visible ≈ 400 Myr from a single ChemEvol run (Section 3.3). The visibility timescale depends on the assumed initial cold gas mass (5×10^8 Msun), initial metallicity (0.1 Zsun), host stellar mass (2.6×10^10 Msun) and stellar metallicity (0.89 Zsun), and a Gaussian SFH with sigma=500 Myr that the authors call 'somewhat arbitrary'. No error bars or sensitivity analysis are provided; the paper explicitly states that altering the SFH, initial metallicity, or IMF can change the derived timescale. Because the correction factor is 5, even a factor-of-two uncertainty in t_visible changes the claim from >18% to >75%, spanning the qualitative conclusion. Additionally, the dilution alternative (stellar mass loss mixed with low-metallicity CGM gas) is explicitly acknowledged in Section 4 as not ruleable-out with these data, directly weakening the attribution to discrete external accretion events rather than in-situ mixing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper combines SDSS DR7 spectroscopic catalogues (MPA-JHU gas-phase metallicities, masses and star formation rates; FIREFLY stellar metallicities) with Galaxy Zoo morphologies to compare the gas-phase metallicity distributions of star-forming elliptical and spiral galaxies. It reports that 7.4% of 567 star-forming ellipticals lie at least 2 sigma below the Tremonti et al. (2004) mass-metallicity relation, versus 1.7% of spirals, with a bootstrap probability below 1e-6. It further reports a Spearman correlation of 0.45 between the gas-to-stellar metallicity ratio and the residual from the mass-metallicity relation, and uses a ChemEvol chemical evolution model to estimate a visibility timescale of about 400 Myr for accreted low-metallicity gas. Combining the observed low-metallicity fraction with this timescale via Equation (1), the authors infer that more than 37% of gas-rich early-type galaxies have accreted their star-forming gas from an external low-metallicity source.","tokens_in":1501,"tokens_out":1799,"duration_ms":66471,"significance":"If the result holds, the paper offers a novel and independent constraint on the gas supply mechanisms of early-type galaxies, with a clean observational sample and a statistical excess that appears robust. The bootstrap significance of the 7.4% versus 1.7% excess and the strong correlation between low gas-phase metallicity and low gas-to-stellar metallicity ratio are genuine strengths; the inference is not circular because the observed fraction, the chemical evolution timescale, and the mass-metallicity baseline are independent inputs. However, the quantitative headline 'at least 37%' is derived from a single toy-model visibility timescale with no propagated error bars, and the paper explicitly acknowledges a mixing/dilution alternative that it cannot exclude. The central qualitative conclusion is plausible and well aligned with independent kinematic studies, but the manuscript currently overstates the precision and uniqueness of the accretion interpretation.","major_comments":[{"comment":"The headline fraction f_true = (t_total/t_visible) f_visible depends linearly on a visibility timescale t_visible = 400 Myr that comes from a single ChemEvol run with explicitly ad hoc inputs: a Gaussian SFH with sigma = 500 Myr described as 'somewhat arbitrary', initial gas metallicity 0.1 Zsun, initial gas mass 5e8 Msun, and no outflows. The correction factor is 2 Gyr / 400 Myr = 5, so even a factor-of-two uncertainty in t_visible moves the inferred fraction from about 18% to about 75%, changing the qualitative strength of the claim. Because the authors state that altering the SFH, initial metallicity, or IMF can change the derived timescale, the paper should either provide a sensitivity analysis (e.g., a grid over these inputs) or present the 37% value as an order-of-magnitude estimate rather than a quantitative prediction.","section":"Sec. 3.3, Eq. (1)"},{"comment":"The authors explicitly state that they 'cannot rule this possibility out' for high-metallicity stellar mass loss being diluted by lower-metallicity circumgalactic gas. That alternative is also capable of producing both low gas-phase metallicity relative to the mass-metallicity relation and low gas-to-stellar metallicity ratios, so the statement in Section 3.2 that the material 'must have come from an external source' overstates the uniqueness of the accretion interpretation. Moreover, if some of the 42 low-metallicity outliers are dilution rather than accretion cases, the inferred f_true from Equation (1) is not a lower limit on external accretion but an upper limit on the truly accreted fraction. The authors should either bound the dilution contribution with additional data (e.g., resolved kinematics) or soften the attribution throughout the abstract and conclusions.","section":"Sec. 4"},{"comment":"The 7.4% fraction is measured among the 567 Galaxy Zoo ellipticals that have detectable star-forming gas and measurable gas-phase metallicities, yet Equation (1) is applied to 'gas rich ETGs' without a quantitative selection correction. The authors note that star-forming-dominated objects may have a lower accreted fraction than objects whose ionisation is dominated by old stars (citing Belfiore et al. 2017), but they do not propagate this into the estimate. The abstract and conclusions therefore state the 37% result with more generality than the sample selection strictly supports; the claim should be phrased as applying to star-forming gas-rich early-type galaxies of this selection, or a quantitative selection-bias term should be added to Equation (1).","section":"Sec. 3.1 and Sec. 4"}],"minor_comments":[{"comment":"The paper uses '>37%', 'at least 37%', and '37.5%' interchangeably; one rounded value with an explicit uncertainty range should be adopted consistently.","section":"Abstract and Sec. 4"},{"comment":"There is a typo, 'we make used of the Sloan Digital Sky Survey', which should read 'we make use of'; similar minor grammar issues appear in the Figure 3 caption ('typical have') and Figure 4 caption ('caused by 5e8 Msun of cold gas', missing 'of').","section":"Sec. 2"},{"comment":"The significance of the Spearman correlation is quoted as '>10 sigma'; because this is not a standard Gaussian test statistic, the authors should specify whether the significance comes from a permutation/bootstrap test or from a normal approximation of the null distribution.","section":"Sec. 3.2"},{"comment":"A machine-readable table listing the 42 low-metallicity outliers with their SDSS identifiers, masses, gas-phase and stellar metallicities, and SFRs would substantially improve reproducibility, since all population-level claims presently rest on aggregate statistics.","section":"Sec. 3.1"}],"recommendation":"major_revision","confidential_remarks":"This is a well-scoped and interesting letter, and the observational excess is likely robust. The main issue is not the statistical analysis but the unquantified systematic uncertainty in the toy-model timescale that drives the central 37% claim. I consider this fixable within the scope of the manuscript through a sensitivity analysis and a more careful statement of the dilution alternative; hence major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth your time. The genuinely new thing here is the population-level gas-phase metallicity distribution for star-forming ellipticals from SDSS plus Galaxy Zoo: 7.4% (42/567) lie at least 2 sigma below the Tremonti mass-metallicity relation, versus 1.7% of spirals, bootstrap p less than 1e-6. That excess is robust and is not just a restatement of earlier small-sample work. The correlation between gas-to-stellar metallicity ratio and position below the mass-metallicity relation (Spearman 0.45, more than 10 sigma) is also a clean observational clue that the low-metallicity gas is not stellar mass loss. Credit where due: the paper uses public data, the statistics are straightforward, and the authors are unusually blunt about their model's shortcomings.\n\nThe soft spot is exactly where the stress test says it is. The 7.4% is a lower limit; converting it to greater than 37% via f_true = (t_total/t_visible) f_visible requires t_visible of about 400 Myr from a single ChemEvol run with hand-picked inputs: initial gas mass 5e8 solar masses, metallicity 0.1 solar, Gaussian SFH with 500 Myr width, which they call 'somewhat arbitrary'. No error bars or sensitivity scan are provided. If t_visible varies by a factor of two, the inferred fraction swings from roughly 18% to 75%. The qualitative conclusion that accretion matters is probably safe, but the number 37% should be read as illustrative. Their own caveat about dilution of stellar mass loss by low-metallicity CGM gas is a real alternative, and they concede it cannot be ruled out with these data.\n\nOne smaller concern: the sample is selected to have star-formation-dominated emission, so it probes gas-rich, currently forming ETGs. The authors discuss why this might bias the accreted fraction low, but it is still a selected subset.\n\nNone of this undermines the paper. The central observational detection is sound, the model dependence is openly flagged, and the agreement with independent kinematic estimates (roughly 42-45% accretion from Davis et al. 2011 and Bryant et al. 2019) is reasonable supporting context. The citation pattern looks appropriate; self-citations are to relevant prior work.\n\nRecommendation: send it to review. The referee should ask for a sensitivity analysis of the ChemEvol inputs and a clearer statement that 37% is a model-dependent estimate, not a measured fraction. The paper deserves serious refereeing because it adds a new, clean population-level constraint on a long-debated question. I would cite the 7.4% tail result; I would not cite the 37% number without a caveat.","headline":"Robust new low-metallicity tail in star-forming ellipticals, but the >37% accretion fraction is a toy-model extrapolation that should not be treated as firm.","tokens_in":10959,"tokens_out":2082,"would_cite":true,"duration_ms":19973,"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":"This paper claims that more than a third of star-forming elliptical galaxies acquire the gas for their residual star formation from outside the galaxy, not from recycled stellar mass loss.","keywords":["early-type galaxies","gas-phase metallicity","mass-metallicity relation","residual star formation","gas accretion","Galaxy Zoo","chemical evolution model","stellar metallicity"],"falsifier":"Resolved optical or cold-gas spectroscopy of a handful of the 42 low-metallicity ellipticals could settle the matter: if the gas is kinematically aligned with the stars and its abundance pattern matches enriched stellar mass loss, the external-accretion attribution collapses; if the gas is misaligned or disturbed and uniformly metal-poor, the paper's interpretation stands.","tokens_in":9898,"feed_emoji":"🔭","tokens_out":7698,"duration_ms":66066,"temperature":0.7,"pith_summary":"The paper sets out to determine where elliptical galaxies obtain the gas that powers their low-level, residual star formation. By combining SDSS spectra with Galaxy Zoo morphologies, it assembles 567 star-forming ellipticals and finds that about 7.4% of them have gas-phase metallicities far below the mass–metallicity relation, while their gas is also more metal-poor than their stars. A simple chemical evolution model shows that this low-metallicity signature disappears within roughly 400 million years once star formation begins, so the observed 7.4% is only a snapshot. Correcting for this visibility effect, the authors conclude that at least 37% of gas-rich ellipticals have accreted their star-forming gas from an external, low-metallicity source. If true, cosmological gas accretion and minor mergers, not recycled stellar mass loss, dominate the fuel supply for residual star formation in early-type galaxies.","feed_headline":"At least 37% of star-forming ellipticals fuel up from outside","feed_subtitle":"Accretion and minor mergers, not stellar mass loss, fuel residual star formation in ellipticals.","key_machinery":"The machinery is a two-part comparison. First, each galaxy is placed on the gas-phase mass–metallicity relation through the residual $\\Delta({\\rm O/H})$, the offset from the Tremonti et al. (2004) relation, and the ratio of gas-phase to stellar metallicity from FIREFLY; an object that is both well below the relation and less enriched in gas than in stars is identified as an accretion candidate. Second, the ChemEvol one-zone chemical evolution model (Morgan & Edmunds 2003; Rowlands et al. 2014; De Vis et al. 2017) tracks how a $5\\times10^8\\,M_\\odot$ reservoir of $0.1\\,Z_\\odot$ gas is enriched by a Gaussian star-formation episode, yielding the visibility timescale $t_{\\rm visible}\\approx400$ Myr. Equation 1, $f_{\\rm true} = (t_{\\rm total}/t_{\\rm visible}) f_{\\rm visible}$, then lifts the observed 7.4% fraction to the claimed at-least-37% true accretion fraction.","core_discovery":"Star-forming elliptical galaxies contain a population of low gas-phase metallicity outliers: 7.4% lie at least $2\\sigma$ below the Tremonti et al. (2004) mass–metallicity relation, compared with 1.7% of spirals, and these outliers have gas that is less enriched than their stellar photospheres. This combination is the paper's central evidence that the gas was accreted from outside rather than produced by stellar mass loss. Because chemical enrichment erases the low-metallicity signature in about 400 Myr, the paper converts the observed fraction into a true fraction using $f_{\\rm true} = (t_{\\rm total}/t_{\\rm visible}) f_{\\rm visible}$ with $t_{\\rm total} = 2$ Gyr, obtaining that at least 37% of gas-rich early-type galaxies have accreted their star-forming gas from an external low-metallicity source. The paper presents this as a lower limit that increases if depletion times in ellipticals are longer.","pith_inferences":["The ~400 Myr enrichment clock implies that many ellipticals observed today at normal gas metallicity may have accreted their fuel within the past gigayear; static surveys therefore underestimate the instantaneous accretion rate.","The same visibility correction could be applied to other quenched populations such as lenticulars and red spirals, where a similar low-metallicity tail would yield their external accretion fractions.","A direct test would measure resolved gas metallicities and kinematics in the 42 outliers: if the metal-poor gas is also kinematically misaligned, the external origin is confirmed; if it is co-rotating and enriched, dilution of stellar mass loss is the explanation.","The dilution mechanism predicts that the visibility of an accretion event depends on the host's pre-existing gas mass, so low-metallicity outliers should have systematically lower molecular gas fractions than the rest of the star-forming ETG population at fixed stellar mass."],"forward_implications":["The fraction of early-type galaxies whose residual star formation is externally fuelled is at least 37%, not the few percent that the raw metallicity-outlier count implies.","Longer gas depletion times in ellipticals, as some observations suggest, would push the accreted fraction above 37%.","The estimate agrees with independent kinematic studies reporting external gas in about 42–45% of early-type galaxies, supporting the use of gas-phase metallicity distributions as a population-level probe.","Low-metallicity gas is 4.4 times more common in ellipticals than in spirals because ellipticals have small pre-existing gas reservoirs, so the same accreted mass causes a much larger dilution.","Because the sample requires star formation to dominate the ionisation, accreting systems hidden by old-star or AGN ionisation are missed, so the derived fraction is likely an underestimate rather than an overestimate."],"supporting_citations":[{"why":"Defines the mass–metallicity relation and its scatter used to identify the 7.4% low-metallicity outliers.","marker":"Tremonti et al. (2004)"},{"why":"Provides the emission-line ratio criteria that restrict the sample to galaxies where ionisation is dominated by star formation.","marker":"Kauffmann et al. (2003b)"},{"why":"Supplies the MPA-JHU estimates of gas-phase metallicities, star formation rates, and stellar masses used throughout the analysis.","marker":"Brinchmann et al. (2004)"},{"why":"Galaxy Zoo citizen-scientist classifications that define the elliptical galaxy sample.","marker":"Lintott et al. (2008, 2011)"},{"why":"FIREFLY stellar metallicities used for the gas-to-stellar metallicity comparison.","marker":"Wilkinson et al. (2017)"},{"why":"The ChemEvol chemical evolution model chain used to compute the ~400 Myr enrichment timescale.","marker":"Morgan & Edmunds 2003; Rowlands et al. 2014; De Vis et al. 2017"},{"why":"Provides the 2 Gyr depletion time assumed for t_total when converting the visible fraction to the true fraction.","marker":"Bigiel et al. (2011)"},{"why":"Independent kinematic estimate that about 42% of local early-type galaxies have externally supplied gas, used as a comparison for the 37% result.","marker":"Davis et al. (2011)"},{"why":"Independent kinematic estimate of about 45% external gas in early-type galaxies, also used as a comparison for the 37% result.","marker":"Bryant et al. (2019)"}],"fun_headline_variants":["Accreted gas fuels 37% of star-forming ellipticals","Low-metal gas reveals external fuel in ellipticals","37% of ellipticals get star-forming gas from outside","Gas from accretion, not stellar loss, feeds ellipticals"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The estimate that at least 37% of gas-rich ellipticals accrete their fuel rests on the toy model's ~400 Myr visibility timescale, which the authors describe as dependent on a 'somewhat arbitrary' star-formation history, and on the assumption that the metal-poor gas is truly accreted rather than diluted stellar mass loss.","fun_headline_variants_meta":{"raw":{"variants":["Accreted gas fuels 37% of star-forming ellipticals","Low-metal gas reveals external fuel in ellipticals","37% of ellipticals get star-forming gas from outside","Gas from accretion, not stellar loss, feeds ellipticals"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000354,"raw_usage":{"total_tokens":1905,"prompt_tokens":904,"completion_tokens":1001,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":520,"completion_tokens_details":{"reasoning_tokens":934}},"tokens_in":520,"tokens_out":1001,"duration_ms":10213,"temperature":1.0,"reasoning_tokens":934,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:24:18.914536+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Resolved optical or cold-gas spectroscopy of a handful of the 42 low-metallicity ellipticals could settle the matter: if the gas is kinematically aligned with the stars and its abundance pattern matches enriched stellar mass loss, the external-accretion attribution collapses; if the gas is misaligned or disturbed and uniformly metal-poor, the paper's interpretation stands.","supporting_citations":[],"review_version":1}