{"id":"c1167f6c-604a-4df0-b70b-7d29b1e872c8","arxiv_id":"2412.07176","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"CO luminosity and WISE 12 micron luminosity are tightly correlated in early-type galaxies, with steeper slopes and a constant CO(2-1)-to-CO(1-0) ratio.","lead":"Researchers compiled CO and WISE infrared observations for 352 nearby early-type galaxies and found that molecular gas emission correlates tightly with 12 micron light, but with a steeper relation than in star-forming galaxies. Because WISE mapped the entire sky, the result could let astronomers estimate cold gas in quiescent galaxies without expensive radio observations.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Censoring changes the fitted slopes from 1.14/1.19 to 1.24/1.52, so the detection-only r>0.9, scatter<0.1 dex, and constant-R21 claims are not established for the gas-poor ETGs the paper targets.","rationale":"The reader's weakest assumption is exactly where the argument is most vulnerable. The headline numbers come from detections, but the application is to gas-poor galaxies. The Table 1 censored fits are already in the paper; the key additional point is that the same censoring that changes the slopes also undermines the constant-R21 claim, because equal slopes are the evidence for it. The 68-galaxy subset with both lines detected is even more strongly selected: galaxies with low R21 will preferentially drop out of the CO(2-1) detection sample, biasing the comparison. A formal joint censored fit of both lines would show whether the population-level slopes are actually equal. The Appendix A old-star correction is more fragile, since it relies on the non-detections' L12-M* relation and is described as preliminary, but it is not the basis of the headline empirical claim; the steeper slope in ETGs is present even in detection-only fits. Thus the censoring/representativeness issue is the single most load-bearing concern. Because the paper reports the censored fits honestly and the empirical correlation for detected ETGs is still interesting, a conditional verdict remains appropriate; no verdict change is needed.","tokens_in":33578,"tokens_out":6326,"duration_ms":72004,"concrete_test":"Run a joint Bayesian censored regression on the full Table 2 data, treating the 258/262 non-detections as 5-sigma upper limits and fitting both CO lines simultaneously, then compare the equal-slope (constant R21) model against the model with separate slopes using a likelihood-ratio or posterior predictive check. If the slope difference (1.24 vs 1.52) is significant at >95% or the censored-model Spearman r falls below 0.9, the detection-only headline claims are not robust to the gas-poor population. This requires no new data, only the published Table 2 fluxes and upper limits.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claims (r > 0.9, intrinsic scatter < 0.1 dex, and constant R21) are estimated from the 82 CO(1-0) and 76 CO(2-1) detections, yet the paper's stated application is molecular-gas estimation in gas-poor ETGs, a population that is mostly non-detections (352 total, 258/262 upper limits). Table 1 shows this censoring is not negligible: adding upper limits changes the CO(1-0) slope from 1.14 +/- 0.06 to 1.24 +/- 0.07 and the CO(2-1) slope from 1.19 +/- 0.06 to 1.52 +/- 0.10, and reduces Spearman r from 0.91/0.92 to 0.87/0.82. The constant-R21 claim rests on the detection-only slopes being equal (1.14 vs 1.19; 1.16 vs 1.16 for the 68 galaxies with both lines), but the censored fits imply slopes that differ by about 2.3 sigma, which would predict R21 increasing with L12. The paper is transparent in Section 3.2 that non-detections could substantially alter the best-fit parameters, but the abstract presents only the detection-only numbers. The load-bearing assumption is that the detected galaxies are representative of the full ETG population; if the typical gas-poor ETG falls below the detection-only relation, the headline correlation, scatter, and R21 conclusions do not support the claimed practical utility.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper compiles CO(1-0) and CO(2-1) measurements for 352 early-type galaxies (ETGs) from public surveys, matching them with WISE 12 micron luminosities measured in the CO beams. For the 82 CO(1-0) and 76 CO(2-1) detections, it reports tight power-law correlations with Spearman r ~ 0.9 and intrinsic scatter ~ 0.1 dex, slopes that are steeper than those of star-forming galaxies, and consistent slopes for the two CO lines. The authors then analyse residuals from the Gao et al. (2019) star-forming relation and propose either subtracting the 12 micron emission from old stars or adding a constant CO brightness density as ways to reconcile the ETG relation with the star-forming relation.","tokens_in":33969,"tokens_out":5691,"duration_ms":59332,"significance":"If the detection-only correlations apply to the wider ETG population, the paper would provide a practical single-band estimator of molecular gas in gas-poor galaxies and a new constraint on excitation conditions in ETGs. The compiled catalogue in Table 2 and the explicit upper-limit fits are useful assets, and the paper is transparent about the detection fraction and about using LinMix with censored data. However, the headline claims rest on a minority of the sample (82/352 and 76/352 detections), and the censored fits in Table 1 materially change the key slopes and the implied R21 behaviour. The significance of the paper is therefore conditional on reframing the claims and on a more careful treatment of selection and censoring.","major_comments":[{"comment":"The abstract's headline numbers (r > 0.9, intrinsic scatter < 0.1 dex, and the estimator in Eq. 4) are computed from the 82 CO(1-0) and 76 CO(2-1) detections only. Table 1 shows that adding the 258/262 non-detections as 5-sigma upper limits changes the CO(1-0) slope from 1.14 +/- 0.06 to 1.24 +/- 0.07 and the CO(2-1) slope from 1.19 +/- 0.06 to 1.52 +/- 0.10, while Spearman r drops to 0.87/0.82 and the intrinsic scatter for CO(1-0) rises to 0.16 dex. Because the paper's stated application is molecular-gas estimation in gas-poor ETGs, most of which are non-detections, the detection-only relation should not be presented as the population relation without either demonstrating that the detected galaxies are representative or making the censored fit the primary result.","section":"Section 3.1, Table 1"},{"comment":"The constant-R21 conclusion is based on the detection-only slopes (1.16 +/- 0.06 vs 1.16 +/- 0.07 for the 68 galaxies with both lines). The censored fits (1.24 +/- 0.07 vs 1.52 +/- 0.10) differ by roughly 2.8 sigma, which would instead predict R21 increasing with 12 micron luminosity. Section 3.2 acknowledges that non-detections could substantially alter the best-fit parameters, but the abstract and the summary list the constant-R21 result without this caveat. Please either restrict the claim to the detected population or provide a censoring-robust estimate of the R21 behaviour.","section":"Section 3.2, Table 1"},{"comment":"The old-star subtraction and the constant CO brightness densities (2.8 and 4.4 K km/s) are fitted to the same deviations that they are subsequently used to remove. The improvement after correction is therefore a post-hoc fit, not an independent validation. To support the physical interpretation and the claimed practical correction, the authors should test the corrections on an independent sample (e.g., high-resolution CO maps or an out-of-sample ETG set) or clearly label these corrections as illustrative toy models rather than validated recipes.","section":"Section 4, Appendices A and B"},{"comment":"The claim that the ETG CO(2-1) slope is steeper than the star-forming slope is not strongly established by the detection-only numbers: 1.19 +/- 0.06 versus 1.11 from Gao et al. (2019) is only about a 1.3-sigma difference when only the ETG uncertainty is considered. Please propagate the uncertainty of the reference relation, or fit both samples in a common framework, before claiming a steeper slope for CO(2-1).","section":"Section 3.2, Eq. (5)"}],"minor_comments":[{"comment":"The quantity log(Mmol / [K km/s pc^2]) in Eq. (4) is dimensionally inconsistent; the argument of the logarithm should presumably be Mmol/Msun. Please correct the units.","section":"Eq. (4)"},{"comment":"Table 2 is referred to as 'Table B' in Section 2.1; please renumber or refer to it consistently. Also check entries such as the CO(2-1) upper limit for NGC4143, which appears implausibly large and may be a column or units typo.","section":"Section 2.1, Table 2"},{"comment":"There are several grammar slips: 'adding an constant CO brightness density' should be 'adding a constant CO brightness density', and the Figure 2 caption 'arrows mean 5-sigma CO upper limits' should be 'arrows denote 5-sigma CO upper limits'.","section":"Abstract and Figure 2 caption"},{"comment":"The statement that adopting different 2-5 sigma upper limits does not 'significantly alter' the slopes should be quantified; a sentence or supplementary table giving the resulting slopes would allow readers to verify this claim.","section":"Footnote 1"},{"comment":"The sentence 'the 5-sigma CO upper limits for non-detections are close to the best-fitting line for detections, suggesting that these non-detections fall significantly below the fit' is internally confusing; please clarify whether the upper limits lie on, above, or below the detection relation.","section":"Section 3.2"}],"recommendation":"major_revision","confidential_remarks":"The paper is honest about the censoring issue in Section 3.2, so the main problem is a mismatch between the qualified body text and the unqualified abstract/summary. I would encourage a revision that makes the censored analysis a central element and that clearly separates the detection-only empirical relation from the population-level claims. The proposed corrections in Appendices A and B should also be framed as exploratory, not as validated calibrations."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a useful first ETG calibration of the WISE 12 µm–CO scaling relations, but the headline claims are built on the detection-only subsample, and the paper's own Table 1 shows that including the 5σ upper limits changes the story. The central result—steep slopes and constant R21—does not yet apply to the gas-poor ETGs the paper wants to serve.\n\nWhat's genuinely new: they assembled 352 ETGs from ATLAS3D, MASSIVE, and other surveys, homogenized the IRAM conversions, and measured WISE 12 µm in the CO beams. For the 82 CO(1-0) and 76 CO(2-1) detections the correlations are tight (r≈0.9, intrinsic scatter ~0.1 dex), and it's a fair point that these are steeper than the Gao et al. (2019) SFG relations. The paper is also transparent in §3.2 that non-detections could substantially change the fits—that sentence is the most important one in the paper.\n\nThe soft spots are real but not fatal. The censoring issue is load-bearing: adding upper limits changes the CO(2-1) slope from 1.19 to 1.52, and the two slopes then differ by ~2.3σ, which would predict R21 increasing with L12 rather than staying constant. The abstract and summary present only the detection-only numbers. The Appendices A and B corrections (old-star subtraction, constant CO brightness density) are post-hoc—they are fitted to remove the deviation from Gao et al. (2019)—so they read as plausible interpretations, not independent support. Also, the claimed absence of dependence on color, sSFR, morphology is tested only on detections; selection effects could hide a trend. Minor: Table 2 contains a few implausible upper limits (NGC4143, NGC4203, NGC5866 CO(2-1) values) that should be fixed.\n\nWho's this for? Observers who want a first-pass molecular gas estimator for ETGs from all-sky WISE data, and anyone working on censored regression in galaxy scaling relations. It deserves a serious referee: the dataset and question are important, and the paper is honest enough to include the numbers that undermine its own abstract. But it needs major revision—proper handling of censoring (or at least prominent presentation of both fits), tempered claims, and independent validation before it can be used as a calibration.\n\nRecommendation: send to review, with a clear request to address the censoring and reframe the conclusions. I'd engage with it after revision.","headline":"Useful first ETG calibration, but the headline correlation rests on detections only; including upper limits shifts the slopes and undercuts the constant-R21 claim.","tokens_in":34526,"tokens_out":2633,"would_cite":false,"duration_ms":25968,"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":"WISE 12-micron luminosity is a tight proxy for CO emission in early-type galaxies.","keywords":["early-type galaxies","molecular gas","CO luminosity","WISE 12 micron","mid-infrared scaling relations","galaxy quenching","CO-dark gas","line ratio R21"],"falsifier":"Conduct a deep CO survey of the 258 CO(1-0) non-detected galaxies in the sample, sensitive enough to detect them if they lie on the detection-only relation; if the majority are not detected at those predicted levels, or fall on the star-forming galaxy relation instead, the claimed steep slope for the early-type population is refuted. A second independent check is high-resolution matched-beam CO(1-0) and CO(2-1) mapping of gas-rich early-type galaxies to see whether $R_{21}$ is truly constant or varies with local conditions as it does in star-forming galaxies.","tokens_in":33400,"feed_emoji":"🌌","tokens_out":9324,"duration_ms":84358,"temperature":0.7,"pith_summary":"The paper asks whether a single mid-infrared band, the 12 micron channel of the all-sky WISE survey, can serve as a practical proxy for the molecular gas content of early-type galaxies, where CO observations are scarce and often yield non-detections. Using 352 nearby early-type galaxies assembled mainly from volume-limited surveys, it finds that for the 82 CO(1-0) and 76 CO(2-1) detections, CO luminosity and 12 micron luminosity are tightly correlated, with correlation coefficients above 0.9 and intrinsic scatter near 0.1 dex. The slopes are steeper than the corresponding relations in star-forming galaxies, and the CO(1-0) and CO(2-1) slopes are identical, indicating that the line ratio $R_{21}$ does not depend on mid-infrared luminosity in these galaxies. The paper shows that the difference from star-forming galaxies can be largely removed either by subtracting estimated 12 micron emission from old stellar populations or by adding a constant CO brightness density, and offers these corrected relations as estimators of molecular gas in gas-poor galaxies.","feed_headline":"WISE 12-micron light traces CO in early-type galaxies","feed_subtitle":"One all-sky infrared band can estimate molecular gas where CO surveys are sparse and mostly empty.","key_machinery":"The central object is the $L_{\\rm CO}$--$L_{12\\,\\mu\\mathrm{m}}$ scaling relation, a power law in log-log space fitted with a Bayesian linear regression that accounts for measurement errors on both axes. The paper's diagnostic is the comparison of slopes between CO(1-0) and CO(2-1); identical slopes imply a constant line ratio $R_{21}$. Two correction mechanisms are used to interpret the steeper slope: subtracting the 12 micron emission expected from old stellar populations (following an established method) and adding a constant CO brightness density intended to represent CO-dark or diffuse molecular gas. The latter is presented as the average additional brightness, $2.8$ and $4.4\\ \\mathrm{K\\ km\\ s^{-1}}$ for CO(1-0) and CO(2-1), required to remove the correlation between the deviation and the molecular gas surface density.","core_discovery":"On its own terms, the paper establishes that the empirical power-law relations between CO line luminosity and WISE 12 micron luminosity known for star-forming galaxies also hold in early-type galaxies, but with distinct parameters. For CO(1-0), the best fit is $\\log L_{\\rm CO(1-0)} = (1.14\\pm0.06)\\log L_{12\\,\\mu\\mathrm{m}} - (1.59\\pm0.50)$, with intrinsic scatter 0.09 dex and Spearman $r=0.91$; for CO(2-1), $\\log L_{\\rm CO(2-1)} = (1.19\\pm0.06)\\log L_{12\\,\\mu\\mathrm{m}} - (2.00\\pm0.49)$, with scatter 0.10 dex and $r=0.92$. Purely on the 68 galaxies detected in both lines, the two slopes agree at $1.16\\pm0.06$ and $1.16\\pm0.07$, from which the paper concludes that $R_{21}$ is independent of mid-infrared luminosity in early-type galaxies, in contrast to star-forming galaxies where $R_{21}$ rises with mid-infrared luminosity. The deviations of individual early-type galaxies from the star-forming relations do not correlate with colour, morphology, or specific star formation rate, but do correlate with molecular gas surface density; this dependence is eliminated either by subtracting the old-star contribution to the 12 micron flux or by adding a constant CO brightness density ($2.8$ and $4.4\\ \\mathrm{K\\ km\\ s^{-1}}$ for the two lines). These two corrections are preliminary but both bring the early-type relations into agreement with the star-forming ones, and the paper argues that the corrected relations can be used to estimate molecular gas masses in gas-poor systems with small scatter.","pith_inferences":["One editorial extension: the paper's headline slopes come from detections only; Table 1 shows that adding the 258 CO(1-0) and 262 CO(2-1) non-detections as upper limits changes the slopes to 1.24 and 1.52, so the population-level relation for all early-type galaxies remains less settled than the detection-based claim.","If the constant CO-dark brightness floor is real, deep CO mapping of the optically faint outskirts of early-type galaxies should reveal a widespread, low-level CO component of roughly 2-4 K km/s, which would be directly testable.","If constant $R_{21}$ holds in resolved data, it would imply that CO excitation in early-type galaxies is governed by a mechanism other than local star formation surface density, such as cosmic-ray heating or turbulent pressure.","The full-sky coverage of WISE means that, if these relations generalise, molecular gas estimates could be produced for many thousands of early-type galaxies across current optical and near-infrared surveys, enabling statistical studies of quenching without new CO campaigns."],"forward_implications":["If the relation holds beyond the detections, a single WISE 12 micron measurement gives molecular gas masses for early-type galaxies to within roughly 0.1 dex of intrinsic scatter, with no need for expensive CO observations.","The constant $R_{21}$ means CO(2-1) can be used interchangeably with CO(1-0) as a gas tracer in these galaxies, removing excitation uncertainty from the conversion.","The steeper slope implies that at a given 12 micron luminosity an early-type galaxy contains less CO-bright gas than a star-forming galaxy, so applying the star-forming calibration would systematically overestimate molecular gas content in gas-poor systems.","Because the deviations show no dependence on colour, morphology, or specific star formation rate, the correction can be applied without knowledge of other galaxy properties.","Both proposed corrections (old-star subtraction and constant CO-dark gas brightness) recover agreement with the star-forming relations, leaving open which physical process dominates."],"supporting_citations":[{"why":"Supplies the ATLAS3D CO(1-0) and CO(2-1) fluxes for most of the sample, including the non-detections treated as upper limits.","marker":"Young et al. 2011"},{"why":"Provides the CLoGS CO measurements for group-dominant early-type galaxies, a major source of detections.","marker":"O'Sullivan et al. 2018"},{"why":"Contributes the SAURON early-type galaxy CO luminosities used in the sample.","marker":"Combes et al. 2007"},{"why":"Defines the star-forming galaxy $L_{\\rm CO}$--$L_{12\\,\\mu\\mathrm{m}}$ relations that the new relations are compared against and from which deviations are computed.","marker":"Gao et al. 2019"},{"why":"Establishes the method for subtracting old-star 12 micron emission used in the first correction test.","marker":"Davis et al. 2014"},{"why":"Supplies the CO luminosity formula and the conversion factor $\\alpha_{\\rm CO}$ used to derive molecular gas masses.","marker":"Bolatto et al. 2013"},{"why":"Provides the Bayesian linear regression (LinMix) used for the correlation fits with errors on both variables.","marker":"Kelly 2007"},{"why":"Gives the method for measuring WISE 12 micron luminosities within the CO beams rather than for the whole galaxy.","marker":"Chown et al. 2021"},{"why":"Large-scale benchmark for the CO-mid-IR relation and for the $R_{21}$ dependence on star formation surface density in star-forming galaxies.","marker":"Leroy et al. 2023a"},{"why":"Provides the MASSIVE survey CO data for the most massive early-type galaxies in the sample.","marker":"Davis et al. 2019"}],"fun_headline_variants":["WISE 12 micron traces CO in early-type galaxies","Tight CO-WISE 12 micron link in early-type galaxies","Steeper CO-12 micron slopes in early-type galaxies","WISE 12 micron as CO proxy in early-type galaxies"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The headline correlations use only the ~80 galaxies with detected CO, and if the typical gas-poor early-type galaxy lies far below that detection-only relation, the steeper slope and constant $R_{21}$ would overstate the population behaviour.","fun_headline_variants_meta":{"raw":{"variants":["WISE 12 micron traces CO in early-type galaxies","Tight CO-WISE 12 micron link in early-type galaxies","Steeper CO-12 micron slopes in early-type galaxies","WISE 12 micron as CO proxy in early-type galaxies"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001269,"raw_usage":{"total_tokens":5369,"prompt_tokens":1296,"completion_tokens":4073,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":912,"completion_tokens_details":{"reasoning_tokens":4003}},"tokens_in":912,"tokens_out":4073,"duration_ms":30597,"temperature":1.0,"reasoning_tokens":4003,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T19:04:19.273416+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Conduct a deep CO survey of the 258 CO(1-0) non-detected galaxies in the sample, sensitive enough to detect them if they lie on the detection-only relation; if the majority are not detected at those predicted levels, or fall on the star-forming galaxy relation instead, the claimed steep slope for the early-type population is refuted. A second independent check is high-resolution matched-beam CO(1-0) and CO(2-1) mapping of gas-rich early-type galaxies to see whether $R_{21}$ is truly constant or varies with local conditions as it does in star-forming galaxies.","supporting_citations":[{"cited_title":"2019, ApJ, 887, 172","cited_arxiv_id":null,"evidence_quote":"Defines the star-forming galaxy $L_{\\rm CO}$--$L_{12\\,\\mu\\mathrm{m}}$ relations that the new relations are compared against and from which deviations are computed."},{"cited_title":"A., Young, L","cited_arxiv_id":null,"evidence_quote":"Establishes the method for subtracting old-star 12 micron emission used in the first correction test."},{"cited_title":"A., Greene, J","cited_arxiv_id":null,"evidence_quote":"Provides the MASSIVE survey CO data for the most massive early-type galaxies in the sample."}],"review_version":1}