{"id":"b49edccc-9d78-480e-a8eb-af1960369c3f","arxiv_id":"2412.06155","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A two-parameter model shows that ultra diffuse galaxies with globular cluster mass fractions near 10% likely formed clusters very efficiently while destroying only a modest fraction of them.","lead":"The paper proposes a simple model to explain why some ultra diffuse galaxies carry unusually large globular cluster systems, concluding that very efficient cluster formation combined with modest destruction is the likely cause. The model links the cluster-to-stellar mass ratio to galaxy stellar populations, and the authors compare it with a sample of 13 UDGs.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"GC destruction fraction d=0.7–0.9 is borrowed from classical dwarf simulations, but UDGs are less dense and the same simulation predicts lower d for them, so the inferred high GC formation efficiency may be overstated.","rationale":"The reader's weakest assumption correctly identifies the destruction fraction d as the most fragile link. My analysis agrees: Section 5 adopts d = 0.7–0.9 from a simulation calibrated to classical dwarfs on the standard stellar mass–halo mass relation, while Section 4 of the same paper uses that simulation's trend of lower disruption in lower-density galaxies to explain UDG GC richness. The tension is direct. The sensitivity of Equation 6 is large: the required c varies by more than a factor of three between d = 0.7 and d = 0.3, so the 'very high efficiency' statement is not robust to the plausible range of d for UDGs. A dedicated simulation run for UDG-like halos would settle the issue. If d is found to be < 0.7, the paper's headline should be softened but not rejected, since even c ≈ 0.2 is still several times higher than the c ≈ 0.03 inferred for classical dwarfs; the qualitative conclusion that GC-rich UDGs required elevated GC formation efficiency would survive. If d is confirmed at 0.7–0.9, the current conclusion stands. The paper is transparent, quotes uncertainties, and acknowledges the simulation caveat, and the JWST high-redshift cluster fractions provide independent support for large c. No critical mathematical errors were found; the model is simple but appropriately so. The conditional verdict is appropriate: the authors should either run the suggested test or explicitly justify why the classical dwarf d range applies to UDGs despite the opposite trend they cite. I therefore recommend no change to the reader's verdict (UNCHANGED).","tokens_in":18046,"tokens_out":13862,"duration_ms":122092,"concrete_test":"Run the Moreno-Hilario et al. (2024) N-body GC disruption model on dwarf halos with UDG-like properties (effective radius > 1.5 kpc, central surface brightness > 24 mag arcsec^-2, cored dark matter density profile, and M_halo/M_* elevated as inferred for GC-rich UDGs), using the same initial power-law cluster mass function. Measure the surviving GC mass fraction after 10–12 Gyr to obtain d. If the resulting d is below 0.7, Equation 6 gives c = 1/((1-d)/y - d) = 0.13–0.22 for y = 0.10, falling below the paper's '≥40%' threshold and undermining the strong version of the claim; if d remains 0.7–0.9, the central claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing premise is the adopted GC destruction fraction d = 0.7–0.9 (Section 5), taken from the Moreno-Hilario et al. (2024) simulation of classical dwarf galaxies. The paper itself states (Section 4) that this simulation finds lower disruption rates for lower-mass, lower-density galaxies, and uses that trend to explain the GC richness of UDGs. Yet it then fixes d = 0.7–0.9 for UDGs, the classical dwarf range, rather than a lower value appropriate to low-surface-density systems. This matters because Equation 6, M_GC/M_* = (1-d)/(1/c+d), is highly sensitive to d near the observed 10% target: for d = 0.7, c ≈ 0.43; for d = 0.5, c ≈ 0.22; for d = 0.3, c ≈ 0.13. If the true d for UDGs is at the lower end of this range, as the cited simulation's own trend suggests, the paper's quantitative inference 'very high GC formation efficiencies (≥40%)' is no longer required; c would be comparable to the cluster mass fractions (30–70%) directly seen in JWST lensed galaxies, and the 'failed galaxy' interpretation would rest on less extreme efficiency. Conversely, if d = 0.9, Equation 6 requires c ≈ 10 (1000%), which is implausible and not supported by any observation. The assumption is therefore genuinely load-bearing and insufficiently justified for the specific UDG population, a point the paper acknowledges only qualitatively (Section 5: 'their model galaxies... follow the standard stellar mass–halo mass relation by design'). The paper is otherwise internally consistent and transparent about this dependence.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper addresses the origin of ultra diffuse galaxies (UDGs) with unusually massive globular cluster (GC) systems. It constructs a simple two-parameter model in which the present-day GC-to-stellar mass ratio, M_GC/M_*, depends on an initial GC formation efficiency c and a destruction fraction d (Eq. 6). The model assumes early quenching and that disrupted GCs contribute their stars to the host galaxy. Using GC destruction fractions d = 0.7–0.9 from the Moreno-Hilario et al. (2024) simulations and formation efficiencies suggested by recent JWST observations of lensed galaxies, the authors conclude that UDGs with M_GC/M_* ≈ 10% require very high formation efficiencies (≥40%) combined with modest destruction. They further compare stellar population properties ([M/H], age, [Mg/Fe]) of a small UDG sample with model tracks and report a loose trend toward more GC-like populations with increasing M_GC/M_*.","tokens_in":18442,"tokens_out":3856,"duration_ms":38351,"significance":"If the central inference holds, the paper offers a compact, physically motivated framework for connecting GC richness to the integrated stellar populations of UDGs, and it usefully highlights the role of GC destruction in shaping present-day ratios. The model is transparent and falsifiable, and the authors are commendably explicit about the limitations of their sample and assumptions. The use of JWST lensed-galaxy cluster mass fractions as empirical anchors for c is a strength. However, the quantitative conclusion is currently weakly constrained because it hinges on a destruction fraction that is borrowed from classical dwarf simulations, and the empirical comparison is affected by small sample size, selection biases, and a model zero point defined from the same data. These issues leave the main claim defensible but not yet firmly established.","major_comments":[{"comment":"The adopted destruction fraction d = 0.7–0.9 is load-bearing for the paper's main quantitative claim, yet it is taken directly from Moreno-Hilario et al. (2024), whose simulated galaxies follow the standard stellar mass–halo mass relation by the authors' own admission. Section 4 states that this same simulation finds lower disruption rates in lower-mass, lower-density galaxies, which is used to explain why UDGs have high M_GC/M_* today. Applying the classical dwarf range d = 0.7–0.9 to UDGs is therefore inconsistent with the cited trend. The sensitivity of Eq. (6) is large: for M_GC/M_* = 10%, d = 0.7 requires c ≈ 0.43, d = 0.5 requires c ≈ 0.22, and d = 0.3 requires c ≈ 0.15. If the true UDG destruction fraction is lower, the claim that 'very high GC formation efficiencies (≥40%)' are required is no longer supported; c would fall in the range already inferred from JWST lensed galaxies. The authors should either justify a UDG-specific d range or present the inferred c as a function of d across a wider interval.","section":"Section 5, Eq. (6)"},{"comment":"The model tracks in Fig. 2 are anchored by defining the M_GC/M_* = 0 stellar population as the average of the five sample UDGs with M_GC/M_* < 1.5%, with the high-ratio endpoint set to the assumed mean stellar population of old, metal-poor GCs. Because the low-ratio zero point comes from the same sample that is then compared with the model, the predicted trend toward more GC-like populations with increasing M_GC/M_* is partly built into the model construction rather than emerging as an independent test. This is acknowledged as a limitation ('This can be improved in the future...'), but it should be stated more prominently, and an external zero point based on classical dwarf stellar populations should be explored to see whether the trend survives.","section":"Section 6, zero point and Fig. 2"},{"comment":"The empirical support for the claimed trends is weak. The sample contains only 12 UDGs, mostly in high-density environments, with heterogeneous GC count and stellar population measurements, and the catalogue itself is biased against GC-poor UDGs (as stated in Section 6). The claimed metallicity decrease of ~0.45 ± 0.1 dex and the weak age trend are not quantified by any fit or rank correlation; inspection of Table 1 shows substantial scatter (e.g., DF17 at M_GC/M_* = 2.1% has [M/H] = –0.83 while DF44 at 4.9% has [M/H] = –1.33, yet PUDG-R84 at 3.9% has [M/H] = –1.48). The [Mg/Fe] panel shows no clear trend. The statement that 'the current data loosely follow the model' would be more convincing with a quantitative significance estimate and an explicit discussion of how selection biases affect the comparison.","section":"Table 1 and Section 6"}],"minor_comments":[{"comment":"For NGVSUDG-20, '11 GC candidates with a large uncertainty of ±8.6' should specify whether this is a Poisson uncertainty, a total uncertainty, or a confidence interval, and the sign convention should be clarified.","section":"Section 2"},{"comment":"The caption would benefit from stating the range of c shown and from marking the specific c values at which the curves intersect the 10% dashed line, since those intersections carry the paper's main quantitative message.","section":"Fig. 1 caption"},{"comment":"The long-dashed blue line representing a constant GC-like stellar population is not defined by an equation or by explicit assumptions in the text; please state how this limiting case is constructed.","section":"Fig. 2 caption"},{"comment":"The top axis showing approximate S_N values is not introduced in the text; define the conversion from M_GC/M_* to S_N and state the assumed mass-to-light ratio and GC mean mass used for this axis.","section":"Section 6, Fig. 2"},{"comment":"For Eridanus II, the statement that M_GC/M_* ≈ 4% 'from its only GC' should clarify that this assumes the universal mean GC mass of 2 × 10^5 M_sun; the dependence on this assumption is worth stating.","section":"Section 7"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of MNRAS and the authors are transparent about many limitations. The central risk is that the quantitative conclusion (very high GC formation efficiencies) rests on a destruction fraction that may not apply to UDGs; if the authors can provide a UDG-specific estimate or a sensitivity analysis over a broader d range, the claim can be placed on firmer ground. The zero-point circularity and small-sample issues should be addressed explicitly, but they are not fatal if framed as illustrative rather than as a definitive test. I see no concerns about citation practice or novelty disclosure."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [Colleague],\n\nThe short version: this paper gives a clean two-parameter accounting for why some UDGs have M_GC/M* up to 10%, and connects that ratio to integrated stellar populations in a way nobody has quite laid out before. The central quantitative claim is not robust as stated, but the model is a useful skeleton and deserves a referee.\n\nWhat's new: Equation 6 is trivial algebra, but the synthesis matters. Combining JWST lensed-galaxy constraints on cluster mass fractions at z~6-10 (30-70%) with a dwarf-galaxy destruction simulation to translate today's M_GC/M* into an initial cluster formation efficiency is genuinely useful. The paper also assembles a small but recent sample of UDGs with both GC counts and spectroscopic stellar populations, and it is transparent about scatter and selection effects. The bias against GC-poor UDGs is stated plainly, which earns trust.\n\nThe math is correct as far as it goes, and the citation pattern is fine: the relevant JWST and simulation papers are cited, and borrowed assumptions are flagged. The soft spot is load-bearing: the destruction fraction d=0.7-0.9 comes from Moreno-Hilario et al. (2024), a simulation of classical dwarfs. The paper itself notes that this simulation finds lower destruction rates for lower-density galaxies -- i.e., the UDGs in question -- and then adopts the classical-dwarf range anyway. If the true d for UDGs is 0.4-0.5, reaching M_GC/M* = 10% requires c ~ 0.2-0.3, not the 40%+ the paper emphasizes. The headline inference is therefore sensitive to the least constrained parameter. There is also a partial circularity: the puffy-dwarf zero point for the stellar population tracks is the average of five UDGs from the same sample, so the metallicity trend is a consistency check, not an independent prediction. The data themselves only support a loose trend; age and [Mg/Fe] show weak or absent correlations.\n\nNet: a useful, honest framework, but the 'most likely' conclusion is overstated. The paper is worth publishing if the authors widen the allowed range of d, or better, justify a UDG-specific value, and soften the causal language. I'd send it to a competent referee rather than desk reject it.","headline":"A clean two-parameter model for GC-rich UDGs, but the adopted destruction fraction is likely too high for UDGs, so the claim of very high formation efficiency is not yet supported.","tokens_in":18957,"tokens_out":3127,"would_cite":true,"duration_ms":28871,"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":"Cluster-rich ultra diffuse galaxies are best explained as failed galaxies whose globular clusters formed with extremely high efficiency — at least 40–80 percent of field-star formation — followed by modest cluster destruction, a…","keywords":["ultra diffuse galaxies","globular cluster systems","globular cluster formation efficiency","globular cluster destruction","failed galaxies","stellar populations","galaxy quenching"],"falsifier":"Measure the integrated stellar metallicity of a robust, spectroscopically confirmed UDG with $M_{\\rm GC}/M_*$ near 10 percent: the model predicts field stars there must be strongly metal-poor and GC-like because disrupted clusters dominate the field by equation 7, so finding such a galaxy with normal dwarf-like metallicity for its stellar mass would falsify the claim that high cluster efficiency drives these systems.","tokens_in":17859,"feed_emoji":"🌌","tokens_out":13480,"duration_ms":110237,"temperature":0.7,"pith_summary":"Some ultra diffuse galaxies (UDGs) carry globular cluster systems whose total mass reaches about 10 percent of the galaxy's stellar mass, far above the roughly 0.5–1 percent typical of classical dwarf galaxies. The paper asks whether that excess comes from forming globular clusters with unusually high efficiency or from destroying fewer of them than usual. A two-parameter model — GC formation efficiency $c$ and destruction fraction $d$ — shows that today's ratios require very high formation efficiency (at least 40–80 percent, depending on $d$) combined with destruction fractions near 70–80 percent, rather than low destruction alone. The accompanying stellar population data loosely follow the model's prediction that field stars become more metal-poor and slightly older as the cluster-mass fraction rises, consistent with disrupted clusters seeding the field. If correct, cluster-rich UDGs are failed galaxies that formed almost entirely through globular clusters and then stopped forming stars early.","feed_headline":"Cluster-rich ultra diffuse galaxies are likely 'failed galaxies'","feed_subtitle":"A two-parameter model says they formed clusters at 40-80% efficiency, then stopped making stars.","key_machinery":"The load-bearing object is a two-parameter identity, equation 6: $M_{\\rm GC}/M_* = (1-d)/(1/c+d)$, derived from the assumptions that the galaxy quenches early, that the only field-star growth is through disrupted globular clusters, and that the initial cluster mass is $c$ times the initial field-star mass. A companion expression, equation 7, gives the fraction of the final stellar mass that came from disrupted clusters, $d/(1/c+d)$, and this is what the paper uses to weight the expected stellar-population shift from dwarf-like to globular-cluster-like as the cluster-mass fraction increases. The machinery's work is to turn the observed spread in $M_{\\rm GC}/M_*$ — from about 0 percent for puffy-dwarf UDGs to about 10 percent for failed-galaxy candidates — into a constraint on the combination of formation and destruction that produced it.","core_discovery":"The central claim is that the extreme globular-cluster richness of some UDGs is set at formation, not by survival. From equation 6, the present-day ratio $M_{\\rm GC}/M_* = (1-d)/(1/c+d)$, where $c$ is the initial ratio of globular-cluster mass to field-star mass and $d$ is the fraction of globular-cluster mass destroyed through mass loss and tidal disruption. Reaching $M_{\\rm GC}/M_* \\sim 10\\%$ today requires $c \\ge 0.4$ if $d = 0.7$, $c \\ge 0.8$ if $d = 0.8$, and $c > 1$ if $d = 0.9$. Since destruction fractions of 0.9 would demand impossibly high formation efficiencies, the authors conclude that cluster-rich UDGs most plausibly formed with very high GC formation efficiencies — consistent with JWST detections of high-redshift lensed galaxies in which bound clusters hold 30–70 percent of the stellar mass — and with only modest subsequent destruction. The model also predicts, and the current data loosely show, that as $M_{\\rm GC}/M_*$ rises the stellar populations of UDGs become more metal-poor and slightly older, approaching the properties of old metal-poor globular clusters.","pith_inferences":["A sharper test of the $c$–$d$ degeneracy would be to measure the fraction of GC-like stars in UDG fields: equation 7 shows that at fixed $M_{\\rm GC}/M_*$, a higher disrupted-cluster fraction means a more metal-poor, older field, so abundance patterns could break the degeneracy that a single ratio leaves open.","If future simulations tailored to low-surface-density, cored-halo UDGs yield destruction fractions below 0.7, the required formation efficiency drops, and cluster-rich UDGs could be ordinary dwarfs with unusually high cluster formation rather than fundamentally failed galaxies; the paper's conclusion would then weaken but its framework would still hold.","The model implies that the 'failed galaxy' label is better viewed as a continuum index — the fraction of stars contributed by disrupted globular clusters — rather than a binary classification; this index is measurable from integrated stellar abundances.","Because the JWST cluster-mass fractions are lower limits (fainter clusters are undetected), deeper imaging of the same lensed galaxies could push $c$ toward or above 1, which would make the failed-galaxy scenario even more extreme and would predict that field stars in these galaxies should be almost entirely GC-like."],"forward_implications":["Cluster-rich UDGs with $M_{\\rm GC}/M_* \\sim 10\\%$ are most likely failed galaxies: they formed in massive halos, produced globular clusters with very high efficiency, and quenched before forming most of their field stars.","The field stars of the most cluster-rich UDGs should resemble old, metal-poor, alpha-enhanced globular-cluster populations; NGC5846_UDG1, with $M_{\\rm GC}/M_* = 9.8\\%$, matches this prediction.","The model predicts a continuous trend of decreasing metallicity and slightly increasing age with rising $M_{\\rm GC}/M_*$, rather than a sharp division between puffy dwarfs and failed galaxies; current data loosely follow this trend.","High-redshift lensed galaxies seen by JWST, with 30–70 percent of their stellar mass in bound clusters, may be the formation phase of today's cluster-rich UDGs.","If $d = 0.9$, no plausible $c$ can reproduce the observed 10 percent upper limit, so the data favor destruction fractions near 0.7–0.8 rather than higher ones."],"supporting_citations":[{"why":"Simulation used to set the destruction fraction range $d = 0.7$–$0.9$ for UDGs.","marker":"Moreno-Hilario et al. (2024)"},{"why":"Supplies the most reliable UDG with $M_{\\rm GC}/M_* \\approx 10\\%$, NGC5846_UDG1, the anchor for the high-ratio end.","marker":"Danieli et al. (2022)"},{"why":"Established that Coma UDGs span $M_{\\rm GC}/M_*$ from 0% to about 10% with an average near 5%.","marker":"Forbes et al. (2020)"},{"why":"Provides the catalogue of UDG stellar populations and GC counts used in the comparison.","marker":"Gannon et al. (2024)"},{"why":"JWST lensed galaxy at $z \\approx 10.2$ where bound clusters hold about 30% of the stellar mass, constraining $c$.","marker":"Adamo et al. (2024)"},{"why":"JWST lensed galaxy at $z \\approx 8.3$ with about 49–57% of the stellar mass in clusters, another high-$c$ constraint.","marker":"Mowla et al. (2024)"},{"why":"Lensed 'Cosmic Grapes' galaxy at $z \\approx 6$ with about 70% of the light in clumps, supporting high cluster formation efficiency.","marker":"Fujimoto et al. (2024)"},{"why":"Introduced the 'failed galaxy' scenario and the 2.5% threshold used to separate puffy dwarfs from failed galaxies.","marker":"Peng & Lim (2016)"}],"fun_headline_variants":["Cluster-rich UDGs are born rich, not survivors","Formation efficiency, not survival, explains failed galaxies","High cluster formation efficiency marks failed galaxies","JWST shows failed galaxies formed clusters at 40-80% efficiency"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The model's conclusions rest on assuming that ultra diffuse galaxies destroy 70–90 percent of their globular clusters, a destruction fraction taken from simulations of ordinary dwarf galaxies, even though the cluster-rich UDGs are thought to have unusually heavy dark matter halos compared with their stellar mass.","fun_headline_variants_meta":{"raw":{"variants":["Cluster-rich UDGs are born rich, not survivors","Formation efficiency, not survival, explains failed galaxies","High cluster formation efficiency marks failed galaxies","JWST shows failed galaxies formed clusters at 40-80% efficiency"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000185,"raw_usage":{"total_tokens":1421,"prompt_tokens":1141,"completion_tokens":280,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":757,"completion_tokens_details":{"reasoning_tokens":215}},"tokens_in":757,"tokens_out":280,"duration_ms":3530,"temperature":1.0,"reasoning_tokens":215,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T19:57:39.697825+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the integrated stellar metallicity of a robust, spectroscopically confirmed UDG with $M_{\\rm GC}/M_*$ near 10 percent: the model predicts field stars there must be strongly metal-poor and GC-like because disrupted clusters dominate the field by equation 7, so finding such a galaxy with normal dwarf-like metallicity for its stellar mass would falsify the claim that high cluster efficiency drives these systems.","supporting_citations":[],"review_version":1}