{"id":"c079ad90-d0e4-43ff-a7e6-28373b777c99","arxiv_id":"1908.10374","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Most massive starless clump candidates are not starless: ALMA reveals hidden low-mass protostars and thermal-Jeans-spaced fragmentation in 11 of 12 clumps.","lead":"A high-resolution ALMA survey of 12 massive 'starless' clump candidates finds that 11 of 12 actually contain young, low-luminosity protostars, along with fragmented substructure. The measured spacing of the fragments matches the expected thermal Jeans length, supporting a picture where star clusters begin by thermal gravitational fragmentation.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified","rationale":"The reader identified the Monte Carlo projection prior as the weakest assumption, and I agree that it is the least secure element of the analysis. But I do not find it load-bearing for the paper's central claim. The paper itself provides the uncorrected separations and an alternative spherical correction; both place the median within the stated 0.4-1.6 thermal Jeans range. The claimed median shifts from 0.82 to about 1.1 under the alternative, which does not change the qualitative conclusion that fragmentation occurs near the thermal Jeans length. The 11/12 protostar detection claim is supported by unambiguous CO/SiO bipolar outflows and is independent of the projection correction. The radiative-transfer modeling comparison is appropriately conservative and explicitly lists its limitations. The field-of-view incompleteness beyond delta/lambda ~ 3.1 is acknowledged and affects only the far tail. Therefore the ACCEPT verdict stands without change, though a sensitivity test on the deprojection prior would strengthen the presentation.","tokens_in":41086,"tokens_out":11117,"duration_ms":129407,"concrete_test":"Recompute the ensemble CDF of delta_nns/lambda_j,th from Table 5 under two alternative deprojection priors: (i) no deprojection at all, and (ii) the spherical correction of Eq. 3 with R_s = 0.6 pc. If the median remains between 0.4 and 1.6 and the fraction below one Jeans length stays near 50-65%, the thermal-Jeans conclusion is robust to the projection prior.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I reviewed the central claim that projection-corrected nearest-neighbor separations between the 67 detected sub-structures equal the clump-average thermal Jeans length. The most uncertain step is the Monte Carlo deprojection prior in Section 5.1: isolated sources are assigned line-of-sight offsets from a Gaussian with sigma_z = 0.15 pc, and dendrogram-grouped sources are modeled as filaments with a common inclination capped at a maximum deprojected length D = 0.6 pc. The authors explicitly flag this scheme as 'simple and imperfect.' However, this assumption is not load-bearing: the uncorrected projected median is 0.083 pc = 0.61 lambda_j,th, and the paper's own alternative uniform-sphere correction (Eq. 3 with R_s = 0.38 pc) gives 0.153 pc = 1.13 lambda_j,th. Both fall inside the claimed 0.4-1.6 range and still support 'consistent with the thermal Jeans length.' The large-separation tail is incomplete beyond delta/lambda ~ 3.1 because of the 40-inch field of view (Section 5.2), but this truncates the tail, not the region around the median. I found no internally inconsistent or unsupported step that would overturn the conclusion.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"Svoboda et al. present ALMA 1.3 mm continuum and spectral line observations of 12 high-mass starless clump candidates (SCCs) selected from BGPS and IRDC surveys. They identify 67 dendrogram sub-structures, detect CO/SiO bipolar outflows toward 11 of 12 clumps, and compare selected continuum sources with RADMC-3D starless core models. The central fragmentation analysis measures nearest neighbor separations among the sub-structures, applies a Monte Carlo deprojection with explicit assumptions for isolated and grouped sources, and compares the results to the clump-average thermal Jeans length computed from ATLASGAL 870 μm flux densities with propagated uncertainties. The ensemble median projection-corrected separation is δ_nns/λ_j,th = 0.82 (25–75% percentile interval 0.52–1.25), and the CDF shows 63.3% of separations below one thermal Jeans length, supporting fragmentation at the thermal Jeans scale.","tokens_in":41352,"tokens_out":9020,"duration_ms":87671,"significance":"This is one of the first systematic ALMA surveys of high-mass starless clump candidates at ~3000 au resolution. If the result holds, it provides a robust observational constraint that fragmentation at the clump scale occurs near the thermal Jeans length rather than the turbulent or cylindrical Jeans scale, with implications for the initial conditions of high-mass star and cluster formation. The paper is careful in several respects: point-source completeness and model recovery are quantified, the Monte Carlo propagation of uncertainties in distances, dust temperatures, and densities is well described, the limitations of single-wavelength temperature assumptions and of the starless core model grid are stated explicitly, and the authors refrain from overinterpreting unresolved cores as definitively protostellar. The discovery of low-luminosity protostars in 11 of 12 supposed starless clumps is a significant empirical result that cautions against relying on MIR/FIR survey completeness for identifying truly quiescent clumps.","major_comments":[],"minor_comments":[{"comment":"The paper first reports 67 sub-structures in §2.4, but §5.1 gives 17 isolated and 49 grouped sources, which sum to 66. Since G23605 is excluded from the nearest-neighbor analysis, please state explicitly that 66 of the 67 sources are used, or correct the counts, so the ensemble statistics in Figure 13 can be verified.","section":"§2.4 and §5.1"},{"comment":"The sentence 'Replacing cs with cs,e_ff in Equation 5.2' appears to refer to Eq. (4) for the thermal Jeans length, not to a numbered Eq. (5.2); similarly, §4.4 refers to 'Equation 4.4' when comparing to Eq. (2). Please correct the equation cross-references.","section":"§5.2 and §4.4"},{"comment":"The deprojection scheme is explicitly labeled 'simple and imperfect'; because the numerical medians in Figure 13 depend on the assumed σ_z and D, a brief quantitative robustness check (e.g., varying σ_z and D over plausible ranges, in addition to the already-given uniform-sphere comparison) would help readers assess how sensitive the median δ_nns/λ_j,th is to the prior.","section":"§5.1"},{"comment":"In the conclusions, the statement 'sensitivity ( 50 µJy beam−1) sufficient...' contains a stray opening parenthesis and does not specify that this is the 1σ RMS noise level; please rephrase, e.g., 'with 1σ RMS sensitivity of ~50 µJy beam−1'.","section":"§7"},{"comment":"The phrase 'This ultimately provides a more conservative criteria for rejecting poor fits' is ambiguous: if the Bayesian likelihood is too stringent, the ∆χ²_r < 3 heuristic is presumably less stringent (more permissive). Please reword to avoid confusion.","section":"§4.4"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is a sound and useful contribution. The requested changes are limited to presentation and cross-checking of numbers; I see no concerns about novelty or scope. The count discrepancy and equation cross-reference typos should be fixed before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper is the first targeted ALMA survey of high-mass starless clump candidates, and it delivers two genuinely new results: 11 of 12 clumps show bipolar CO or SiO outflows (i.e., low-luminosity protostars missed by MIR/FIR surveys), and nearest-neighbor separations of the 67 detected substructures land at about the thermal Jeans length (median 0.82 lambda_j,th after projection correction).\n\nThe data work is careful. Reduction is cross-checked against 8-micron extinction, completeness is quantified with radiative transfer models, and Monte Carlo error propagation for the Jeans length is transparent. The paper does not oversell: it flags the single-wavelength temperature assumptions, the incompleteness of the model grid for unresolved cores, and the fact that G28539 remains a possible true starless clump.\n\nThe weakest link is the deprojection in Section 5.1. The priors (sigma_z = 0.15 pc for isolated sources, D = 0.6 pc for filament groups) are admittedly simple. But they are not load-bearing: the uncorrected projected median is 0.61 lambda_j,th, and the uniform-sphere alternative gives 1.13 lambda_j,th. Both fall in the claimed 0.4-1.6 range, so the central conclusion is robust to the correction choice. The tail beyond ~3 lambda_j,th is incomplete due to the primary beam field of view, but that truncates the tail, not the region around the median. The interpretation of unresolved sources as protostellar from poor fits to the starless model grid is, as the authors say, not strictly required - extreme density models could fit - but the free-fall argument plus the outflow detections make it reasonable.\n\nThe sample is small and selected to be the most massive SCCs within 5 kpc, so population claims are made cautiously. That is appropriate.\n\nWho is this for? Anyone working on high-mass star formation initial conditions, fragmentation, or ALMA survey methods. It deserves serious refereeing: the data are public, the analysis is reproducible in principle, and the outflow discovery alone is a solid observational contribution that will drive follow-up. I would send it out.","headline":"A careful ALMA survey showing that supposedly quiescent massive clumps mostly already host low-mass protostars, with fragmentation at the thermal Jeans length; the deprojection assumption is soft but not load-bearing.","tokens_in":41835,"tokens_out":1571,"would_cite":true,"duration_ms":16252,"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":"ALMA observations of twelve high-mass starless clump candidates show that fragmentation occurs at the thermal Jeans length and that 11 of 12 such clumps already contain low-luminosity protostars.","keywords":["high-mass star formation","starless clump candidates","thermal Jeans length","fragmentation","molecular outflows","protostars","ALMA","infrared dark clouds"],"falsifier":"Measure line-of-sight distances directly, for example by fitting dense-gas tracers such as NH3 or N2H+ emission to obtain velocities and solve for the three-dimensional distribution of the sub-structures in a few clumps. If the real line-of-sight spread is much larger than the assumed 0.15 pc Gaussian, or the filament inclinations differ from the Monte Carlo priors, the median $\\delta_{\\mathrm{nns}}/\\lambda_{\\mathrm{j,th}}$ will shift away from 0.82, and the ensemble may instead match the turbulent or cylindrical Jeans scale of roughly 3 $\\lambda_{\\mathrm{j,th}}$; that direct comparison would settle the claim.","tokens_in":40964,"feed_emoji":"🔭","tokens_out":9797,"duration_ms":95988,"temperature":0.7,"pith_summary":"This paper presents a high-resolution millimeter survey of twelve massive clumps that previous surveys flagged as starless, and asks whether such clumps really are quiescent and how they fragment. After correcting the projected separations of the 67 detected sub-structures for geometry, the authors find a median nearest-neighbor spacing of $0.82\\,\\lambda_{\\mathrm{j,th}}$ with 63.3% of separations below one thermal Jeans length. They conclude that the dense gas in these clumps fragments on the thermal Jeans scale, with turbulence and magnetic fields providing only modest support. At the same time, CO and SiO outflows reveal previously undetected low-luminosity protostars in 11 of the 12 clumps, indicating that current infrared surveys miss substantial early star formation. If correct, the results narrow the window in which truly starless massive clumps exist and put the initial fragment spacing of high-mass cluster formation at the thermal Jeans scale.","feed_headline":"Starless clumps fragment at the Jeans scale","feed_subtitle":"ALMA finds 11 of 12 'starless' candidates hide low-luminosity protostars, reshaping how massive stars begin","key_machinery":"The load-bearing quantity is the thermal Jeans length $\\lambda_{\\mathrm{j,th}} = (\\pi c_s^2 / G \\rho_0)^{1/2}$, the minimum wavelength for gravitational fragmentation of a uniform isothermal medium; for these clumps it ranges from about 0.10 to 0.17 pc. The measurement chain is a dendrogram segmentation of the ALMA continuum maps that extracts 67 sub-structures, followed by Monte Carlo trials that assign plausible line-of-sight offsets to isolated sources and filament-like inclinations to branch-linked sources, yielding projection-corrected nearest-neighbor separation distributions that are then scaled by each clump's thermal Jeans length. This construction translates a two-dimensional image of clustered cores into a statistical statement about the intrinsic three-dimensional fragmentation scale.","core_discovery":"The paper's central claim is that fragmentation in high-mass starless clump candidates is controlled by the thermal Jeans length, not by turbulent or magnetic support. In the ALMA 1.3 mm maps, 67 dense sub-structures are extracted with a dendrogram; the nearest-neighbor separations, after Monte Carlo de-projection, have an ensemble median $\\delta_{\\mathrm{nns}}/\\lambda_{\\mathrm{j,th}} = 0.82$ with a (25,75) percentile interval of $0.52$–$1.25$, and 63.3% of separations fall below one thermal Jeans length. The same data show that 11 of the 12 clumps host embedded low- or intermediate-luminosity protostars, traced by 16 bipolar CO outflows and 4 SiO outflows, and that unresolved continuum sources are poorly fit by starless-core radiative-transfer models. Two resolved cores in G28539 are well fit as starless cores, one with a model mass near $29\\,M_\\odot$, leaving that clump as the only true starless candidate in the sample. The authors interpret the ensemble of results as evidence for hierarchical fragmentation in which the highest-density gas fragments thermally before high-mass protostars form.","pith_inferences":["If the assumed projection geometry is close to right, a testable extension is that dense-gas kinematics (e.g., from N2H+ or NH3) should show velocity structure coherent on the Jeans scale, allowing the Monte Carlo correction to be replaced by direct line-of-sight measurements.","The frequent sub-Jeans separations could mean some pairs are still moving together during collapse or that projection correction underestimates clustered sources; comparing relative velocities of close pairs would distinguish contraction and merging from a static fragmented hierarchy.","A similar survey of lower-mass or more distant 70-µm-dark clumps would test whether thermal-Jeans fragmentation is universal among massive starless clumps or only characteristic of the most massive, high-column-density subset studied here.","If the outflow detection rate holds, a quantitative prediction is that roughly 90% of nominally starless massive clumps observed at this resolution and sensitivity will show low-luminosity protostars, shrinking the estimated population of genuinely quiescent massive clumps."],"forward_implications":["If fragmentation is thermal, the initial spacing of massive protocluster cores is set by temperature and density alone, and the thermal Jeans mass of about $1\\,M_\\odot$ aligns naturally with the peak of the stellar initial mass function.","Catalog-level 70 µm 'starless' classifications are not reliable: 11 of 12 such clumps show outflow or compact-source evidence of protostars, so any quiescence claim needs high-resolution millimeter follow-up.","A 'low-mass first' formation route becomes plausible: initially low- to intermediate-mass protostars accrete from the surrounding clump, so high-mass stars need not begin as massive monolithic starless cores.","The bimodal separation distribution and second-nearest-neighbor spacings near the Jeans length support hierarchical fragmentation, with clump-scale cylindrical fragmentation followed by core-scale thermal Jeans fragmentation.","G28539, the sole remaining starless candidate, with its two well-fit high-mass starless core candidates, is the best current target for studying the pre-stellar initial conditions of a massive protocluster."],"supporting_citations":[{"why":"Defines the parent starless clump candidate catalog and the 70 µm luminosity completeness limit that the outflow detections beat.","marker":"Svoboda et al. (2016)"},{"why":"Supplies the additional 70-µm-dark clump candidates used to build the 12-source sample.","marker":"Traficante et al. (2015)"},{"why":"Introduces the dendrogram algorithm used to identify the 67 sub-structures and measure separations.","marker":"Rosolowsky et al. (2008)"},{"why":"Provides the thermal Jeans length formalism and the clump-scale physical context for fragmentation.","marker":"McKee & Ostriker (2007)"},{"why":"Sets the 1.3 mm dust opacity used for masses, densities, and Jeans-length scaling.","marker":"Ossenkopf & Henning (1994)"},{"why":"Supplies the compact Class 0 envelope flux densities and sizes used to identify unresolved ALMA sources as protostar analogs.","marker":"Enoch et al. (2011)"},{"why":"Defines the externally heated Plummer starless-core radiative transfer modeling approach applied to the resolved cores.","marker":"Shirley et al. (2005)"},{"why":"Simulation showing low-mass cores form first and accrete clump gas, the coeval-formation scenario the data are compared with.","marker":"Smith et al. (2009)"},{"why":"Supplies the Gould's Belt protostar luminosity distribution used to argue surveys are incomplete below about 50 solar luminosities.","marker":"Dunham et al. (2014)"}],"fun_headline_variants":["ALMA finds hidden protostars in 'starless' clumps","Starless clumps actually hide low-luminosity protostars","Fragmentation in massive clumps follows Jeans length","ALMA reveals thermal Jeans scaling in starless clumps","Surprise: Most 'starless' clumps host protostars"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire fragmentation-scale conclusion depends on the assumed three-dimensional layout of the sources: isolated sources are modeled with a line-of-sight Gaussian spread of 0.15 pc and dendrogram-linked sources are modeled as filaments with a common inclination capped at a 0.6 pc deprojected length, so if the true geometry differs, the deprojected separations that match the Jeans length could be significantly biased.","fun_headline_variants_meta":{"raw":{"variants":["ALMA finds hidden protostars in 'starless' clumps","Starless clumps actually hide low-luminosity protostars","Fragmentation in massive clumps follows Jeans length","ALMA reveals thermal Jeans scaling in starless clumps","Surprise: Most 'starless' clumps host protostars"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00016,"raw_usage":{"total_tokens":1362,"prompt_tokens":1207,"completion_tokens":155,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":823,"completion_tokens_details":{"reasoning_tokens":65}},"tokens_in":823,"tokens_out":155,"duration_ms":2196,"temperature":1.0,"reasoning_tokens":65,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:45:32.093326+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure line-of-sight distances directly, for example by fitting dense-gas tracers such as NH3 or N2H+ emission to obtain velocities and solve for the three-dimensional distribution of the sub-structures in a few clumps. If the real line-of-sight spread is much larger than the assumed 0.15 pc Gaussian, or the filament inclinations differ from the Monte Carlo priors, the median $\\delta_{\\mathrm{nns}}/\\lambda_{\\mathrm{j,th}}$ will shift away from 0.82, and the ensemble may instead match the turbulent or cylindrical Jeans scale of roughly 3 $\\lambda_{\\mathrm{j,th}}$; that direct comparison would settle the claim.","supporting_citations":[],"review_version":1}