{"id":"3bb92e32-3e40-4d25-89fe-9a854d20a6d2","arxiv_id":"2502.01868","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Resolved ALMA and JWST imaging of nine z=3.09 proto-cluster dusty star-forming galaxies shows predominantly disk-like dust and stellar morphologies with compact cores, interpreted as the early formation of the local morphology-density relation.","lead":"New ALMA and JWST images of nine huge, dusty galaxies in a crowded early-universe region show them as disk-like systems with bright dusty cores, not as the rounded, settled galaxies that dominate modern galaxy clusters. The finding suggests we are watching the first stages of the morphology-density relation, the rule that dense environments today are filled with elliptical galaxies.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 870um Sersic n~1 result may be an artifact of iterative 5-sigma residual masking; without a masking-bias check, the disk-dominance claim is not secure.","rationale":"The reader's weakest assumption correctly identifies the iterative 5-sigma masking of bright residual pixels as the load-bearing step: the abstract's 'masking bright regions' qualifier is exactly where the n~1 result comes from, and the unmasked fits give a significantly higher mean index. My independent reading of Section 3.3 and Table 3 confirms that the disk interpretation is not independently validated by a two-component fit, and the agreement in axis ratio and position angle between ALMA and F444W, while suggestive, could also be produced if both tracers are dominated by the same central structure. The concern is concrete and testable with a simulation that applies the pipeline to known input profiles. The paper otherwise has internal consistency, sensible caveats, and the NIRCam results provide partial independent support for disk-like stellar components in several sources, so the appropriate verdict remains CONDITIONAL: the central measurement needs a masking-bias check before the morphological and environmental conclusions are accepted. I agree with the reader's assessment and recommend no change to the verdict.","tokens_in":25254,"tokens_out":4804,"duration_ms":54954,"concrete_test":"Run an end-to-end recovery test of the exact Section 3.3 pipeline: simulate 870um images from Sersic profiles with n=1, 2, and 4 (with and without a compact core component), convolve with the 0.15'' synthesized beam, add Gaussian noise at the 28 uJy/beam rms, and apply the identical iterative 5-sigma residual-masking fit. If the recovered masked n is within ~0.2 of unity for intrinsically n=4 inputs (or for point-source-plus-wing models), then the reported n=0.86 is a masking artifact and the disk claim fails; if the recovered n tracks the input extended-component n, the concern is resolved. A complementary check is to fit each unmasked 870um image with an explicit two-component model (disk Sersic + unresolved core) and compare Bayesian information criteria against the single-component masked fit.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that ALMA 870um emission is disk-dominated rests on the masked Sersic fits in Section 3.3/Table 3: iterative masking of residual pixels above 5 sigma lowers the mean Sersic index from 1.37 +/- 0.08 (unmasked) to 0.86 +/- 0.05. The mask is applied to characterize a smooth underlying disk in the presence of clumpy structures, but this creates a selection: the fit is performed on the low-surface-brightness outskirts after the bright core and clumps are removed. At 0.15'' resolution with rms 28 uJy/beam, the surviving pixels may not independently constrain the profile shape; a compact high-n Sersic profile whose central peak is masked has shallow outer wings that an exponential can approximate, so n~1 could be produced by the masking procedure rather than by a physical disk. The unmasked n>1 result is then interpreted as an additional 'core component', but the two-component interpretation is not tested with an explicit two-component model. The consistency of axis ratios and position angles between ALMA and F444W (Fig. 8) is supportive but not conclusive, since the same central structure could dominate both. If the masked n~1 is an artifact, the claim that 'dust emission arises from disks' and hence the morphological comparison with local clusters loses its primary quantitative support.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes ALMA 870 um and 1.1 mm images together with JWST/NIRCam (F115W–F444W) images of nine spectroscopically confirmed dusty star-forming galaxies (DSFGs) at z ~ 3.09 in the SSA22 proto-cluster core. Fluxes are measured at multiple resolutions and cross-checked, and Sersic profiles are fit to the ALMA 870 um images both with and without iterative bright-pixel masking, as well as to the NIRCam F444W images. The masked ALMA fits give Sersic indices near unity, which the authors interpret as evidence that dust emission arises from disks with superimposed compact cores; the F444W fits give n ~ 1–3.5, suggesting bulges and disks. Combining dust mass surface density, F444W Sersic index, and F200W–F444W colors, the paper argues that the proto-cluster DSFGs span different evolutionary stages, including red, dusty cores undergoing rapid bulge growth. The paper concludes that these DSFGs are predominantly disk-dominated and interprets this as the early formation stage of the local morphology-density relation.","tokens_in":30,"tokens_out":5643,"duration_ms":98599,"significance":"If the central disk-dominance claim holds, the paper provides a valuable observational datum for the evolution of the morphology-density relation and for the internal structure of DSFGs in a proto-cluster environment. Strengths include the multi-resolution flux cross-checks in Section 3.2, the convolution of model profiles with the synthesized beam in the Sersic fits (Section 3.3), the explicit comparison of masked and unmasked fits, the fixed-n test in Appendix A, and the candid discussion of caveats such as flux recovery, F444W dust extinction, and the small sample size. The comparison of ALMA and F444W sizes, axis ratios, and position angles is a useful addition to the literature, and the paper's emphasis on the coexistence of disks, cores, bars, and clumps is timely.","major_comments":[{"comment":"The central disk-dominance claim rests on the masked Sersic fits. Iterative masking of residual pixels above 5σ removes the bright central core and clumps before fitting, and the reported mean n drops from 1.37 ± 0.08 (unmasked) to 0.86 ± 0.05 (masked). The interpretation that this exposes an underlying exponential disk is not uniquely supported by the data: if the true profile is a compact high-n Sersic with a masked center, the remaining outer wings can be approximated by n ≈ 1. The paper does not provide a recovery simulation showing that the masking-and-fitting procedure returns the input n for single-component Sersic profiles with n spanning 1–4 in the presence of the 0.15″ beam and the measured noise, nor does it fit an explicit two-component (disk + core) model. The fixed-n = 1 test in Appendix A is not an adequate substitute because it forces n = 1 rather than testing whether the residual wings of a high-n profile masquerade as an exponential disk. I recommend adding injection-recovery simulations or explicit two-component fits before the abstract's claim that 'dust emission arises from disks' is accepted.","section":"§3.3, Table 3"},{"comment":"The statement in §3.3 that the fixed-n = 1 unmasked fits 'broadly yielded similar parameters compared with the fit with bright emission masks' is not fully supported by the numbers. For ADF22.A1, Re changes from 4.73 ± 0.53 kpc (masked free-n) to 2.82 ± 0.31 kpc (fixed-n unmasked), and ADF22.A7 changes from 1.53 ± 0.15 to 1.30 ± 0.15 kpc. These differences are several times the quoted uncertainties and indicate that the fixed-n fit is not simply recovering the same disk component. Since the fixed-n fit is the main check on the masking procedure, this discrepancy should be discussed and ideally reconciled with a two-component model.","section":"Appendix A, Table A1"},{"comment":"The concluding claim that 'we are witnessing the early formation of the morphology-density relation' goes beyond what the data directly show. The paper demonstrates that nine DSFGs in the SSA22 proto-cluster core have disk-dominated dust and stellar profiles, and it compares their Sersic indices, sizes, and axis ratios with field DSFGs from Hodge et al. (2025). However, the comparison is not quantitative in terms of environment: the field sample is not matched in stellar mass, redshift, or selection (e.g., L_IR limits), and the paper does not show that the proto-cluster DSFGs are more disk-dominated than field DSFGs at fixed mass. Without such a differential test, the claim about the origin of the morphology-density relation is an interpretation rather than a measurement. I suggest either adding a matched control comparison or softening the conclusion to state that the ancestors of local cluster ellipticals were disk-dominated at z ≈ 3 in this proto-cluster.","section":"§5.3, Abstract"}],"minor_comments":[{"comment":"The sentence listing 5σ limiting magnitudes gives five values for four filters ('29.74, 29.01, 29.15, 29.49, 29.00 AB mag for F115W, F200W, F356W, and F444W'); please correct the number of values or filters.","section":"§2.3"},{"comment":"The phrase 'The recipitable water vapor' should be 'The precipitable water vapor'.","section":"§2.2.1"},{"comment":"There is a typo: 'invisiblein the bluer bands' should read 'invisible in the bluer bands'.","section":"§4.1"},{"comment":"The ADF22.A4 masked fit reports b/a = 1.0 ± 0.1 with an essentially unconstrained position angle (114.1 ± 100.5 deg); it would be helpful to note that PA is meaningless for a round source and to exclude this PA from the correlation analysis in §4.3.","section":"Table 3"},{"comment":"The header 'IRCS IRCS' above the R.A. and Dec. columns appears to be an undefined or leftover coordinate-system label; please define it or remove it.","section":"Table 2"}],"recommendation":"major_revision","confidential_remarks":"To the editor: The masking-bias issue is the main technical risk; I would like to see injection-recovery simulations or a two-component fit before acceptance. The sample size is small but the data are unique and the analysis is generally careful. There is no concern about novelty or scope; the paper fits standard astronomy journals. The authors rely heavily on their own prior papers for target selection and context, but this is legitimate as independent measurements."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a solid observational paper and the central disk-plus-core interpretation survives the masking worry well enough to publish. What is new: first matched-resolution ALMA and JWST morphology for nine z~3.09 DSFGs in a protocluster core, including resolved stellar morphologies (F444W) and sub-kpc dust structures. The data handling is careful: flux cross-checks, beam-convolved Sersic fits, masked/unmasked comparisons, WebbPSF-matched galfit, and clear caveats. The best evidence for disks is not n~1 alone but the consistent axis ratios and position angles between the masked ALMA fits and the F444W stellar light; several sources are strongly flattened, which is not what you'd get if the mask were leaving only the wings of a round unresolved core. That independent consistency is a genuine check on the masking bias.\n\nSoft spots, in proportion. The masking concern is valid but not fatal. n drops from 1.37 to 0.86 when bright residuals are masked, so the disk claim does depend on that choice. A two-component (disk+core) fit on the 870um data, or an injection/recovery test, would close the loop; currently the two-component interpretation is inferred from residual maps and the unmasked n. The bigger soft spot is the environmental conclusion. Nine galaxies, one protocluster core, luminosity-selected, no identically processed field control. The comparison to Hodge et al. (2025) is suggestive but not a matched baseline, and the wording 'witnessing the early formation of the morphology-density relation' overstates what a sample this size can establish. The authors flag most limitations in the text; that honesty is real. No circularity issue; the self-citations are legitimate since the sample and photometric context come from earlier independent papers.\n\nWho it's for: DSFG structure people, JWST/ALMA morphology folks, and anyone building the z~3 environmental picture. It deserves a serious referee. A referee should ask for a masking-bias test and a tempered abstract, but the core measurements will stand.","headline":"Solid matched-resolution ALMA+JWST morphology of z~3 protocluster DSFGs; the disk interpretation is well supported, while the morphology-density-relation narrative overreaches the small sample.","tokens_in":26127,"tokens_out":2982,"would_cite":true,"duration_ms":32600,"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":"The paper argues that z≈3 dusty starbursts in a proto-cluster core are predominantly disk-like galaxies with compact cores, so the local morphology-density relation must form after that epoch.","keywords":["dusty star-forming galaxies","submillimeter galaxies","proto-clusters","SSA22","Sersic profiles","galaxy morphology","cosmic web","JWST NIRCam"],"falsifier":"A single deep ALMA 870-micron observation of one of these DSFGs at about 0.05 to 0.08 arcseconds that resolves out the extended component and shows the flux concentrated in unresolved cores and clumps, with no exponential disk remaining in the residual after masking, would falsify the disk claim; alternatively, [C II] or CO kinematics showing no ordered rotation across the apparent disk would weaken the late-type interpretation.","tokens_in":1737,"feed_emoji":"🌌","tokens_out":3134,"duration_ms":76570,"temperature":0.7,"pith_summary":"This paper uses matched ALMA and JWST images, at angular resolutions down to 0.05 arcseconds, to dissect nine dusty star-forming galaxies embedded in the cosmic-web filaments of the SSA22 proto-cluster core at z≈3.09. It claims that once bright compact regions are masked, the 870-micron dust emission follows an exponential disk profile, meaning the dust is spread across disks with superimposed compact cores, clumps, and bar-like ridges. JWST F444W images show Sersic indices between about 1 and 3, indicating stellar bulges coexisting with disks, and the close agreement of axis ratios and position angles between the dust and stellar fits argues that the dust truly traces the disks. The authors conclude that the most massive galaxies in this dense high-redshift environment are predominantly late-type, in sharp contrast to the early-type dominance of local cluster cores, and interpret this as witnessing the early assembly phase of the morphology-density relation.","feed_headline":"Proto-cluster starbursts at z≈3 are mostly disks, not bulges","feed_subtitle":"ALMA and JWST show exponential dust disks with compact cores, placing the ancestors of local ellipticals in a late-type phase.","key_machinery":"The central tool is two-dimensional Sersic profile fitting, applied to ALMA 870-micron images at about 0.15 arcseconds and to JWST F444W images at matched resolution, with iterative masking of residual pixels above 5-sigma in the ALMA fits to expose an underlying smooth disk component. The Sersic index n characterizes profile concentration: n≈1 corresponds to an exponential disk and high n to a concentrated spheroid, so comparing n, effective radius, axis ratio, and position angle between the two tracers lets the paper test whether dust and stellar emission share the same geometry. A second element is the core dust mass surface density estimated from 0.05-arcsecond 1.1-millimeter maps, combined with the F444W Sersic index and F200W–F444W color, to place the galaxies in an evolutionary diagram of dust-obscured bulge growth.","core_discovery":"The central claim is that, in the core of the SSA22 proto-cluster at z≈3.09, dusty star-forming galaxies have predominantly disk-like morphologies with superimposed compact cores. For the six brightest sources, the 870-micron Sersic index measured after iteratively masking bright residual pixels is close to unity, with a mean n=0.86±0.05, indicating an exponential disk; without the masks the mean rises to n=1.37±0.08, quantifying the extra concentrated core light. The JWST F444W Sersic indices range from about 1 to 3, pointing to coexisting bulges and stellar disks, and the ALMA and F444W measurements agree well in axis ratio and position angle, supporting the interpretation that dust emission originates across the disks. The paper therefore argues that these DSFGs are late-type galaxies undergoing rapid, heavily obscured growth, and that the sharp contrast with the early-type morphology-density relation seen in local clusters means we are observing the early formation of that relation.","pith_inferences":["If these disky DSFGs are truly the ancestors of local ellipticals, their transformation into spheroids must be completed between z≈3 and z≈1; counting mergers or measuring disk settling in proto-cluster cores over that redshift range would test this.","The tight ALMA–F444W axis-ratio correlation seen here, not seen in the general field, could reflect shared gas inflow geometry along cosmic-web filaments; a larger proto-cluster sample can check whether the alignment is systematic.","ADF22.A11's n≈4 F444W profile with faint dust may be a snapshot of inside-out quenching; near-infrared spectroscopy of its gas could confirm that the center is already passive.","The 5-sigma masking step biases the fitted Sersic index downward by construction, so independent checks with visibility-plane modeling or deeper stacking of the extended component are needed before the disk conclusion is generalized."],"forward_implications":["The most massive galaxies in the proto-cluster core at z≈3 are predominantly late-type, so the morphology-density relation must emerge after this epoch.","Dust continuum emission in DSFGs traces star formation across the disk rather than only in compact starburst cores, since the ALMA and F444W axis ratios and position angles agree.","Compact red dusty cores in some DSFGs mark an ongoing, dust-obscured bulge-growth phase, often accompanied by X-ray AGNs, suggesting simultaneous bulge and black-hole growth.","ADF22.A11 may be a transitional object where the center has already quenched, with n≈4 in F444W, while the disk still forms stars, offering evidence for inside-out quenching.","Bars and offset ridges in dust emission in at least ADF22.A1, A3, and A5 indicate that secular bar-driven gas inflow contributes to the starbursts."],"supporting_citations":[{"why":"established the spectroscopic sample of DSFGs at z≈3.1 in the SSA22 proto-cluster that this paper builds on.","marker":"Umehata et al. 2015"},{"why":"showed these DSFGs are embedded in Ly-alpha cosmic-web filaments, providing the environmental context for the morphology-density claim.","marker":"Umehata et al. 2019"},{"why":"supplied the iterative 5-sigma masking recipe for Sersic fits and the general-field comparison showing n≈0.9 disks.","marker":"Hodge et al. 2019"},{"why":"provided stacked detection of faint extended components and dust size measurements supporting the disk interpretation.","marker":"Gullberg et al. 2019"},{"why":"supplied matched ALMA and JWST morphology measurements for general-field DSFGs that the ADF22 results are compared with.","marker":"Hodge et al. 2025"},{"why":"defined the local morphology-density relation that the paper claims is seen in its early formation stage.","marker":"Dressler 1980"},{"why":"presented the ADF22-WEB project and the detailed bar and disk case of ADF22.A1 that anchors the disk and bar interpretation.","marker":"Umehata et al. 2024"},{"why":"provided the X-ray AGN identification and SED modeling used to interpret the red cores and AGN hosts.","marker":"Monson et al. 2023"}],"fun_headline_variants":["z≈3 proto-cluster DSFGs are disks, not bulges","ALMA+JWST: dusty starbursts in z≈3 cluster are disk-like","Proto-cluster galaxies at z≈3 are disks, not ellipticals","JWST+ALMA reveal z≈3 cluster starbursts as disk-like","In a z≈3 proto-cluster, dusty starbursts are mostly disks"],"cache_read_input_tokens":28160,"weakest_assumption_plain":"The central assumption is that iteratively masking the brightest 5-sigma pixels in the ALMA images reveals a real pre-existing smooth exponential disk, rather than manufacturing one by fitting only the faint wings of unresolved compact cores.","fun_headline_variants_meta":{"raw":{"variants":["z≈3 proto-cluster DSFGs are disks, not bulges","ALMA+JWST: dusty starbursts in z≈3 cluster are disk-like","Proto-cluster galaxies at z≈3 are disks, not ellipticals","JWST+ALMA reveal z≈3 cluster starbursts as disk-like","In a z≈3 proto-cluster, dusty starbursts are mostly disks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000784,"raw_usage":{"total_tokens":3549,"prompt_tokens":1122,"completion_tokens":2427,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":738,"completion_tokens_details":{"reasoning_tokens":2319}},"tokens_in":738,"tokens_out":2427,"duration_ms":15454,"temperature":1.0,"reasoning_tokens":2319,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T14:10:25.572255+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A single deep ALMA 870-micron observation of one of these DSFGs at about 0.05 to 0.08 arcseconds that resolves out the extended component and shows the flux concentrated in unresolved cores and clumps, with no exponential disk remaining in the residual after masking, would falsify the disk claim; alternatively, [C II] or CO kinematics showing no ordered rotation across the apparent disk would weaken the late-type interpretation.","supporting_citations":[],"review_version":1}