{"id":"d49e20b0-dfad-4fc8-8931-a58717e11356","arxiv_id":"2505.04299","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"In 1829 WALLABY hydrogen detections, 315 are low surface brightness galaxies and 55 have no optical counterpart; after reliability vetting, 25 isolated dark sources remain as dark galaxy candidates.","lead":"Giant radio telescopes found 55 galaxies that are invisible, or nearly invisible, in deep optical images, but have lots of hydrogen gas. After careful checks, 25 of these are new candidate 'dark galaxies' that could change how we think galaxy formation works.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unquantified SoFiA false-positive rate is the core risk to the 25 non-tidal dark-source candidates; the paper openly acknowledges this, so its central claim is appropriately conditional.","rationale":"The reader's weakest_assumption is exactly the load-bearing concern I identify: the unknown SoFiA false-positive rate. My read of the full text confirms that the paper does not quantify this rate and explicitly defers it to future work, while simultaneously using the 38/25/68 per cent numbers as the basis for the 'significant gap' statement. That is a genuine soft spot in the chain from data to headline interpretation. Nevertheless, the paper's wording is carefully conditional throughout — candidates, not confirmed detections; follow-up observations explicitly required — and it provides some external anchors (three sources with Parkes/HIPASS detections, one source with a faint co-added optical counterpart). For an observational pilot paper, this level of hedging is appropriate, and the stated conclusion is not overclaimed. The concern does not reveal an internal inconsistency or a fatal flaw; it identifies the main scientific uncertainty that follow-up observations are designed to settle. Therefore I do not recommend changing the reader's ACCEPT verdict, although confidence should remain moderate. I agree with the reader's identification of the same weakest assumption.","tokens_in":36964,"tokens_out":5576,"duration_ms":60738,"concrete_test":"Generate noise-only data cubes matched to the three pilot fields (same rms noise, channel width, and RFI flagging) and run the identical SoFiA source-finding and the same SNR > 4.9 plus visual-vetting chain; count how many spurious candidates survive the full 'strong dark source' selection. If the inferred false-discovery rate is above roughly 10–20 per cent of 38, the 25 non-tidal candidates and the 68 per cent high-mass fraction are not robust; if it is below a few per cent, the central reliability concern is retired.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result — 38 strong dark HI detections, 25 of them non-tidal, with 68 per cent of the non-tidal sources above M_HI = 10^9 M_sun — depends on the assumption that the vast majority of these WALLABY detections are genuine astrophysical sources. Section 3.3 explicitly states that 'the false positive rate of the SoFiA implementation in the WALLABY pipeline still requires further investigation' and that catalogue notes caution against trusting these sources without further analysis. The paper's vetting (peak SNR > 4.9, visual inspection of cubes and spectra, RFI checks) reduces but does not quantify the false-positive rate; no injected-source, blank-sky, or inverted-cube test is reported. If a meaningful fraction of the 38 'strong' sources are noise artifacts or residual continuum/RFI features, both the count of 25 isolated candidates and the 68 per cent high-mass fraction would change. The paper itself flags that seven dark sources are within a factor of two of the detection limit and 'could simply be false positives', and it does not state whether these seven are among the strong or uncertain set. A secondary weakness is that the tidal/non-tidal split relies on proximity to NED galaxies within ±300 km/s and qualitative HI asymmetry; incomplete parent-galaxy catalogs could misclassify some tidal debris as isolated. However, the authors consistently present the final interpretation as conditional ('if these dark galaxy candidates are revealed to be genuine detections'), and the three Parkes/HIPASS coincidences plus the one faint optical counterpart provide partial external support. The argument is therefore not internally inconsistent; its weakest load-bearing point is the unmeasured false-positive rate.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper cross-matches 1829 WALLABY HI detections in the Hydra, NGC 4808, and NGC 5044 pilot fields against DESI Legacy Survey DR10 imaging. It classifies 315 detections as optically low surface brightness galaxies (LSBGs; mean g-band surface brightness within 1 Re fainter than 23 mag arcsec−2) and 55 as optically dark HI sources. After reliability vetting, the authors retain 38 strong dark-source candidates, 13 of which show tidal signatures and 25 of which are presented as isolated dark galaxy candidates. The paper measures HI sizes and masses, derives stellar masses and star formation rates from Sérsic-model photometry with external calibrations, and compares the LSBGs against xGASS, ALFALFA, and other WALLABY samples. The headline quantitative claim is that 68 per cent of the 25 non-tidal dark sources have M_HI > 10^9 Msun, which would indicate a significant gap in galaxy formation models if these are genuine detections. The authors consistently frame the dark-source interpretation as conditional and call for deeper HI and optical follow-up.","tokens_in":37257,"tokens_out":8662,"duration_ms":91984,"significance":"If the dark candidates are genuine, this is one of the largest and most homogeneous samples of isolated HI-rich, optically dark galaxy candidates to date, and it directly challenges theoretical expectations that dark galaxies should be predominantly low-mass systems. The LSBG sample is also valuable: it is HI-selected, spans five orders of magnitude in stellar mass, and demonstrates that gas-rich low-surface-brightness systems have low star formation efficiencies. The paper is commendably careful in several ways: it separates strong from uncertain dark detections, uses stellar mass and SFR calibrations derived for LSBGs from independent work, compares against external HI-selected samples, and states its main conclusions conditionally. The principal limitation, acknowledged by the authors, is that the SoFiA false-positive rate is not quantified; the central 38/25/68 per cent numbers depend on the assumption that most of these detections are astrophysical. Because the paper is a candidate catalogue rather than a confirmation paper, this limitation does not invalidate the work, but it is load-bearing for the headline claim.","major_comments":[{"comment":"The paper states that seven dark sources have HI masses less than a factor of two above the detection limit and 'could simply be false positives', but it does not state whether these seven are among the 38 strong candidates or the 17 uncertain detections. Since the subsequent central counts (13 tidal, 25 isolated, and the 68 per cent fraction with M_HI > 10^9 Msun) are computed from the strong set, the reader cannot assess how much of the headline result rests on sources the authors themselves regard as potentially unreliable. Please state the overlap explicitly, and if any of the seven are in the strong set, recompute the counts and percentages with and without them.","section":"Section 3.3"},{"comment":"The reliability of the 38 strong dark detections rests on a peak SNR threshold of 4.9 plus visual inspection of cubes and spectra, but no blank-sky, inverted-cube, or injected-source test is reported, and the paper acknowledges that the SoFiA false-positive rate 'still requires further investigation'. Because the central claim is the existence of 25 isolated dark galaxy candidates, this is a load-bearing uncertainty. A quantitative statement of expected contamination—for example, from pipeline validation tests or a conservative upper limit on false positives among the 38—would materially strengthen the paper. The current conditional language is appropriate, but it does not replace a quantitative reliability assessment.","section":"Section 3.3 and Section 6"}],"minor_comments":[{"comment":"The caption contains a typo: 'pannels' should be 'panels'.","section":"Appendix A caption"},{"comment":"The phrase 'haromonics of radionavigation satellites' contains a typo; it should be 'harmonics'.","section":"Section 3.3"},{"comment":"The entries in Table 3 run together in the rendered text (for example, '22.2±2.08.95±0.03' and '4.6±1.37.97±0.07'), making it difficult to separate D_HI, log M_HI, log M*, and log SFR. Please add clear column spacing or separators.","section":"Table 3"},{"comment":"The row for WALLABY J133747-175606 appears to lack the stellar-mass upper limit and SFR upper limit entries that are present for the other strong candidates; please add the missing values or explain why they are omitted.","section":"Table 3"},{"comment":"The extrapolation to roughly 570 isolated dark sources across the full survey uses the three pilot fields, but 34 of the 38 strong dark detections are in the larger NGC 5044 field; given the small-number statistics and field-to-field variation, the extrapolation deserves a more explicit caveat or a simple Poisson uncertainty.","section":"Section 6"}],"recommendation":"major_revision","confidential_remarks":"The paper is an honest and well-crafted candidate catalogue, and the authors' conditional framing is appropriate given the unquantified false-positive rate. The main revision I would require is the clarification of how the seven near-detection-limit sources overlap with the strong/uncertain split, and preferably some quantitative contamination estimate. If the authors can show that none of the seven are in the strong sample, or that the headline statistics are robust to their inclusion, the paper could be accepted without new observations."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is the first systematic optical classification of the full WALLABY pilot sample in the Hydra, NGC 4808 and NGC 5044 fields, and it delivers a useful catalog of 315 LSBGs and 55 dark HI candidates. The 25 isolated non-tidal dark sources with high HI masses are the headline, but the authors frame them exactly as candidates requiring follow-up.\n\nWhat's genuinely good: the paper is honest about its own weaknesses. It explicitly separates 38 strong from 17 uncertain dark detections, notes that the SoFiA false-positive rate is not yet quantified, flags seven sources within a factor of two of the detection limit as possible false positives, and identifies RFI-affected candidates (all already in the uncertain bin). The photometry uses ASTROPHOT Sersic fits in a consistent multi-band manner, stellar masses use an LSBG-specific scaling relation (Du et al. 2020), and comparisons are made against external samples (xGASS, ALFALFA, Wang et al. 2016). The LSBG sample itself is a valuable resource, spanning five orders of magnitude in stellar mass. The scaling relations (HI size-mass, SFE, sSFR) are sensible, and the paper finds the LSBGs follow the Wang et al. relation tightly, which is a nice result.\n\nSoft spots, in proportion: the load-bearing concern is the false-positive rate of the SoFiA detections. The paper acknowledges it but does not quantify it; no injected-source or blank-sky tests are reported. If a meaningful fraction of the 38 'strong' sources are artifacts, the 25 isolated candidates and the 68% high-mass fraction would change. The paper's conditional language ('if these dark galaxy candidates are revealed to be genuine detections') is appropriate, so this is a limitation rather than an internal inconsistency. Also, the tidal/non-tidal classification relies on proximity to NED galaxies within ±300 km/s and qualitative HI asymmetry; incomplete parent catalogs could misclassify some tidal debris as isolated. Minor: the paper doesn't state whether the seven near-limit sources are among the strong or uncertain set, and the derived catalog is only available on request, which is a reproducibility minus.\n\nWho it's for: astronomers working on HI-selected galaxies, LSBGs, and dark galaxy candidates. It deserves serious peer review—the catalog and the carefully qualified interpretation are worth publishing. I'd recommend accepting after minor revisions, mainly asking for a quantitative false-positive estimate or at least a clear statement of which sources are near the limit and whether any independent confirmation (e.g., the three Parkes coincidences) strengthens the strong set.","headline":"A careful, well-hedged census of optically dark and low-surface-brightness HI sources in WALLABY pilot data; the 25 isolated dark candidates are interesting but the unknown SoFiA false-positive rate is the main caveat, and the paper already says so.","tokens_in":37950,"tokens_out":2445,"would_cite":true,"duration_ms":23663,"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":"A census of 1829 HI detections yields 25 isolated, optically dark galaxy candidates.","keywords":["neutral hydrogen","dark galaxies","low surface brightness galaxies","WALLABY survey","ASKAP","galaxy formation","HI size-mass relation","DESI Legacy Surveys"],"falsifier":"Point a high-sensitivity, high-resolution radio telescope such as MeerKAT at the 25 non-tidal dark candidates and integrate until each putative HI line is either confirmed or ruled out at $5\\sigma$. If most candidates fail to reproduce, show continuum artefacts, or reveal asymmetric debris tied to neighbours, the count of 25 isolated dark galaxies and the 68 per cent figure above $10^9\\,M_\\odot$ collapse. Deep optical imaging that reveals stellar counterparts for a large fraction would likewise rule out the dark-galaxy interpretation.","tokens_in":36783,"feed_emoji":"🌌","tokens_out":6847,"duration_ms":61297,"temperature":0.7,"pith_summary":"The paper asks what fraction of gas-rich galaxies are invisible in optical light and what those objects are. Examining 1829 neutral hydrogen (HI) detections from three WALLABY pilot fields against deep Legacy Survey imaging, it identifies 315 low surface brightness galaxies (17 per cent) and 55 optically 'dark' HI sources. After reliability vetting, 38 dark sources remain strong candidates; 13 show tidal signatures, leaving 25 with no tidal origin. Those 25 are candidates for isolated galaxies with high HI masses but very low stellar masses and star-formation rates. If genuine, they point to a galaxy population that current formation models do not predict in large numbers.","feed_headline":"25 dark HI sources may be isolated, gas-heavy galaxies","feed_subtitle":"If confirmed, 68 percent hold over a billion suns of hydrogen, challenging galaxy formation models.","key_machinery":"The machinery is a matched HI-optical census. WALLABY's 30-arcsecond ASKAP beams and the SoFiA source-finding pipeline produce the HI catalogue; the DESI Legacy Imaging Surveys DR10 images provide the optical depth that defines 'dark'; ASTROPHOT Sersic models supply photometry, surface brightnesses, stellar masses and star-formation rates; and a reliability filter (peak signal-to-noise above 4.9, visual inspection of cubes and spectra, radio-frequency-interference checks, tidal-neighbour analysis) separates 38 strong dark candidates from 17 uncertain detections. The HI size-mass relation from the literature acts as a physical consistency check: resolved dark sources and LSBGs fall on the same relation as normal galaxies.","core_discovery":"The central claim is that a substantial population of gas-rich objects exists with no detectable optical counterpart, and that the non-tidal members are mostly massive: 68 per cent of the 25 non-tidal dark candidates have HI masses above $10^9\\,M_\\odot$, with stellar mass upper limits implying $M_{\\rm HI}/M_* > 187$ for all and $> 500$ for 89 per cent of them. The paper argues that these objects are either extreme low surface brightness galaxies hidden below current optical limits, genuine dark galaxies, or artefacts, and that deeper HI and optical follow-up is required to decide. It also establishes that the well-resolved LSBGs and dark sources follow the standard HI size-mass relation, so their optical faintness coexists with typical HI properties. The authors state that if the dark candidates are confirmed, this indicates a significant gap in the current understanding of galaxy formation.","pith_inferences":["If follow-up confirms the candidates, their high HI masses would clash with simulations that place most dark galaxies in halos below about $10^9\\,M_\\odot$, suggesting either a new formation channel for massive 'failed' halos or an extreme, gas-rich extension of the LSBG population.","The contrast with ALFALFA, which found almost no dark galaxy candidates, may stem from WALLABY's better angular resolution reducing source confusion; the same approach on the remaining pilot fields would test whether this population is ubiquitous.","Stacking the deep optical images at the HI positions could reveal cumulative ultra-faint stellar light below the single-source detection limit, offering a test of the dark-galaxy interpretation before deeper imaging arrives.","Applying the same reliability vetting to the full survey's detections could turn the 25 candidates into a statistical sample, allowing environment and halo-mass comparisons with simulations."],"forward_implications":["Among 1829 HI detections, 17 per cent are low surface brightness galaxies and 3 per cent are optically dark, so blind HI surveys substantially enlarge the optically faint galaxy census.","Seventy-five per cent of the LSBGs and all 55 dark HI sources were previously uncatalogued, showing that existing southern-sky catalogues miss most of this population.","The gas-rich LSBGs have low star-formation efficiencies while spanning five orders of magnitude in stellar mass, from irregular dwarfs to massive diffuse spirals.","If the non-tidal dark sources are not preferentially distributed, the full WALLABY survey should yield roughly 570 isolated dark HI sources.","The well-resolved LSBGs and dark sources follow the HI size-mass relation, so extreme optical faintness does not imply unusual HI structure."],"supporting_citations":[{"why":"Describes WALLABY, the survey whose three pilot fields supply the 1829 HI detections analysed here.","marker":"Koribalski et al. 2020"},{"why":"Phase 1 pilot data release providing the Hydra field catalogue and flux corrections.","marker":"Westmeier et al. 2022"},{"why":"Phase 2 pilot data release providing the NGC 4808 and NGC 5044 catalogues and flux corrections.","marker":"Murugeshan et al. 2024"},{"why":"The DESI Legacy Imaging Surveys, the optical data in which counterparts are sought and 'dark' is defined.","marker":"Dey et al. 2019"},{"why":"Measured the Legacy Survey surface brightness depths used to set the dark-source optical limits.","marker":"O'Beirne et al. 2024"},{"why":"The ALFALFA almost-dark galaxy sample against which the WALLABY dark sources are compared in HI mass and line width.","marker":"Leisman et al. 2017"},{"why":"The HI size-mass relation that the resolved LSBGs and dark sources are shown to follow.","marker":"Wang et al. 2016"},{"why":"Theoretical prediction of dark galaxy number densities that the high-mass candidates may contradict.","marker":"Jimenez & Heavens 2020"},{"why":"IllustrisTNG-based predictions for dark galaxy formation and environments used to frame the candidates.","marker":"Lee et al. 2024"}],"fun_headline_variants":["25 dark HI sources may be isolated, gas-rich galaxies","Most dark HI candidates are massive, isolated galaxies","Dark HI objects: most are massive, isolated galaxies","Gas-rich but optically dark: 25 isolated galaxy candidates"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the 38 'strong' HI detections without optical counterparts are real astrophysical sources rather than noise, interference, or imaging artefacts; the paper notes that the false positive rate of its source-finding pipeline still requires further investigation.","fun_headline_variants_meta":{"raw":{"variants":["25 dark HI sources may be isolated, gas-rich galaxies","Most dark HI candidates are massive, isolated galaxies","Dark HI objects: most are massive, isolated galaxies","Gas-rich but optically dark: 25 isolated galaxy candidates"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000846,"raw_usage":{"total_tokens":3718,"prompt_tokens":1018,"completion_tokens":2700,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":634,"completion_tokens_details":{"reasoning_tokens":2636}},"tokens_in":634,"tokens_out":2700,"duration_ms":17699,"temperature":1.0,"reasoning_tokens":2636,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:32:18.787619+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Point a high-sensitivity, high-resolution radio telescope such as MeerKAT at the 25 non-tidal dark candidates and integrate until each putative HI line is either confirmed or ruled out at $5\\sigma$. If most candidates fail to reproduce, show continuum artefacts, or reveal asymmetric debris tied to neighbours, the count of 25 isolated dark galaxies and the 68 per cent figure above $10^9\\,M_\\odot$ collapse. Deep optical imaging that reveals stellar counterparts for a large fraction would likewise rule out the dark-galaxy interpretation.","supporting_citations":[],"review_version":1}