{"id":"f82c3c62-cbaa-4425-8aa4-441ad32803c0","arxiv_id":"2412.12573","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Using marked correlation functions, UV magnitude and color of z=3-5 Lyman-break galaxies show strong environmental correlations, while SFR and stellar mass show weak ones.","lead":"At redshifts 3 to 5, the UV brightness and color of Lyman-break galaxies are strongly tied to their surroundings, while star formation rate and stellar mass show only weak links. The result introduces a high-redshift environmental probe and a target for galaxy formation simulations.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The apparent-magnitude and dropout-color marks are strongly redshift-dependent within the broad LBG selection windows; the reported MCF excess may reflect mark-redshift clustering rather than environmental dependence of intrinsic galaxy properties.","rationale":"The reader's weakest assumption identifies exactly the same load-bearing concern: the marks used for the principal result, apparent UV magnitude and dropout color, are plausibly correlated with redshift inside the broad LBG photo-z window, and this can inflate or fully generate the MCF signal without any environmental dependence. My independent reading of Sections 2.3, 3.1, and the robustness tests in Section 4.6 confirms that the paper does not break this degeneracy. The MCF ratio removes the overall clustering amplitude but not the mark-redshift covariance; the interlopers test changes the sample edges but not the internal N(z)-mark relation, and the magnitude-noise test only weakens the coupling. This is not an internal inconsistency, but it is a genuine correctness risk for the central claim that UV magnitude and color are environmental tracers. The SFR and M* results are less affected by this particular degeneracy but are also based on UV-incomplete samples and are not a substitute for the missing control on the primary marks. The paper has useful strengths: it is a first application of MCF at z~3-5, uses large HSC/CLAUDS samples, and includes several honest caveats. However, the key measurement should be accepted only conditionally on a test that removes the mark-redshift covariance. Since the reader already assigned CONDITIONAL with this concern, I see no need to change the verdict; I would keep it conditional pending the narrow-photo-z-bin or absolute-magnitude rank test.","tokens_in":18992,"tokens_out":3424,"duration_ms":36809,"concrete_test":"Recompute the magnitude- and color-marked MCFs after assigning ranks within narrow photometric-redshift bins, e.g., Delta z = 0.3, for the z~4 g-dropout sample. Also compare the mean photo-z as a function of apparent-magnitude rank: if the mean photo-z changes substantially across the rank range, the mark is not intrinsic. If the MCF drops from its reported ~1.2-1.3 level toward unity in the within-bin ranking, the signal is a selection artifact. A complementary check is to use absolute UV magnitude M_UV = m_UV - DM(z_phot) as the mark and recompute the MCF; if the signal largely disappears, the claimed environmental dependence is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that rank-ordered marked correlation functions for UV magnitude and dropout color are significantly above unity and therefore trace environment at z~3-5. The load-bearing assumption is that these marks encode intrinsic galaxy properties rather than redshift within the selected sample. That assumption is not secure.\n\nSection 3.1 assigns ranks using the apparent UV magnitude and the dropout color across the full sample at each redshift. Within a dropout-selected sample, both marks are expected to correlate with redshift: the sample is apparent-magnitude limited, so fainter apparent magnitudes preferentially select higher redshifts near the flux limit, and the Lyman-break color itself is a monotonic function of redshift across the selection window (Eqs. 1-3). Photometric redshifts in Section 2.1 have scatter sigma_z ~ 0.1-0.15, so the resulting N(z) is broad.\n\nThe MCF estimator in Eq. 5, M(theta) = WW(theta)/DD(theta), divides out the average pair clustering but does not remove correlation between marks and redshift. If galaxies at similar redshift cluster angularly, pairs at small theta preferentially share similar redshifts and therefore similar marks. The mean mark product for such pairs exceeds the global mean even if the intrinsic property has no environmental dependence. Thus a positive MCF signal can be generated entirely by the mark-redshift covariance.\n\nSection 2.3 argues that the ratio form makes the MCF 'less affected by N(z)'. This is only true for an overall normalization of the redshift distribution; it does not address the differential mark-redshift correlation that drives the pair weights. The robustness tests in Fig. 6 (changing z95 interloper cuts) and Fig. 7 (adding magnitude noise) do not break this degeneracy: the interloper test leaves the broad N(z) inside the selection window intact, and the noise test only dilutes the mark-redshift coupling rather than removing it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper measures rank-ordered marked correlation functions (MCFs) for Lyman-break galaxies at z~3, 4, and 5, using HSC-SSP and CLAUDS data. Marks are apparent UV magnitude, dropout color, stellar mass, and star formation rate. The authors report that the UV magnitude and color MCFs deviate significantly above unity on scales from a few arcseconds to hundreds of arcseconds, with stronger signals in brighter samples, whereas the M* and SFR MCFs remain near unity. They also compare samples matched in absolute magnitude, effective large-scale bias, and effective halo mass across redshifts, finding redshift evolution in the marked clustering strength. The paper interprets these signals as evidence that UV magnitude and color are strong tracers of the high-redshift environment.","tokens_in":19252,"tokens_out":2039,"duration_ms":20671,"significance":"If the main claim is correct, this is one of the first measurements of the environmental dependence of LBG properties using the MCF at z~3-5, and it extends MCF studies from the local universe to high redshift with a much larger survey area than previous environment studies at these redshifts. The paper has several concrete strengths: the MCF is computed directly from pair counts with a random-catalogue-free estimator; the authors check robustness to low-z interloper cuts (Fig. 6, left) and to magnitude noise (Fig. 6, right and Fig. 7); and they compare samples matched in bias and halo mass across redshifts. However, the central interpretation depends on an unverified assumption that the apparent-magnitude and dropout-color marks trace intrinsic galaxy properties rather than redshift within the broad photometric-redshift selection windows. The paper does not currently establish this, which is a load-bearing gap for the main conclusions.","major_comments":[{"comment":"The central claim that UV magnitude and dropout color are effective tracers of environment is not yet secure against a redshift-mark covariance. Marks are assigned using apparent UV magnitude and dropout color across the full sample (Section 3.1), and both quantities correlate with redshift inside the broad photo-z windows: the sample is apparent-magnitude limited, and the Lyman-break color is itself a monotonic redshift indicator over the selection region (Eqs. 1-3). Since galaxies at similar redshift cluster angularly, pairs at small θ preferentially share similar marks, so the ratio WW(θ)/DD(θ) in Eq. (5) can exceed unity even if the intrinsic property has no environmental dependence. The statement in Section 2.3 that the MCF is 'less affected by N(z)' addresses only the overall normalization of the angular correlation function, not covariance between marks and redshift. The paper should test this directly, for example by re-ranking galaxies within narrow photo-z bins, using absolute UV magnitude as the mark, or shuffling redshifts in a mock sample; without such a test, the interpretation of the large MCF amplitudes as environmental dependence of intrinsic properties is not established.","section":"§2.3 and §3.1, Eq. (5)"},{"comment":"The redshift evolution of the MCF extracted by comparing samples with matched absolute magnitude, effective bias, or effective halo mass may also be affected by the mark-redshift covariance, because the width of N(z) and the photo-z scatter vary between the z~3, 4, and 5 samples (Section 2.1, Fig. 1). If the MCF excess is partly driven by redshift-sorted pairs, the differences in amplitude across redshifts could reflect differences in the N(z) width rather than genuine evolution of environmental dependence. The authors should either quantify this effect for their cross-redshift comparisons or restrict the cross-redshift comparison to a control test that removes the redshift-mark correlation.","section":"§4.3-§4.5, Figs. 3-5"}],"minor_comments":[{"comment":"There are several typographical issues, including the section title 'THE DATE' (should be 'THE DATA'), 'explicitely' (Section 3), and 'weighing' where 'weighting' is meant (Section 3). The manuscript would benefit from a careful proofread.","section":"Throughout"},{"comment":"The caption lists multiple magnitude thresholds for each redshift but the mapping between panels and thresholds is not self-explanatory; please clarify which panel corresponds to which survey layer and threshold, or annotate the panels directly.","section":"Fig. 2 caption"},{"comment":"The discussion of why SFR does not trace environment is plausible but speculative; explicitly noting that the SFR/M* marks are measured with larger scatter (as acknowledged in Section 4.6) would help the reader weigh this argument before the explanation based on star-formation history.","section":"Section 4.1, paragraph on SFR"}],"recommendation":"major_revision","confidential_remarks":"The paper is appropriate for the Open Journal of Astrophysics in scope, and the measurement is a useful contribution to high-redshift environment studies. The main technical concern is the mark-redshift covariance described in the major comments; it is a correctness issue for the central claim rather than a presentation issue. I would not recommend rejection, because the problem is addressable with additional control statistics within the current data, but the manuscript needs a major revision to either demonstrate that the effect is subdominant or substantially reframe the conclusions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a genuinely new measurement—the first marked correlation function for Lyman-break galaxies at z=3-5—and it's carefully executed. The soft spot is the one you'd expect: the headline marks, apparent UV magnitude and dropout color, are entangled with redshift inside the broad photo-z windows.\n\nWhat it does well: the estimator is standard, the errors are jackknife and scrambled-mark, and they test interloper cuts and magnitude noise. The comparison of samples matched by effective bias and halo mass across redshift is a thoughtful way to ask whether the environmental signal evolves, and they're upfront that the SFR/stellar-mass samples are UV-incomplete. That last point means the weak MCFs for those properties are not a surprise, but they also don't serve as a clean control.\n\nThe load-bearing concern is the mark-redshift degeneracy. Within each dropout-selected sample, apparent magnitude correlates with redshift because of the flux limit, and the dropout color is a redshift indicator by construction. Galaxies at similar redshift cluster angularly, so pairs at small separation share similar marks even if the intrinsic properties have no environmental dependence. The ratio estimator divides out the overall N(z) normalization, but not the mark-redshift covariance. The interloper and noise tests change sample boundaries or add scatter; they don't break the covariance. The paper's statement that the MCF is 'less affected by N(z)' is too quick.\n\nHow much does this matter? The small-scale signal could well be physical—central-satellite effects are a plausible source—but as written the paper can't separate that from the mark-redshift effect. The redshift evolution claims are similarly vulnerable.\n\nI'd send this to a serious referee. The measurements are useful and the method is standard. What's missing is a control: re-rank within narrow photo-z bins, use absolute magnitudes or rest-frame colors as marks, or explicitly model the mark-redshift covariance. If the signal survives, it's a solid result. If not, the paper still documents an important systematic for future MCF work. Bring it to the reading group—the degeneracy discussion is worth having.","headline":"A genuinely new first measurement of MCFs for z~3-5 LBGs, but the UV magnitude and color marks are entangled with redshift inside broad photo-z windows, so the headline signal needs a control before I'd trust it.","tokens_in":636,"tokens_out":1497,"would_cite":false,"duration_ms":36630,"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 claims that UV magnitude and color of Lyman-break galaxies at $z \\sim 3$–$5$ are strong environmental tracers, while stellar mass and star-formation rate are not.","keywords":["marked correlation function","Lyman-break galaxies","galaxy environment","UV luminosity","galaxy color","star formation rate","stellar mass","high-redshift galaxy clustering"],"falsifier":"Re-rank the galaxies by absolute UV magnitude or within narrow photometric-redshift bins and recompute the same marked correlation functions; if the UV-magnitude and color MCFs drop toward unity while the same pairs and the same redshift distribution are kept, the claimed environmental dependence would be shown to be a selection artifact.","tokens_in":18761,"feed_emoji":"🔭","tokens_out":9687,"duration_ms":82015,"temperature":0.7,"pith_summary":"The paper asks which observable properties of Lyman-break galaxies at redshifts 3 to 5 are tied to their surrounding environment, using marked correlation functions measured from large optical survey data. It finds that UV magnitude and UV dropout color are strong environmental tracers: galaxies with brighter UV light or with more extreme dropout colors are more likely to be found near similar galaxies at separations from a few arcseconds to hundreds of arcseconds, beyond what ordinary clustering predicts. Star-formation rate and stellar mass, by contrast, show only weak or marginal marked clustering, so they are poorer tracers of the environment at these redshifts. The signal is stronger for brighter samples, grows from $z \\sim 5$ to $z \\sim 3$ when samples are matched in absolute magnitude or halo mass, and is larger than $z \\sim 0$ measurements, giving a survey-scale benchmark for how environment shapes young galaxies.","feed_headline":"Brighter UV galaxies cluster more at z=3-5; mass and SFR do not","feed_subtitle":"UV magnitude and color, not stellar mass or star formation, trace the environment of galaxies at z ~ 3-5.","key_machinery":"The central object is the rank-ordered marked correlation function, defined as $M(\\theta) = (1+W(\\theta))/(1+\\omega(\\theta)) \\equiv WW(\\theta)/DD(\\theta)$, where each galaxy receives a mark equal to the percentile rank of a chosen property, rescaled to lie between 0 and 2. The ratio of weighted to unweighted pair counts cancels survey geometry and, to first order, the redshift distribution $N(z)$, allowing different galaxy properties to be compared on the same footing. A value above unity means that pairs with high mark values are overrepresented relative to ordinary angular clustering, and a value below unity means they are underrepresented.","core_discovery":"The central claim is that, in magnitude-selected Lyman-break galaxy samples at $z \\sim 3$, $4$, and $5$, apparent UV magnitude and UV dropout color are strongly correlated with environment, while stellar mass and star-formation rate are not. Concretely, when galaxies are rank-ordered by UV magnitude or color and the marked correlation function is measured, the statistic significantly exceeds unity on scales up to roughly 400 arcseconds for the brightest samples, indicating an excess of pairs of similarly bright or similarly colored galaxies over the expectation from the two-point angular correlation function. The effect is stronger for brighter threshold samples at every redshift, persists after adding magnitude noise, and is robust to changing the lower-redshift interloper cut. The same-absolute-magnitude comparison shows the signal generally strengthens from $z \\sim 5$ to $z \\sim 3$, samples with the same effective halo mass also show stronger marked clustering at lower redshift, and samples matched in effective large-scale bias show comparable large-scale signals at $z \\sim 4$ and $5$.","pith_inferences":["The authors leave implicit that, because the samples are selected by apparent UV magnitude and dropout color and the photometric-redshift window is broad, the rank ordering itself may partly order galaxies by redshift; recomputing the MCF after splitting into narrow photometric-redshift slices, or ranking by absolute magnitude, would test whether the environmental signal is intrinsic.","A further inference is that the comparison with $z \\sim 0$ marked correlation functions is not apples-to-apples, since the low-redshift studies use different marks, depths, and selection functions; part of the stronger high-redshift amplitude could reflect these methodological differences rather than true evolution.","One testable extension is to repeat the analysis with dust-corrected UV luminosities and infrared- or radio-based star-formation rates; if the weak SFR signal is caused by scatter in SED-derived values, the MCF for these alternative marks should rise toward the UV-magnitude signal."],"forward_implications":["If the claim holds, UV magnitude and color can serve as practical environmental tracers for Lyman-break galaxies at $z \\sim 3$–$5$ without relying on expensive spectroscopy.","The signals persisting to separations of hundreds of arcseconds imply that environmental influence on UV properties extends beyond individual dark matter halos, pointing to two-halo conformity or assembly bias.","The weak stellar-mass and star-formation marked clustering implies that, at these redshifts, photometrically derived mass and star-formation rate are not reliable indicators of environment, in contrast to the local Universe.","The observed trend that marked clustering strengthens from $z \\sim 5$ to $z \\sim 3$ for samples matched in absolute magnitude or halo mass suggests environmental correlations grow as cosmic structure develops."],"supporting_citations":[{"why":"Supplies the HSC-SSP galaxy catalogues, masks, and random catalogues from which the LBG samples and pair counts are built.","marker":"Aihara et al. 2022"},{"why":"Provides the CFHT U-band deep imaging that enables U-dropout selection of the $z \\sim 3$ sample.","marker":"Sawicki et al. 2019"},{"why":"Supplies the combined CLAUDS/HSC-SSP catalogue with photometric redshifts and derived stellar masses and star-formation rates for the ultra-deep fields.","marker":"Desprez et al. 2023"},{"why":"Defines the dropout colour cuts, magnitude thresholds, and bright-source masks used to construct the g- and r-dropout samples.","marker":"Harikane et al. 2022"},{"why":"Introduces the rank-ordered marked correlation function used to compare different galaxy properties as marks.","marker":"Skibba et al. 2013"},{"why":"Introduces the marked correlation function statistic that the analysis applies to high-redshift LBGs.","marker":"Beisbart & Kerscher 2000"},{"why":"Supplies the WW/DD estimator and the mark-scrambling method used to assign statistical uncertainties to the MCF.","marker":"Sheth et al. 2005"},{"why":"Provides the $z \\sim 0$ GAMA/WISE MCF measurements that the paper compares against to argue the high-redshift signals are stronger.","marker":"Sureshkumar et al. 2021"},{"why":"Supplies the photometric-redshift lower-bound criteria used to remove low-z interlopers from the dropout samples.","marker":"Toshikawa et al. 2024"}],"fun_headline_variants":["UV brightness traces environment at z=3-5; mass and SFR don't","Bright LBGs cluster; mass and SFR do not at high z","Galaxy UV magnitude reveals environment, not stellar mass","At z=3-5, UV color marks environment, not SFR","Marked correlations show UV magnitude clusters, mass does not"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that ranking a dropout-selected galaxy sample by apparent UV magnitude and color orders galaxies by intrinsic brightness and color rather than by redshift; if the ranking mostly tracks redshift, the marked correlation signal could appear without any true environmental dependence.","fun_headline_variants_meta":{"raw":{"variants":["UV brightness traces environment at z=3-5; mass and SFR don't","Bright LBGs cluster; mass and SFR do not at high z","Galaxy UV magnitude reveals environment, not stellar mass","At z=3-5, UV color marks environment, not SFR","Marked correlations show UV magnitude clusters, mass does not"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000343,"raw_usage":{"total_tokens":1942,"prompt_tokens":1055,"completion_tokens":887,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":671,"completion_tokens_details":{"reasoning_tokens":794}},"tokens_in":671,"tokens_out":887,"duration_ms":7710,"temperature":1.0,"reasoning_tokens":794,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T13:57:09.291162+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-rank the galaxies by absolute UV magnitude or within narrow photometric-redshift bins and recompute the same marked correlation functions; if the UV-magnitude and color MCFs drop toward unity while the same pairs and the same redshift distribution are kept, the claimed environmental dependence would be shown to be a selection artifact.","supporting_citations":[],"review_version":1}