{"id":"d7a81fe4-ee25-455c-9bfc-9279440deb1f","arxiv_id":"2607.28732","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Host galaxies of extended, low-redshift quasars are mostly quiescent and often post-starburst, unlike the star-forming hosts found for compact quasars.","lead":"Using DESI spectra and Subaru images, the authors split more than 1,000 extended quasars into the light of the black hole and the light of its host galaxy. Most of these host galaxies turn out to be quiet, old, and no longer forming many stars — and about a quarter show post-starburst marks.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"SFR validation for the quiescent/post-starburst classification rests on only 8 mock recoveries; the low-sSFR regime where the 83%/23% fractions live is not independently calibrated.","rationale":"The reader's weakest assumption—that the mock validation is model-internal—is exactly the source of the concern I identify. My reading sharpens it by pointing to the specific evidentiary gap: the SFR recovery validation in Figure 9 uses only 8 confident recoveries, and no low-sSFR recovery statistics are provided. Since the headline claims are quantitative fractions of quiescent and post-starburst hosts, the accuracy of sSFR in the low-sSFR regime is the load-bearing condition. I considered whether the morphological selection effect is more fundamental, but the authors explicitly acknowledge this selection (Section 5.4), so it does not invalidate the descriptive claim for the selected sample. The MBH–σ relation is already presented as non-robust, and the MBH–Mstar relation is a secondary result. Therefore the central quantitative claims stand or fall on the SFR validation. The recommended check is feasible with modest additional computing and would either confirm the current conditional verdict or, if bias is found, force a revision of the quiescent/post-starburst fractions. I do not change the reader's CONDITIONAL verdict because, absent the test, the concern is unresolved but not demonstrated.","tokens_in":29657,"tokens_out":4047,"duration_ms":49091,"concrete_test":"Re-run the mock SFR recovery test splitting the sample by input log sSFR and host-galaxy fraction, with at least 100 simulated quiescent/post-starburst hosts generated using an independent stellar population synthesis code (e.g., BC03) and a different dust attenuation law from that used in the fitting. If the median recovered log sSFR for true log sSFR < −10.5 is biased by more than 0.3 dex, or if the recovered HδA for post-starburst templates is biased by more than 1 Å, the 83%/23% fractions are not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that most extended quasars are hosted by quiescent galaxies (83%) and a large fraction are post-starburst (23%)—depends directly on the sSFR values assigned by the spectrophotometric decomposition. The mock validation in Section 4 is the only quantitative support for these SFRs, yet the SFR recovery test (Figure 9) shows only N=8 'confident recoveries.' The paper does not report recovery statistics for the low-sSFR regime (log sSFR < −10.5) where the vast majority of the real sample lies, and many quiescent hosts are in the upper-limit region (Figure 12 shaded region). Because the mocks are generated with the same Bagpipes non-parametric SFH family (and Galsim) used for fitting, a systematic mismatch between the true host SEDs and the adopted templates—e.g., in dust attenuation, nebular emission, or SFH shape—would not be detected. If the pipeline biases sSFRs systematically low for weakly star-forming hosts, the 83%/23% fractions could be substantially inflated. The HδA measurement, which defines the post-starburst criterion, is also not tested on mock spectra in the presence of AGN contamination. This is the most load-bearing weakness because all headline quantitative conclusions inherit it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a spectrophotometric decomposition technique that jointly fits DESI spectra and Subaru HSC images for 1126 type 1 quasars at 0.14 Å and log sSFR<−10.94), and that the host stellar masses and the MBH–M⋆ relation are broadly consistent with local scaling relations. The method iterates between an AGN spectral model (power-law continuum, Fe pseudo-continuum, Balmer continuum, Gaussian emission lines) and a Bagpipes non-parametric SFH stellar population model, using GalfitM photometric decomposition as a prior. Mock tests are used to argue that the decomposition recovers AGN luminosities, host SFRs, and stellar velocity dispersions with acceptable fidelity.","tokens_in":29946,"tokens_out":2839,"duration_ms":34876,"significance":"If the central claims hold, the paper would add an important data point to the quasar-host co-evolution debate: a sizable population of low-redshift quasars with quiescent and post-starburst hosts, complementing the well-studied compact quasars whose hosts are star-forming. The sample size (1126 objects) is substantial for spectroscopically decomposed quasar hosts, and the joint use of imaging priors with spectral fitting is a genuinely useful methodological direction. The paper is also honest in acknowledging that the morphological selection pre-selects a particular host population. However, the quantitative headline fractions (83% quiescent, 23% post-starburst) are only as reliable as the SFR calibration, and the current validation is not strong enough to support them at face value. The MBH–M⋆ analysis is more robust and less dependent on the disputed SFR values.","major_comments":[{"comment":"The SFR recovery test that underpins the quiescent/post-starburst classification contains only N=8 'confident recoveries,' and the mocks are generated with Bagpipes (plus Galsim) and then fitted with Bagpipes. This demonstrates template self-consistency, not external accuracy. The real sample is concentrated at log sSFR<−10.5 (Fig. 12), but no recovery statistics are shown for that low-sSFR regime. If the pipeline systematically biases weak hosts toward even lower sSFR, the 83% and 23% fractions would be inflated. I recommend adding an independent calibration—e.g., mock galaxies from a different SPS code (or with a different dust/nebular prescription) and reporting recovery rates as a function of input sSFR, including upper-limit behavior.","section":"Sec. 4, Fig. 9"},{"comment":"The 83% quiescent fraction appears to include objects whose sSFRs are upper limits: Fig. 12 explicitly states that points in the shaded region 'only indicate upper limits,' and this shaded region covers the low-sSFR area where the quiescent population lies. Treating upper limits as detections can artificially boost the quiescent fraction. The paper should report the quiescent fraction both with and without upper limits, or use a survival-analysis estimator, and state how many of the 1126 objects are non-detections in SFR.","section":"Sec. 5.2/5.4, Fig. 12"},{"comment":"The sample is selected to have 'non-psf' (extended) morphologies in the Legacy Survey and to be bright (λLλ(5100 Å)>10^43.3 erg s−1). As the paper itself notes, extended morphology is correlated with large effective radius, high stellar mass, and quiescence. The comparison with previous compact-quasar studies is therefore not apples-to-apples: the claim that extended quasar hosts are 'quite different' from compact quasar hosts conflates selection with a physical difference. To make the comparison meaningful, the authors should at least match the two populations in redshift, AGN luminosity, and host stellar mass, or explicitly frame the result as characterizing a pre-selected morphological class rather than as evidence for diversity among quasars in general.","section":"Sec. 2.3/5.4"},{"comment":"The post-starburst criterion (HδA>4 Å and log sSFR<−10.94) is measured on decomposed host spectra, but no mock test is presented for HδA recovery in the presence of residual AGN contamination. The Lick index is sensitive to continuum placement and to any residual AGN power-law or emission-line leakage. Given that the 23% post-starburst fraction is one of the two central quantitative claims, the paper should validate HδA recovery on mocks with injected AGN at the actual host-fraction and S/N distributions of the sample, and show how the fraction changes under reasonable continuum-placement uncertainties.","section":"Sec. 5.3, Fig. 15"},{"comment":"The MBH–M⋆ slope recovery test is constructed by generating mock galaxies that follow the local relation and then injecting the same systematic uncertainties (σsys,M⋆=0.2 dex, σsys,MBH=0.3 dex) that are used in the ODR fit. This is not an independent validation of the fitted slope; it shows that a steep input relation can survive these uncertainties, not that the data themselves require a steep relation. The paper should also present the best-fit slope without the adopted systematic floors, or a sensitivity analysis over a range of floor values, so the reader can see how much of the 'consistency with local relations' is built into the uncertainty model.","section":"Sec. 6.1, Fig. 17"}],"minor_comments":[{"comment":"The phrase 'extended objects' should be defined precisely at first use; it is only fully explained in Sec. 2.3. Also 'Mgii' and 'Caii' are typeset inconsistently (Mgii, Caii, CaI, Mgi) throughout; standard 'Mg II'/'Ca II' would be clearer.","section":"Abstract/Introduction"},{"comment":"The sentence 'its spectra were reduced with pipeline from Guy et al. (2023)' should be 'with the pipeline from Guy et al. (2023).' Similarly, 'The Signal-to-Noise ratios (S/N) of the spectra are calculated as flux divided by the flux error (converted from ivar), which is 1σ value' is awkward and should be rephrased.","section":"Sec. 2.1"},{"comment":"The 20% systematic error floor is motivated, but it is not stated whether the 20% is applied in quadrature to the statistical errors or as a multiplicative fractional error in the likelihood. This should be clarified for reproducibility.","section":"Sec. 3.1"},{"comment":"The text says 'Statistical analysis for the decomposition results of the second mock dataset is displayed in Appendix A (Figure B1)' but the figure appears in Appendix B. The cross-reference should be corrected.","section":"Sec. 4/Appendix B"},{"comment":"The color bar in the upper panel is labeled 'AGN Density' while the panel shows sSFR on the y-axis; the color scale presumably encodes black hole mass (as stated in the caption), so the label is confusing and should be corrected.","section":"Fig. 15"},{"comment":"The virial mass calibration uses a=6.91 from Ho & Kim (2015). Since the host galaxy decomposition changes the continuum luminosity, the authors should state whether the systematic uncertainty in the virial factor is propagated into the MBH errors used in the scaling-relation fits. This is relevant to the claimed 1σ agreement with local relations.","section":"Sec. 5.1"}],"recommendation":"major_revision","confidential_remarks":"The paper has a promising data set and a thoughtful joint decomposition methodology, but the headline fractions (83% quiescent, 23% post-starburst) currently rest on a model-internal mock calibration and on treating upper limits as detections. The authors themselves acknowledge the selection effect, which is good, but the abstract overstates the contrast with compact quasars. I would like to see the SFR validation strengthened (independent templates or at least low-sSFR recovery statistics) and the upper-limit treatment clarified before the quantitative claims are accepted. The MBH–M⋆ part is more solid and could be published even if the SFR classification is softened."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this paper gives us the first large systematic look at hosts of extended quasars at z<1, and the central descriptive claim—that most of these hosts are quiescent, with a large post-starburst minority—is probably real for the population as selected. The selection and the SFR calibration, however, are weaker than the headline fractions suggest, so the numbers should be treated as provisional until the pipeline is validated against independent SED/emission-line constraints.\n\nWhat's new: the sample itself. Over a thousand type 1 quasars with resolved hosts, analyzed with a sensible joint imaging+spectral decomposition. The authors use HSC photometry as a prior on the host flux in DESI spectra, iterate to convergence, and show plausible example decompositions and stacked spectra with strong Balmer absorption. That is a solid methodological step, and the result that extended-host quasars are not on the star-forming main sequence is a real contrast with the compact-quasar studies. They also do a reasonable job of checking the AGN continuum recovery with mocks and with the L5100–LHalpha relation.\n\nWhere I'd be cautious. First, the 83% quiescent and 23% post-starburst fractions rest on sSFR values that are partly upper limits, and the SFR recovery test has eight 'confident recoveries'—all generated with Bagpipes and fitted with Bagpipes. That validates internal consistency, not external accuracy. If the true hosts have different dust/SFH shapes, the low-sSFR tail could be biased low. The HδA measurement is not tested on mocks with AGN contamination, which matters for the post-starburst cut. Second, the sample is selected to be extended/non-PSF, which by construction favors large, bulge-dominated, likely quiescent hosts. The authors say this explicitly and frame the work as complementary to compact-quasar studies, so it's not a hidden flaw; but it does mean the 'diversity' and 'negative feedback' language in the discussion goes beyond what the data can support. Third, the MBH–Mstar slope of 0.93 only becomes 'canonical' after adding 0.2/0.3 dex systematic floors and restricting to a dense core; the MBH–sigma relation is non-robust by their own admission. Those scaling-relation claims should be treated as indicative.\n\nWho should read it: anyone studying quasar host demographics or AGN feedback at low redshift. The pipeline is a useful template, though no code or catalog is released, which limits immediate reuse.\n\nPeer review: send it out. A good referee will ask for an independent SFR calibration (e.g., stack by [OII] or far-IR where available), mock tests that inject non-Bagpipes SEDs, and a matched control sample of compact quasars from the same DESI/HSC footprint. The descriptive core should survive; the fractions and the evolutionary claims need the extra work.","headline":"First large systematic look at extended quasar hosts at z<1 finds them mostly quiescent, but the headline fractions rest on selection and model-internal SFR validation.","tokens_in":30618,"tokens_out":3428,"would_cite":true,"duration_ms":31901,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper argues that most quasars with extended optical morphologies at redshifts 0.1–1 are hosted by quiescent, old-stellar-population galaxies—about 83% quiescent and 23% post-starburst—contrary to the star-forming hosts typical of comp","keywords":["AGN host galaxies","Active galactic nuclei","Quasars","Supermassive black holes","Quiescent galaxies","Post-starburst galaxies","Extended quasar morphology","Spectrophotometric decomposition"],"falsifier":"Generate mock extended quasars with a stellar-population library, dust prescription, and star-formation histories different from the one used in the fitting, then run the decomposition; if the recovered quiescent and post-starburst fractions shift by more than the quoted statistical uncertainties, the classification is template-dependent. Observationally, measure star formation in a subset of the same hosts with an indicator that bypasses template fitting—such as resolved H-alpha emission or far-infrared/radio continuum—and compare the resulting fractions.","tokens_in":29497,"feed_emoji":"🔭","tokens_out":7822,"duration_ms":69839,"temperature":0.7,"pith_summary":"At intermediate redshifts (0.1 < z < 1), quasars whose host galaxies are resolved in optical images are often assumed to resemble the compact quasars studied before—actively star-forming. This paper argues the opposite for extended quasars. By fitting images and spectra together, it separates the bright point-like quasar from the surrounding galaxy and finds that roughly 83% of the hosts are quiescent and about 23% are post-starburst, meaning they recently stopped forming stars and now show strong hydrogen absorption. The hosts are massive and old, and the central black holes follow the local black-hole–stellar-mass relation. If right, this shows that low-redshift quasar hosts are more diverse than earlier samples suggested, with extended morphology selecting a later, quenched evolutionary stage.","feed_headline":"83% of extended quasars live in quiescent hosts","feed_subtitle":"Joint image-spectrum fits reveal old stellar populations and a 23% post-starburst share in low-redshift quasars.","key_machinery":"The central mechanism is an iterative spectrophotometric decomposition: five-band optical imaging is fit with a point-source component plus a generalized elliptical surface-brightness profile, and photometry of the residual image gives the host-to-quasar flux ratio in each band. These image-derived ratios are used as priors when fitting the quasar spectrum, which is modeled as an AGN component (broken power-law continuum, blended iron pseudo-continuum, hydrogen emission lines) plus a host stellar component with a star-formation history divided into age bins. Host and AGN are refit alternately until derived parameters change by less than 5%, breaking the usual degeneracy between the featurele","core_discovery":"On the paper's own terms, the discovery is that the host galaxies of extended quasars at 0.1 4 Å, together with log sSFR < -10.94). The hosts are massive—median stellar mass near 10^10.6 solar masses—with old stellar populations and red colors, and the sample is clearly bimodal in star-formation rate, with no substantial population on the star-forming main sequence. Black-hole masses measured from the decomposed quasar light track host stellar mass with a slope broadly consistent with local relations once systematic uncertainties are included, while the","pith_inferences":["Editorial inference: the 83% and 23% figures are conditional on the morphological selection and do not apply to compact quasars; the combined picture suggests a diversity axis tied to host structure rather than only luminosity or redshift.","Editorial inference: the same decomposition procedure could be applied to other imaging-plus-spectroscopy AGN samples to map how the quiescent fraction varies with host size and bulge prominence, with a testable prediction that it rises with effective radius.","Editorial inference: the inferred sequence—star-forming compact hosts giving way to post-starburst extended hosts and then to quiescent hosts—could be tested with spatially resolved stellar-population ages, which should show younger cores or tidal features in the post-starburst systems.","Editorial inference: if real hosts contain dusty or nebular emission not captured by the template family, the quiescent fraction would be overestimated; resolving this requires star-formation indicators that bypass template fitting, such as infrared luminosity, resolved H-alpha, or radio continuum."],"forward_implications":["A previously undercounted population of active black holes resides in passive galaxies, so AGN activity and strong star formation are not necessarily simultaneous in low-redshift quasars.","The 23% post-starburst share implies that a substantial subset of extended quasars is caught shortly after a burst of star formation was shut off, making them direct laboratories for quenching and feedback.","The host stellar masses peak near 10^10.6 solar masses and match the mass function of quiescent galaxies, so selecting extended morphology effectively selects massive, structurally evolved, quenched hosts.","The black-hole–stellar-mass relation in this sample is consistent with local relations once systematic uncertainties are included, arguing against strong redshift evolution in this restricted population.","The absence of a measurable black-hole–velocity-dispersion relation here should not be read as a physical absence, because the same large uncertainties flatten a known input relation in mock tests."],"fun_headline_variants":["Extended quasars hide quiet hosts, not star-forming ones","Post-starburst hosts common among extended quasars","Extended quasars reveal quiescent, old host galaxies","Surprise: extended quasars live in quiet galaxies","Bimodal star formation: extended quasars are not like compact ones"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The quiescent and post-starburst classifications depend on the assumed star-formation-history templates, dust law, and priors used to model the host spectra; the mock tests are generated with the same model family, so they validate internal consistency, not absolute accuracy of the derived star-formation rates.","fun_headline_variants_meta":{"raw":{"variants":["Extended quasars hide quiet hosts, not star-forming ones","Post-starburst hosts common among extended quasars","Extended quasars reveal quiescent, old host galaxies","Surprise: extended quasars live in quiet galaxies","Bimodal star formation: extended quasars are not like compact ones"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000471,"raw_usage":{"total_tokens":2212,"prompt_tokens":810,"completion_tokens":1402,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":554,"completion_tokens_details":{"reasoning_tokens":1320}},"tokens_in":554,"tokens_out":1402,"duration_ms":8413,"temperature":1.0,"reasoning_tokens":1320,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T00:32:04.943131+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Generate mock extended quasars with a stellar-population library, dust prescription, and star-formation histories different from the one used in the fitting, then run the decomposition; if the recovered quiescent and post-starburst fractions shift by more than the quoted statistical uncertainties, the classification is template-dependent. Observationally, measure star formation in a subset of the same hosts with an indicator that bypasses template fitting—such as resolved H-alpha emission or far-infrared/radio continuum—and compare the resulting fractions.","supporting_citations":[],"review_version":1}