{"id":"4b31e420-5b88-4885-a72d-d481bf792e7c","arxiv_id":"2411.16006","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Radio galaxies split cleanly by hydrogen-alpha equivalent width, and only the most highly excited ones show Eddington ratios above 0.01, but this depends on a model-based bolometric calibration.","lead":"This paper sorts 16,803 radio galaxies from SDSS into 'optically active' and 'optically inactive' groups using the strength of their hydrogen-alpha emission, then compares their masses, star formation, and radio power. It also builds a new model-based way to estimate black-hole power from optical lines, leading to the claim that only the most highly excited radio galaxies are efficiently accreting.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central Eddington-ratio claim hinges entirely on the unvalidated Appendix B bolometric calibration; absent independent X-ray/mid-IR checks, it is not established.","rationale":"The reader's weakest_assumption is the Appendix B bolometric calibration, and I agree that this is the single most load-bearing link in the argument. The paper's central claim is that only VHERGs have Eddington ratios above 0.01, and this is determined by dividing Lbol from Eq. B.4 by LEdd. The model-dependent conversion (SEDs, dust, mixing, covering factor) sets the zero-point and slope of the Lbol-L(H-alpha) relation, and the paper provides no independent check. The paper itself flags this uncertainty in the abstract and conclusion, but the claim is nevertheless stated as a finding; the responsible verdict is conditional on external validation. I do not see a more fundamental flaw: the OPARG/OPIRG classification is physically motivated, the comparative statistics are carefully corrected for Malmquist bias, and the VHERG identification is an interesting, falsifiable hypothesis. The absence of code/data and the in-preparation ROGUE II catalogue are secondary. The concrete test described above would settle whether the Eddington-ratio dichotomy survives an independent bolometric calibration; until then, the reader's CONDITIONAL verdict stands.","tokens_in":32641,"tokens_out":3714,"duration_ms":37962,"concrete_test":"Cross-match the OPARG sample with WISE mid-IR detections (W3/W4) and X-ray catalogs (2RXS, XMM slew, or Swift BAT for the brightest subset). For the matched objects, estimate Lbol independently from the 12-micron luminosity (e.g., Spinoglio et al. 2024, or a standard Lbol ~ 10 L_12 relation) and from absorption-corrected 2-10 keV luminosity using bolometric corrections of 20-30, then compare these with the Appendix B values (Eq. B.4 divided by 0.65). If the median offset exceeds ~0.5 dex, or if the fraction of non-VHERG OPARGs with lambda > 0.01 changes materially, the headline claim is not robust to the calibration.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline result ('only VHERGs have Eddington ratios higher than 10^-2') is explicitly conditional on the bolometric calibration of Appendix B, as the paper itself concedes in Section 10 ('If our estimates of the bolometric luminosities are correct...'). The calibration converts BPT position and extinction-corrected L(H-alpha) into Lbol via Eq. B.4, combined with a covering factor of 0.65 (Section 7.1). Every Eddington ratio in Section 7.2 inherits this conversion, so any systematic offset in Eq. B.4 propagates directly into the claimed radiative-efficiency dichotomy. The surfaces for eta and Lmod_bol/L(H-alpha) are fitted to Cloudy model grids that assume a constant density of 10^3 cm^-3, a fixed set of SEDs, a specific dust/depletion treatment, and a mixing prescription between AGN and H II region models; the paper provides no external validation against independent bolometric indicators (X-ray, mid-IR) and no propagated uncertainties. Since the VHERG region (log[O III]/H-beta >= 0.8) sits at the high-Lbol end of the fitted surface, the conclusion that only this small BPT-top subgroup is radiatively efficient could be at least partly an artifact of the model surface rather than a genuine property of the sources. The reader's weakest_assumption identifies exactly this fragility, and the paper's own caveat makes it load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses the ROGUE I and II radio catalogues matched to the SDSS main galaxy sample to define two classes of radio galaxies: optically inactive radio galaxies (OPIRGs, W(H-alpha)<3 A) and optically active radio galaxies (OPARGs, W(H-alpha)>=3 A). After applying Vmax completeness corrections, the authors compare stellar masses, black hole masses, radio luminosities, stellar ages, and dust properties of the two classes, and place them in the context of the full SDSS galaxy population. They then identify a subgroup of OPARGs at the top of the AGN wing of the BPT diagram, call them very-high-excitation radio galaxies (VHERGs), and compute Eddington ratios using a new photoionization-model-based bolometric calibration described in Appendix B. The central claim is that only VHERGs have Eddington ratios above 10^-2, so that only a small fraction of canonical HERGs are radiatively efficient.","tokens_in":33004,"tokens_out":6593,"duration_ms":67394,"significance":"If the bolometric calibration is correct, the result would challenge the standard HERG/LERG division as a proxy for radiative-mode versus jet-mode accretion and would identify the radiatively efficient population with a small BPT-selected subgroup rather than with the whole high-excitation class. The strengths of the paper include the large, visually classified ROGUE sample, the use of Vmax corrections for a magnitude-limited and flux-limited sample, the MaNGA-based check of contamination in low-W(H-alpha) galaxies, and the explicit presentation of the fitted polynomial formulae for the bolometric correction. The weakness is that the headline conclusion is conditional on an unvalidated model-dependent calibration, and the paper itself concedes in Section 10 that the conclusion holds only if the bolometric-luminosity estimates are correct. With independent validation and sensitivity tests, the result would be an important contribution; as it stands, the significance is high but provisional.","major_comments":[{"comment":"The central claim that only VHERGs have Eddington ratios above 10^-2 rests entirely on the calibration of Eq. (B.4), which maps the BPT coordinates to log Lmod_bol/LH-alpha, combined with the adopted covering factor of 0.65 in Section 7.1. The paper provides no external validation of this calibration against independent bolometric indicators such as X-ray or mid-infrared luminosities, and no propagated uncertainties from the model grid (SED choice, density, dust/depletion, abundances, mixing prescription, covering factor) into the Eddington ratios reported in Section 7.2 and Table 1. Because the VHERG region is located at the high-excitation end of the fitted surface, a systematic error in that part of the surface would directly change the inferred fraction of radiatively efficient radio galaxies. I request an external validation for at least a subsample and a sensitivity analysis that varies the most important model parameters, especially the covering factor and the treatment of dust.","section":"Appendix B, Eqs. (B.3)-(B.4); Section 7.1"},{"comment":"The VHERG class is defined by the condition log [OIII]/H-beta >= 0.8, while Eq. (B.4) assigns log Lmod_bol/LH-alpha as an increasing function of y = log [OIII]/H-beta (positive coefficients in y and y^3). The conclusion that VHERGs preferentially have high Eddington ratios is therefore partly built into the calibration: selecting objects on the same coordinate that enters the bolometric correction tends to select objects with a larger assigned Lbol for a given H-alpha luminosity. This is not a purely circular argument, because the Eddington ratio also depends on the extinction-corrected H-alpha luminosity and on MBH, but the effect should be quantified. I request a robustness test in which the Eddington ratios are recomputed with a BPT-independent bolometric correction (e.g., Lbol = 600 L[OIII] or Lbol = 3500 L[OIII]) and with an independent bolometric indicator, to show how much of the VHERG excess survives.","section":"Section 7.2; Appendix B, Eq. (B.4)"},{"comment":"Several quantitative claims are made on the basis of visual inspection rather than statistical tests. In particular, the statement that the radio-luminosity distributions of OPARGs and OPIRGs are 'undistinguishable' (Section 4.2 and the abstract) is not supported by any two-sample test; with 16,803 objects, even small distribution differences can be highly significant, and percentile overlap is not a substitute for a Kolmogorov-Smirnov or Anderson-Darling test. Similarly, the claimed displacement of the VHERG subgroup to the left of the main AGN wing in Section 6 is not quantified. These tests are needed to establish the secondary claims and to support the interpretation that the VHERG location is special.","section":"Section 4.2; Section 6; Figures 5-11"},{"comment":"The abstract and Section 10 state that 'Only very-high excitation radio galaxies (VHERGs) have Eddington ratios higher than 10^-2', but Section 7.2 states that 'Almost all radio galaxies with Eddington ratios lambda larger than 0.01 are at the top right of the BPT diagram.' These statements are not equivalent. The paper should quantify the fraction of objects with lambda > 0.01 that fall inside and outside the VHERG region defined by log [OIII]/H-beta >= 0.8, and align the wording of the abstract and conclusions with the actual numbers. This is directly relevant to the headline claim.","section":"Abstract; Section 7.2; Section 10"}],"minor_comments":[{"comment":"The text contains incomplete citations, e.g., '?Best & Heckman 2012; ?' in the discussion of the DLM diagram; these should be completed.","section":"Section 2.3"},{"comment":"The notation 'Lmod_bol/LH-alpha' is ambiguous: it is not clear from the text whether the ratio is taken with respect to the AGN H-alpha luminosity or the total H-alpha luminosity after mixing with H II regions. Since the fitted formula is applied to observed total H-alpha luminosities, this point must be clarified explicitly.","section":"Appendix B, Eq. (B.4) and Figure B.2"},{"comment":"The H II region models assume a solar-metallicity stellar population even when the nebular oxygen abundance is sub-solar or super-solar, and the mixing prescription assumes the same O/H for the AGN and H II regions. These choices should be justified and their impact on the fitted surfaces discussed.","section":"Appendix B, Section B.2"},{"comment":"The black hole masses are derived from the Tremaine et al. (2002) relation using starlight stellar velocity dispersions; the paper does not discuss possible systematics from fibre-aperture effects or from the use of a different MBH-sigma relation. A brief statement of the expected systematic uncertainty would be useful.","section":"Section 7.2"},{"comment":"The caption contains a typo: 'OPIGRs' should be 'OPIRGs'. There are also scattered typographical issues such as 'di fferent' and 'Objets' that should be corrected during editing.","section":"Figure 19 caption"},{"comment":"The statement that HOLMES do not strongly affect the bolometric luminosity is plausible but is argued qualitatively; a quantitative estimate of the HOLMES contribution for objects near W(H-alpha)=3 A would be more convincing.","section":"Appendix B, Section B.4"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely to be of interest to the AGN community, but the central Eddington-ratio result is currently conditional on a calibration that is not independently validated. The authors should be encouraged to make the Cloudy model grid and the fitting code publicly available, since the polynomial coefficients alone are not sufficient for full reproducibility. The incomplete references and the fact that ROGUE II is listed as '2024 in preparation' should be checked by the editor during revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth reading for the OPARG/OPIRG classification and the population comparisons, but the central claim that only VHERGs are radiatively efficient is not established. It rests entirely on the Appendix B bolometric calibration, which is model-dependent and never validated against independent bolometric indicators.\n\nWhat is genuinely new and good: the W(Halpha) >= 3 A threshold is physically motivated by the HOLMES cutoff, and they justify it with MaNGA (about 4% of low-W(Halpha) galaxies show weak AGN evidence). The sample is large (16,803 radio AGNs) and the Vmax corrections are standard and carefully applied. The result that radio luminosity distributions of optically active and inactive radio galaxies are indistinguishable, while masses and stellar ages differ, is solid and interesting. The VHERG subpopulation at the top of the AGN wing is also a real observational find: these objects have harder ionizing fields, high HeII/Hbeta, and high equivalent widths. The Appendix B method is clearly described and provides useful formulae, even if the coefficients are not final.\n\nThe soft spot is the load-bearing bolometric correction. Because the polynomial in Eq. B.4 gives systematically higher Lbol/LHalpha toward the high-excitation BPT region, the statement that only VHERGs exceed lambda = 0.01 is partly built into the calibration. The paper's own 'if our estimates are correct' caveat in the abstract and conclusions is honest but understates the problem. There is no external check against X-ray or mid-IR bolometric luminosities, no propagated uncertainties, and the covering factor of 0.65 is a single number with no sensitivity analysis. The constant density of 10^3 cm^-3, the SED set, and the HII/AGN mixing prescription are all reasonable but untested choices. Minor issues: the 'undistinguishable' radio luminosity distributions are not tested for significance, and the unresolved reference placeholders and missing code/data limit reproducibility.\n\nWho is this for? AGN and radio-galaxy researchers. The classification is a useful alternative to the loose HERG/LERG definitions, and the population-level comparisons deserve engagement. The Eddington-ratio claim should be treated as a hypothesis, not a result, until the calibration is validated.\n\nSend it to peer review, but require a much stronger validation section: compare with independent bolometric indicators for a subset, propagate uncertainties, and show that the qualitative conclusion survives plausible changes in covering factor, density, and SED. If the authors can do that, this becomes a strong paper. As it stands, the headline is not yet supported.","headline":"Useful W(Halpha)-based radio-galaxy classification and population comparisons, but the headline Eddington-ratio claim rests on an unvalidated, BPT-dependent bolometric calibration that needs serious scrutiny.","tokens_in":33528,"tokens_out":2747,"would_cite":true,"duration_ms":27002,"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":"The paper argues that only very-high-excitation radio galaxies have Eddington ratios above 0.01, which would mean the standard radiative-mode versus jet-mode division of radio AGN needs revision.","keywords":["radio galaxies","active galactic nuclei","Eddington ratio","BPT diagram","bolometric luminosity","photoionization models","SDSS","accretion modes"],"falsifier":"Measure bolometric luminosities of a sample of OPARGs with log([O iii]/Hβ)<0.8 using independent tracers such as X-ray or mid-infrared luminosity; if a substantial fraction of these objects have Eddington ratios above $10^{-2}$, the claim that only VHERGs are radiatively efficient is falsified.","tokens_in":32449,"feed_emoji":"📡","tokens_out":10818,"duration_ms":88331,"temperature":0.7,"pith_summary":"The paper proposes a physically motivated split of radio galaxies into optically inactive (OPIRGs, H$\\alpha$ equivalent width below 3 Å) and optically active (OPARGs, above 3 Å) classes, using 16,803 radio-loud active galactic nuclei (AGN) from the ROGUE catalogues in the SDSS main galaxy sample. It then devises a photoionization-based method to estimate bolometric luminosities from optical spectra, subtracting the contribution of young stars to the emission lines. With those luminosities, it finds that only a small sub-group of OPARGs at the top of the BPT diagram (the standard AGN emission-line diagnostic), the very high-excitation radio galaxies (VHERGs), have Eddington ratios above $10^{-2}$. The paper concludes that the commonly used HERG/LERG classification overstates the size of the radiatively efficient population and that the radiative-mode versus jet-mode picture of radio galaxies needs revision.","feed_headline":"Only a small slice of radio galaxies is radiatively efficient","feed_subtitle":"A new bolometric correction from SDSS spectra leaves only a BPT-top subgroup above the 0.01 Eddington threshold.","key_machinery":"The central machinery is a calibration that converts an optical spectrum into an AGN bolometric luminosity. A grid of photoionization models for AGN and H II regions is mixed in varying proportions; two polynomial surfaces, one giving the AGN fraction of H$\\alpha$ and one giving $L_{\\mathrm{bol}}/L_{\\mathrm{H}\\alpha}$ as functions of log([N ii]/H$\\alpha$) and log([O iii]/H$\\beta$), are fitted to the model grid. Applying the surfaces to BPT position, dividing by a covering factor of 0.65, and combining with black hole masses from velocity dispersions yields the Eddington ratios that drive the paper's conclusion. Simpler regressions from $L_{[\\mathrm{O\\,iii}]}$ to $L_{\\mathrm{bol}}$ are also provided.","core_discovery":"Using the W(H$\\alpha$)$\\ge$3 Å threshold to define optically active radio galaxies, the paper finds 2,721 OPARGs and 14,082 OPIRGs. After Malmquist correction, the radio luminosity distributions of the two classes are indistinguishable, while OPIRGs host more massive black holes and stellar masses, and OPARGs show recent star formation. Placing OPARGs on the BPT diagram reveals a distinct sub-family at the top of the AGN wing, slightly left of the main AGN sequence, with the highest [O iii]/[O ii], He ii/H$\\beta$, H$\\alpha$ luminosity and equivalent width, indicating a harder ionizing field and higher ionization parameter. The paper's bolometric-luminosity method, which mixes AGN and H II region photoionization models according to BPT position, yields Eddington ratios that exceed $10^{-2}$ only for these VHERGs, defined by log([O iii]/H$\\beta$)$\\ge$0.8. Thus most canonical HERGs fall below the threshold generally taken to mark radiatively efficient accretion, and the radiatively efficient radio-loud population is a small, high-excitation subset.","pith_inferences":["If the VHERG threshold is stable, it could serve as a cheap single-diagnostic selector for radiatively efficient radio AGN in surveys without full spectral modelling.","The same bolometric-correction machinery could be tested against X-ray or mid-infrared AGN luminosities; disagreement would reveal which model ingredient dominates.","The paper's result suggests that the canonical $10^{-2}$ Eddington threshold, if correct, should be applied to VHERG-like objects only, so previous demographic studies of AGN accretion modes may need re-binning.","Since VHERGs have low mechanical-to-radiative output, their jets may be produced by a different mechanism than those of OPIRGs, a prediction that high-resolution radio observations could test."],"forward_implications":["The HERG class as usually defined contains many objects accreting below $10^{-2}$ Eddington, so surveys that select HERGs by W[O iii]$>$5 Å are not selecting radiatively efficient AGN.","Radio luminosity alone cannot distinguish accretion modes: OPARGs and OPIRGs have indistinguishable $L_{1.4}$ distributions.","True radiatively efficient radio-loud AGN are concentrated in a small BPT-top region, so studies of quasar-mode feedback should target VHERGs.","The new $L_{\\mathrm{bol}}$–$L_{[\\mathrm{O\\,iii}]}$ relations give lower bolometric luminosities than several earlier corrections, shifting Eddington-ratio estimates for type II AGN downward.","OPARGs with recent star formation and active nuclei support cold-gas fuelling of radio AGN, tying the optically active class to gas-rich galaxies."],"supporting_citations":[{"why":"Provides the ROGUE I/II radio-source catalogues from which the 16,803 radio AGN sample is drawn.","marker":"Kozieł-Wierzbowska et al. 2020"},{"why":"Supplies the DLM diagram criterion used to separate radio AGN from star-forming radio sources.","marker":"Kozieł-Wierzbowska et al. 2021"},{"why":"Shows that galaxies with W(Hα)<3 Å are ionized by hot old stars rather than AGN, the physical basis of the OPIRG/OPARG split.","marker":"Cid Fernandes et al. 2011"},{"why":"Provides the observed AGN SEDs used as input to the photoionization model grid for bolometric corrections.","marker":"Ferland et al. 2020"},{"why":"Justifies the covering factor 0.65 used to scale model bolometric luminosities.","marker":"Stalevski et al. 2016"},{"why":"Defines the canonical Eddington-ratio threshold $10^{-2}$ for radiatively efficient accretion and the previous bolometric corrections being challenged.","marker":"Heckman & Best 2014"},{"why":"The main earlier HERG/LERG sample and Eddington-scaled accretion-rate results that the paper compares and revises.","marker":"Best & Heckman 2012"},{"why":"Provides the popular $L_{\\mathrm{bol}}=600\\,L_{[\\mathrm{O\\,iii}]}$ bolometric correction that the paper shows overestimates $L_{\\mathrm{bol}}$.","marker":"Kauffmann & Heckman 2009"}],"fun_headline_variants":["Most radio galaxies are not radiatively efficient","Only extreme-excitation radio AGN shine efficiently","New bolometric method shrinks radiatively efficient AGN club","Radio galaxies: radiative efficiency reserved for the top BPT wing","VHERGs only: Eddington ratio above 0.01 in radio AGN"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the model grid that turns measured line strengths into bolometric luminosity, especially the assumed covering factor of 0.65; if the true covering factor or input SEDs are systematically different, every Eddington ratio and the conclusion about which radio galaxies are radiatively efficient shifts accordingly.","fun_headline_variants_meta":{"raw":{"variants":["Most radio galaxies are not radiatively efficient","Only extreme-excitation radio AGN shine efficiently","New bolometric method shrinks radiatively efficient AGN club","Radio galaxies: radiative efficiency reserved for the top BPT wing","VHERGs only: Eddington ratio above 0.01 in radio AGN"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000471,"raw_usage":{"total_tokens":2461,"prompt_tokens":1181,"completion_tokens":1280,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":797,"completion_tokens_details":{"reasoning_tokens":1193}},"tokens_in":797,"tokens_out":1280,"duration_ms":9806,"temperature":1.0,"reasoning_tokens":1193,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:38:55.684786+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure bolometric luminosities of a sample of OPARGs with log([O iii]/Hβ)<0.8 using independent tracers such as X-ray or mid-infrared luminosity; if a substantial fraction of these objects have Eddington ratios above $10^{-2}$, the claim that only VHERGs are radiatively efficient is falsified.","supporting_citations":[{"cited_title":"2020, , 247, 53","cited_arxiv_id":null,"evidence_quote":"Provides the ROGUE I/II radio-source catalogues from which the 16,803 radio AGN sample is drawn."},{"cited_title":"2021, , 910, 64","cited_arxiv_id":null,"evidence_quote":"Supplies the DLM diagram criterion used to separate radio AGN from star-forming radio sources."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the canonical Eddington-ratio threshold $10^{-2}$ for radiatively efficient accretion and the previous bolometric corrections being challenged."},{"cited_title":"& Heckman , T","cited_arxiv_id":null,"evidence_quote":"Provides the popular $L_{\\mathrm{bol}}=600\\,L_{[\\mathrm{O\\,iii}]}$ bolometric correction that the paper shows overestimates $L_{\\mathrm{bol}}$."}],"review_version":1}