{"id":"a2cd7c50-e9fa-45d3-9496-f09365484548","arxiv_id":"2608.13259","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Most Galactic Be X-ray binaries are caught by eROSITA at 10^33 to 10^35 erg/s outside outburst, pointing to stable low-level accretion as the usual state.","lead":"Using four eROSITA all-sky scans, this paper maps the faintest X-ray luminosities ever systematically measured for known high-mass X-ray binaries in the Milky Way. The key finding is that most Be-neutron-star binaries are detected at intermediate luminosity between outbursts, implying that low-level accretion is common, not rare.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claim that >80% of BeXRBs exceed Be-star luminosities by an order of magnitude rests on a small 500 pc volume-limited Be-star baseline whose bright end is not robust; the full Be-star sample overlaps the BeXRB regime.","rationale":"The paper is a careful, observationally grounded study: it delivers the first eROSITA logN-logL for known Galactic HMXBs down to about 10^32 erg/s, provides a robust comparison to RXTE/ASM and MAXI, and convincingly separates the luminosity regimes of BeXRBs and SgXBs. The detection of many BeXRBs at 10^33 to 10^34 erg/s in single snapshots is a real result, corroborated for a subset by NuSTAR hard X-ray detections. The stress-test concern focuses on the quantitative inference from this result: the claim that more than 80% of BeXRBs are accreting outside outburst relies on the excess of BeXRB luminosities over the expected Be-star donor emission. That expectation is set by the 500 pc volume-limited Be-star sample, whose observed maximum around 10^32 erg/s is used as the effective floor. Because this sample contains only tens to a few hundred stars, its maximum is not a reliable estimator of the true bright end of the Be-star X-ray luminosity distribution. The paper itself notes the overlap between the total Be-star sample and BeXRBs in the 10^32 to 10^33 erg/s range, yet the headline claim uses the more restrictive volume-limited maximum. A volume-corrected Be-star luminosity function or a bootstrap uncertainty on the observed maximum would directly test whether the order-of-magnitude excess and the associated 80% fraction are robust. If the revised baseline moves to about 10^33 erg/s, the accretion-norm claim would need to be weakened or rephrased as 'accretion is common but the fraction is not constrained to 80%.' This does not invalidate the paper's central measurement, but it does mean the strongest claim is conditional on the Be-star baseline, exactly as the reader's verdict stated. I therefore see no reason to change the CONDITIONAL verdict, and I agree with the reader's weakest-assumption identification.","tokens_in":33972,"tokens_out":9634,"duration_ms":96727,"concrete_test":"Recompute the Be-star baseline using a 1/Vmax volume-corrected luminosity function from the full eROSITA Be-star sample, or bootstrap-resample the 500 pc subsample to estimate the 95% upper bound on the maximum Be-star luminosity. Then count the fraction of eROSITA-detected BeXRBs whose luminosity exceeds 10 times that revised baseline, and also recompute the fraction using the total Be-star sample maximum (shown in Fig. 6) as the baseline. If the fraction falls below about 60%, the 'accretion is the norm' claim needs a quantitative caveat; if it remains above 80%, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Abstract and Sect. 5.1) is that more than 80% of BeXRBs are detected at luminosities at least an order of magnitude above the most X-ray-luminous Be stars. The baseline is the volume-limited 500 pc Be-star subsample of Fig. 6, whose distribution 'ends at about 10^32 erg/s'. This baseline is load-bearing for the accretion inference, but it is statistically fragile. A 500 pc sphere contains only of order 10^2 Be stars from the BeSS sample, so the observed maximum is a noisy estimator of the intrinsic upper end of the Be-star X-ray luminosity function; a rare, luminous Be star could easily be absent. Indeed, the total (non-volume-corrected) Be-star sample in the same figure extends above 10^33 erg/s and overlaps the BeXRB regime, a fact the paper acknowledges. The 'order of magnitude' threshold is thus defined relative to a small-sample maximum, not a robust estimate of the expected bright end. Section 4 explicitly uses the 500 pc subsample 'in lieu of a volume-corrected distribution.' If the true Be-star luminosity function has a tail to about 10^33 erg/s, the appropriate 10x threshold becomes about 10^34 erg/s, and the reported 80% fraction would drop substantially, as most eROSITA BeXB detections cluster around 10^33 to 10^34 erg/s (Fig. 6). This directly weakens the headline statement that accretion outside outburst is likely the norm for a vast majority of BeXRBs. The concern is not that the eROSITA detections are wrong, but that the inferred accretion rate is over-stated because the comparison baseline is not demonstrated to be unbiased.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents eROSITA eRASS1-4 observations of 72 known HMXBs in the Western Galactic hemisphere, measuring 0.2-10 keV fluxes and luminosities down to ~10^32 erg/s, with careful pile-up corrections, fixed/modeled spectral parameters, and upper limits for non-detections. It constructs logN-logL distributions for the full sample and for subclasses (BeXRBs, SgXBs, SFXTs), compares them with variability-aware resamplings of RXTE/ASM and MAXI light curves, and compares BeXRB luminosities with those of isolated Be stars. The paper reports that BeXRBs dominate the low-luminosity end, that combining the four eRASS surveys yields a detection fraction of about 80% for known BeXRBs, and interprets the luminosity excess over isolated Be stars as evidence that low-luminosity accretion outside outburst is likely the norm, discussing cold-disk accretion and the propeller mechanism.","tokens_in":34283,"tokens_out":10712,"duration_ms":108767,"significance":"If the central claim is robust, this is an important result: it is the first wide-area survey characterization of Galactic HMXBs at 10^32-10^34 erg/s, with implications for the HMXB XLF, the quiescent behavior of BeXRBs, and accretion theory at low mass-accretion rates. The paper's strengths include a carefully vetted and reclassified HMXB catalog, explicit treatment of pile-up, cross-instrument consistency checks against RXTE/ASM and MAXI, and the release of a source catalog as auxiliary material. However, the headline accretion conclusion rests on a statistically fragile comparison to a small volume-limited Be-star sample and, for most sources, on luminosity arguments rather than direct spectral or timing diagnostics; the significance is therefore conditional on a more robust statistical treatment of the Be-star baseline.","major_comments":[{"comment":"The load-bearing comparison for the abstract's claim that more than 80% of BeXRBs exceed the most X-ray luminous Be stars by an order of magnitude uses the 500 pc volume-limited Be-star subsample as the baseline. This sample contains only of order 10^2 Be stars, so its observed upper end near 10^32 erg/s is a noisy extreme-order statistic; the non-volume-limited Be-star sample in the same figure extends above 10^33 erg/s and overlaps the BeXRB locus, as the paper itself acknowledges. If the intrinsic Be-star X-ray luminosity function has a tail to about 10^33 erg/s, the 'order of magnitude' threshold becomes about 10^34 erg/s, and most eROSITA BeXRB detections, which cluster at 10^33-10^34 erg/s, would no longer exceed it. The manuscript needs a quantitative treatment of this baseline, for example a bootstrap or order-statistics analysis, or a completeness-corrected Be-star XLF, before the headline claim can be supported.","section":"Sect. 4, Fig. 6"},{"comment":"The central physical conclusion that accretion outside outburst is the norm is an extrapolation from snapshot detections for most sources. The paper states in Sect. 5.2 that eROSITA does not provide enough statistics for detailed spectral or timing studies, and the spectral discrimination between a Be star and a low-luminosity BeXRB (Fig. 7) is available for only a minority of sources. Hard X-ray confirmation, shown as blue points in Fig. 8, covers roughly 20% of the sample (Zalot et al. 2026). For the remaining detections, the luminosity excess over isolated Be stars could also be produced by binary-enhanced donor emission, gamma-Cas-like activity, or magnetospheric emission, alternatives that the paper itself discusses for bright Be stars. The conclusion should be reworded as a hypothesis with supporting evidence, and the fraction of sources with direct accretion signatures should be stated separately from the fraction detected at elevated luminosity.","section":"Sect. 5.2, Fig. 8"},{"comment":"The flux and luminosity calibration carries an acknowledged systematic uncertainty of up to a factor of 10 from geometrical and other effects (Sect. 5.2), which is comparable to the 'order of magnitude' excess that underpins the central claim. In addition, fluxes for sources with fewer than about 50 counts are derived by fixing Gamma=1 or kT=1 keV and N_H at catalog values, and upper limits are sensitive to the assumed absorption. Since many sources lie near the 10^33 erg/s threshold used to define the >80% fraction, the authors should propagate these uncertainties into the reported detection fraction and the order-of-magnitude excess fraction, for example by varying spectral parameters and N_H within plausible ranges.","section":"Sect. 3.1, Sect. 5.2"}],"minor_comments":[{"comment":"The abstract refers to 'the Galactic population' and 'known HMXBs' without noting that the analysis covers only the Western Galactic hemisphere; this limitation should be stated explicitly in the abstract.","section":"Abstract"},{"comment":"The column labeled 'KS-Test probability' reports values such as 19.32 and 6.82, which cannot be probabilities; this is likely the KS statistic or a scaled D value, and the column should be relabeled or converted to p-values.","section":"Table C.1"},{"comment":"The text states that the detection fraction increases to 80% when combining eRASS1:4, while the abstract claims that more than 80% of BeXRBs are detected at luminosities at least an order of magnitude higher than the most X-ray luminous Be stars; these are different statements, and not all sources in Table E.1 have minimum luminosities above 10^33 erg/s, so the wording should be reconciled.","section":"Sect. 5.1, Abstract"},{"comment":"Please clarify whether the red dashed curve for undetected BeXRBs is constructed from flux upper limits and how upper limits enter the N(>L) counting; the current description is ambiguous.","section":"Fig. 6"},{"comment":"The phrase 'in lieu of a volume-corrected distribution' highlights a limitation of the Be-star comparison; at minimum, the discussion should address how the distance and absorption distributions of the BeXRB sample differ from those of the 500 pc Be-star sample, since these differences could bias the inferred luminosity excess.","section":"Sect. 4"}],"recommendation":"major_revision","confidential_remarks":"This is a valuable dataset paper with an interpretive claim that currently outruns its statistical support. The main revision should focus on making the Be-star baseline quantitative and on separating the detection fraction from the fraction with direct evidence of accretion. The paper fits the scope of A&A and the catalog release is a useful community resource. The authors should also ensure that results from companion papers such as Zalot et al. (2026) are clearly available, since they carry part of the hard X-ray evidence for the accretion interpretation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper. First, the central measurement is real and valuable: it is the first wide-area logN-logL for known Galactic HMXBs reaching down to about 10^32 erg/s, three orders of magnitude fainter than what RXTE/ASM or MAXI could do. Second, the headline claim that accretion outside outburst is “likely the norm” for >80% of BeXRBs is plausible but less firmly established than the abstract suggests, because it leans on a small, volume-limited sample of Be stars that is not demonstrably unbiased.\n\nThe paper is genuinely well done. The cross-matching, pile-up corrections, and upper-limit treatment are careful. The consistency check against RXTE/ASM and MAXI, using random sampling of the monitoring lightcurves to match eROSITA's snapshot nature, is a thoughtful and convincing way to validate the bright end. The subclass-resolved distributions, with SgXBs populating the high-luminosity end and BeXRBs the low end, are a useful step beyond earlier averaged XLFs. The catalog release is a plus.\n\nThe soft spot is exactly where the stress-test note lands. The “order of magnitude above the most X-ray luminous Be stars” baseline is the 500 pc volume-limited Be-star subsample, which contains only a few hundred stars and ends near 10^32 erg/s. The full (non-volume-corrected) Be-star sample extends above 10^33 erg/s and overlaps the faint BeXRBs. If the true Be-star luminosity function has a tail to 10^33 erg/s, then the 80% detection fraction at luminosities “10x above the Be-star max” would drop. The authors do acknowledge the overlap and use spectral hardness as a discriminator, and they have some NuSTAR follow-up, but the luminosity-only excess is still a load-bearing part of their argument. They should either quantify the uncertainty in the Be-star maximum (e.g., via a luminosity function model or a bootstrap) or soften the specific 80%/10x claim.\n\nOther concerns are minor: the sample is half the sky, distances for some systems are poorly constrained, and the faint-source fluxes assume fixed spectral parameters. None of this breaks the measurement.\n\nBottom line: this paper deserves a serious referee. The logN-logL extension and the snapshot-detection statistics are important and likely correct. The accretion interpretation is probably right, but the Be-star baseline comparison needs strengthening before the headline claim is published as stated.\n\nRecommendation: send it to peer review, but ask the authors to make the Be-star comparison robust and to restrain the abstract until they do.","headline":"A careful eROSITA population study that extends the HMXB luminosity function to 10^32 erg/s and makes a credible but slightly over-stated case that most BeXRBs accrete outside outburst; the measurement is solid, the Be-star baseline needs work.","tokens_in":35019,"tokens_out":2835,"would_cite":true,"duration_ms":29819,"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":"Most Be X-ray binaries are accreting even when not in outburst","keywords":["high-mass X-ray binaries","Be X-ray binaries","luminosity function","logN-logL","accretion","propeller effect","eROSITA","X-ray surveys"],"falsifier":"A decisive test would be a volume-complete comparison: select BeXRBs and isolated Be stars to the same distance limit (say, 3 kpc using Gaia parallaxes), correct both samples for absorption, and compare their hard 2–10 keV luminosity distributions; if the BeXRB excess over the isolated Be-star distribution disappears, the accretion-outside-outburst conclusion would be undermined. A complementary check is to monitor the few BeXRBs with stringent upper limits below about $10^{32}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$ (such as AX J1700.2−4220 and Ginga 0834−430) for months to years with a sensitive instrument: if any of them is caught in a bright outburst after a long quiet period, the propeller picture would have to accommodate a genuine 'off' state, whereas if they are always detected at or above $10^{32}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$ whenever sufficient exposure is accumulated, the inferred luminosity floor would be confirmed.","tokens_in":33729,"feed_emoji":"🔭","tokens_out":19287,"duration_ms":160825,"temperature":0.7,"pith_summary":"The paper uses the first four all-sky surveys of the eROSITA X-ray telescope to measure the X-ray luminosities of every known high-mass X-ray binary in the Western Galactic hemisphere down to about $10^{32}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$, the deepest wide-area census of these objects yet made. At these luminosities the sample is dominated by Be X-ray binaries — systems in which a neutron star orbits a Be-type star with a circumstellar disk. The central finding is that the vast majority, more than 80 percent once the four surveys are combined, are detected in a random snapshot at luminosities of $10^{33}$ to $10^{35}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$, at least an order of magnitude above what their Be star donors alone emit. The authors interpret this excess as ongoing accretion onto the neutron star outside of outburst, meaning the long-standing picture of BeXRBs as 'off' between rare outbursts is wrong for most systems, and establishing a de facto luminosity floor for accreting BeXRBs around $10^{32}$ to $10^{33}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$. The result matters because the low-luminosity end of the Milky Way's HMXB luminosity function was previously inaccessible, and it suggests that some X-ray-luminous 'isolated' Be stars may in fact be hidden accreting binaries.","feed_headline":"Most Be X-ray binaries are accreting even when not in outburst","feed_subtitle":"Faintest census: over 80% of Be X-ray binaries outshine their donors outside outburst, so accretion is the norm.","key_machinery":"The argument is carried by a comparison of two logN-logL distributions: the BeXRBs detected by eROSITA versus the eROSITA-detected Be stars drawn from a catalog of 832 Gaia quality-cut stars in the Western hemisphere (170 detected), where a 500 pc volume-limited subsample stands in lieu of a volume-corrected distribution. The physical interpretation is anchored to two analytic thresholds: the cold disk accretion luminosity, $L_{\\rm cold} \\sim 9\\times10^{33}\\,k^{1.5} B_{12}^{0.86} M_{1.4}^{0.28} R_6^{1.57}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$, below which a poorly ionized disk can still sustain accretion, and the propeller limit, $L_{\\rm prop} \\sim 4\\times10^{37}\\, k^{7/2} B_{12}^{2} P^{-7/3} M_{1.4}^{-2/3} R_6^{5}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$, below which the rotating magnetosphere is expected to inhibit accretion; plotting each system's spin period against its eROSITA luminosity places most BeXRBs between these two limits. On the data side, the load-bearing elements are the four stacked eROSITA surveys (eRASS1–4), the updated HMXB catalog of 72 Western-hemisphere systems, and fluxes from absorbed power-law and blackbody fits corrected for pile-up, complete to about $10^{-13}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}\\,\\mathrm{cm}^{-2}$.","core_discovery":"The paper's central claim is that the low-luminosity regime of the Galactic HMXB population belongs to the Be X-ray binaries, and that these systems are, in a random snapshot of the sky, most often found accreting rather than quiescent. Using the eRASS1–4 data, the authors extend the HMXB logN-logL from around $10^{35}$ down to $10^{32}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$, and show that the distribution requires a broken power law in this range, with the lower-luminosity slope dominated by BeXRBs and the bright end by supergiant systems. Comparing the BeXRB distribution with the eROSITA-detected isolated Be star population — including a volume-limited 500 pc subsample used as a donor baseline — over 80 percent of the BeXRBs lie at luminosities at least an order of magnitude above what their Be donors would produce, in a regime consistent with cold disk accretion onto the neutron star. The four eROSITA scans also reveal a large scatter in the overall logN-logL due to source-intrinsic variability, which the variability-resampled RXTE/ASM and MAXI distributions above $10^{35}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$ confirm is consistent with previous monitoring, while below that level the monitors suffer incompleteness. A small number of bright, hard X-ray 'isolated' Be stars reach luminosities comparable to the faintest BeXRBs, which the authors propose as candidate quiescent or low-luminosity accreting systems.","pith_inferences":["If accretion outside outburst is the norm, the transient/quiescent distinction for BeXRBs is largely an observational bias of monitoring instruments; population synthesis and binary evolution models should treat most BeXRBs as quasi-persistent low-luminosity accretors with outburst duty cycles superimposed, which would change estimates of the integrated accretion luminosity of the Galaxy.","The $10^{32}$–$10^{33}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$ overlap between the most luminous isolated Be stars and the faintest BeXRBs suggests luminosity alone cannot identify hidden compact companions; the discriminating observables are hard-band spectral shape (power-law tail versus purely thermal) and timing (pulsations), a testable prediction for follow-up campaigns with sensitive hard X-ray inst","Correlating eROSITA luminosities with optical measurements of the Be star's circumstellar disk state across the eRASS1–4 sample would test the cold-disk interpretation: if X-rays persist when the disk is depleted, the low-luminosity accretion must be fed by a different reservoir than the classic decretion disk."],"forward_implications":["The Milky Way HMXB luminosity function is time-dependent: the four eROSITA scans differ substantially in shape and break luminosity depending on which sources are flaring, so single-epoch XLFs come with large intrinsic scatter that must be propagated when comparing with external galaxies.","All-sky monitors that only reach $\\gtrsim 10^{35}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$ (RXTE/ASM, MAXI) are strongly biased toward supergiant HMXBs; the dominant BeXRB population is essentially invisible to them outside outburst, which explains the longstanding 'transient' classification.","The propeller mechanism is constrained to at most ~20% of the BeXRB population: most systems are detected in the cold-disk luminosity band, and only a handful of short-spin systems sit near or below the propeller limit with stringent upper limits.","With more than 80% of BeXRBs detected in eRASS1:4 at $10^{33}{-}10^{35}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$, the eROSITA data indicate that stable low-luminosity accretion rather than quiescence is the typical state of these binaries.","The logN-logL of the known sample requires a broken power law with the low-luminosity part dominated by BeXRBs, so extrapolating a single power law from outburst-censored samples would misrepresent both the shape and the subclass composition of the Milky Way HMXB population."],"supporting_citations":[{"why":"Supplies the XRBcats HMXB catalog (source list, absorption, spin and orbital periods) from which the Western-hemisphere sample is built.","marker":"Neumann et al. 2023"},{"why":"Provides the complementary HMXB census used to revise the catalog classification of sources.","marker":"Fortin et al. 2023"},{"why":"Defines the eRASS1 catalog, detection likelihoods, and survey sensitivity that the cross-matching and completeness rely on.","marker":"Merloni et al. 2024"},{"why":"Establishes the method for constructing the isolated Be-star logN-logL that the BeXRB excess is measured against.","marker":"Nazé & Robrade 2023"},{"why":"Defines the cold disk accretion threshold $L_{\\rm cold}$ used to interpret the low-luminosity detections.","marker":"Tsygankov et al. 2017a"},{"why":"Introduces the propeller mechanism whose centrifugal inhibition sets the expected 'off' state luminosity limit.","marker":"Illarionov & Sunyaev 1975"},{"why":"Provides the RXTE/ASM-based Galactic HMXB logN-logS that the eROSITA distribution is compared to and extended beyond.","marker":"Grimm et al. 2002"},{"why":"Gives the previous variability-aware treatment of the ASM luminosity function that the authors reproduce and compare with eROSITA.","marker":"Islam & Paul 2016"},{"why":"Reports the hard X-ray INTEGRAL luminosity distribution with a break that the eROSITA broken-power-law result is consistent with.","marker":"Lutovinov et al. 2013"},{"why":"Predicted the eROSITA HMXB yield and the low-luminosity XLF shape that this survey now tests.","marker":"Doroshenko et al. 2014"}],"fun_headline_variants":["Most Be X-ray binaries accrete even outside outburst","Faint census reveals Be X-ray binaries usually accreting","eROSITA shows over 80% of BeXRBs accreting at low luminosity","Be X-ray binaries: accretion is the norm, even when quiet","Faintest HMXB survey: BeXRBs outshine donors outside outburst"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument that BeXRBs are accreting outside outburst assumes that the luminosity distribution of isolated Be stars, especially the 500 pc volume-limited subsample, is an unbiased baseline for the X-ray output of the Be donor inside a binary, so if binary interactions brighten the donor or the BeXRB distances and absorbing columns differ systematically, the order-of-magnitude excess attributed to accretion could be inflated.","fun_headline_variants_meta":{"raw":{"variants":["Most Be X-ray binaries accrete even outside outburst","Faint census reveals Be X-ray binaries usually accreting","eROSITA shows over 80% of BeXRBs accreting at low luminosity","Be X-ray binaries: accretion is the norm, even when quiet","Faintest HMXB survey: BeXRBs outshine donors outside outburst"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000438,"raw_usage":{"total_tokens":2382,"prompt_tokens":1256,"completion_tokens":1126,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":872,"completion_tokens_details":{"reasoning_tokens":1031}},"tokens_in":872,"tokens_out":1126,"duration_ms":8422,"temperature":1.0,"reasoning_tokens":1031,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:12:13.652767+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be a volume-complete comparison: select BeXRBs and isolated Be stars to the same distance limit (say, 3 kpc using Gaia parallaxes), correct both samples for absorption, and compare their hard 2–10 keV luminosity distributions; if the BeXRB excess over the isolated Be-star distribution disappears, the accretion-outside-outburst conclusion would be undermined. A complementary check is to monitor the few BeXRBs with stringent upper limits below about $10^{32}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$ (such as AX J1700.2−4220 and Ginga 0834−430) for months to years with a sensitive instrument: if any of them is caught in a bright outburst after a long quiet period, the propeller picture would have to accommodate a genuine 'off' state, whereas if they are always detected at or above $10^{32}\\,\\mathrm{erg}\\,\\mathrm{s}^{-1}$ whenever sufficient exposure is accumulated, the inferred luminosity floor would be confirmed.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports the hard X-ray INTEGRAL luminosity distribution with a break that the eROSITA broken-power-law result is consistent with."}],"review_version":1}