{"id":"e2547459-a29d-4464-ab78-c450340af197","arxiv_id":"2505.03972","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"The survey produced a public catalog of 53,579 stars in IC10, including 536 long-period variable candidates, the largest optical LPV sample for this galaxy.","lead":"This paper presents a new optical survey of IC10, the nearest starburst dwarf galaxy, identifying 536 candidate long-period variable stars, mostly evolved red giants and supergiants. The public catalog lets astronomers probe dust production and star formation history in a low-metallicity, starbursting environment.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Recovery statistics in the abstract and Section 5.2.1 are irreconcilable as written; because these numbers are the evidence for a validated detection method, the catalog's headline claim is unsupported until the counts are recomputed and the text corrected.","rationale":"I read the paper as a catalog paper whose central contribution is the INT optical LPV candidate list and a validated detection method. The strongest claim in the abstract is the recovery performance against external catalogs, especially 'we found all the confirmed LPVs that Gaia DR3 detected in IC10' and the 43%, 40%, and 70% recovery fractions. For that claim to hold, the cross-match counts must be self-consistent. They are not: Section 5.2.1's detailed counts (7+11=18 recovered Spitzer variable candidates; 11 of 235 x-AGB = 4.7%) cannot be reconciled with the '110 stars' and '40% x-AGB' statements, and the HST numbers disagree between text and figure caption. These discrepancies sit exactly in the validation layer of the argument, so they are load-bearing. I am not claiming the catalog is wrong; the inconsistency may be a resolvable denominator or wording problem. But as written, a reader cannot verify the central quantitative claims without redoing the cross-match. The L-index threshold concern identified by the reader is real but secondary: it is a standard mirror-Gaussian heuristic with an explicit ~10% contamination estimate, and its failure would change the candidate list size but not the qualitative existence of the catalog. The recovery-statistic conflict, by contrast, directly contradicts the paper's own headline numbers, so I make it the primary concern. I recommend the reader's conditional verdict stands: the paper should be accepted only after the recovery statistics are recomputed, reported in one consistent table, and the abstract is corrected accordingly. This is an internal-consistency issue, not a disagreement with consensus, so no ad hominem or disciplinary judgment is involved.","tokens_in":22810,"tokens_out":9281,"duration_ms":86234,"concrete_test":"Recompute the full cross-match consistency table from the public catalogs. Define the DUSTiNGS footprint within CCD4 and within 2rh; count (a) total DUSTiNGS sources, (b) Spitzer variable sources, (c) x-AGB and less-dusty AGB subsets; then count how many of each appear in the INT photometric catalog and how many are among the 536 INT LPV candidates. Verify whether 110 equals the number of Spitzer variables present in the INT catalog (recovered as sources) and 18 equals the number of those that are INT variables; if so, recompute the abstract percentages using the correct denominator and adjust the text to say 'recovered as sources' versus 'identified as variables.' Repeat for HST/HCV: reconcile the text's 22 variables/13 LPVs with Fig. 21's 34 variables/17 LPVs using the HCV catalog.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—'most comprehensive optical LPV candidate catalog of IC10, with a validated detection method'—rests on recovery statistics against Spitzer/DUSTiNGS, HST, CFHT, and Gaia DR3. Those statistics are not internally consistent. Section 5.2.1 states that Boyer et al. found 235 x-AGB and 22 less-dusty AGB variable candidates in IC10, and that 7 less-dusty AGB and 11 x-AGB variable candidates were recovered among the INT variables in CCD4—i.e., 18 recovered Spitzer variable candidates. Yet the same paragraph says 'Among the 257 variables identified with Spitzer, we recovered 110 stars in our catalog,' and the abstract claims the survey 'identified 43% of the variable stars found with Spitzer' (110/257 = 43%) and 'retrieved 40% of the extremely dusty AGB stars' (as written, 11/235 = 4.7%; a different CCD4-only denominator would need to be stated). The '~68% less dusty AGB' figure conflicts with 7/22 = 32%. The HST recovery numbers also differ between the text (669 SFVCs; 22 variables and 13 LPVs) and the Fig. 21 caption (34 variables and 17 LPVs). These are not cosmetic typos in an ancillary table: the abstract's quantitative success claims are precisely the evidence that the detection method is validated. If the 18-count is the correct number of Spitzer variables detected as variable by INT, the abstract overstates completeness by a large factor; if the 110-count instead refers to Spitzer sources recovered as non-variable objects, the text must say so and the 'identified 43%' wording must be revised. Until one consistent set of counts is provided, the validation argument cannot be checked and the central claim is not supported as written.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents an optical monitoring survey of the starburst dwarf galaxy IC10 using the Isaac Newton Telescope Wide Field Camera, producing a photometric catalog of 53,579 stars in CCD4 (0.07 deg^2) and identifying 536 long-period variable (LPV) candidates, of which 380 lie within two half-light radii. The authors describe their PSF photometry, variability selection based on Stetson J/K/L indices, extinction corrections, and foreground decontamination using Gaia DR3 and TRILEGAL. They validate the catalog by cross-matching with Pan-STARRS, Spitzer/DUSTiNGS, HST/HCV, CFHT carbon stars, and Gaia DR3 LPVs, and they claim recovery fractions such as ~73% of Spitzer sources, 43% of Spitzer variables, 40% of extremely dusty AGB stars, and ~70% of HST variables. The catalog is to be released at CDS. The central claim is that this is the most comprehensive optical LPV candidate catalog of IC10 with a validated detection method.","tokens_in":23188,"tokens_out":3479,"duration_ms":34445,"significance":"If the claimed recovery statistics are correct, this is a valuable contribution: it is the deepest optical LPV census of IC10 to date, it reuses and documents a mature photometric pipeline, it provides a public catalog that will enable star-formation history and dust-production studies, and it demonstrates cross-survey consistency against Spitzer, HST, Gaia, and CFHT data. The paper includes careful completeness simulations and foreground modeling, and the data release is a concrete public asset. However, the significance is currently undermined by internal numerical inconsistencies in the very statistics that are used to validate the detection method, so the headline claims cannot be accepted as written.","major_comments":[{"comment":"The recovery statistics against Spitzer are internally inconsistent. Section 5.2.1 states that among the 257 variables identified with Spitzer, the authors recovered 110 stars in their catalog, and that Boyer et al. found 235 x-AGB and 22 less dusty AGB variable candidates, of which 11 x-AGB and 7 less dusty AGB candidates were recovered among the INT variables in CCD4. These two statements imply 18 recovered Spitzer variables, not 110. The abstract's claims that the survey 'identified 43% of the variable stars found with Spitzer' (110/257) and 'retrieved 40% of the extremely dusty AGB stars' are not supported by the per-class counts as written: 11/235 is 4.7%, and 7/22 is 32%, not the ~68% stated in the text. The authors must recompute these numbers, state the exact denominators for each percentage, and clarify whether the 110 count refers to Spitzer sources recovered as catalog objects (not necessarily as variables) or to variables. Because these percentages are the primary evidence for the validated detection method, this discrepancy must be resolved before the paper can be accepted.","section":"Abstract and Section 5.2.1"},{"comment":"The HST/HCV cross-match results differ between the text and the figure caption. The text says the authors recovered 669 SFVCs and 'identified 22 of our variables among the HCV variables, of which 13 are LPV candidates,' while the Figure 21 caption says '34 of our variables were recovered from the HCV, of which 17 are considered LPV candidates.' These are different numbers for the same quantity, and the discrepancy is not explained. Since the abstract's ~70% HST recovery claim depends on the correct count, the authors must reconcile the text and caption and report a single consistent set of HST recovery statistics.","section":"Section 5.2.3 and Figure 21 caption"},{"comment":"The derivation of the L-index threshold rests on the assumption that the mirrored negative side of the L histogram represents the non-variable population and that a Gaussian fit to it isolates variables. The text states that a threshold of L = 1.05 is chosen 'at which at least 90% of the stellar distribution is above the Gaussian fit' and that '~10% of these might be non-variable.' This wording is ambiguous: if 90% of the distribution is above the fit, the contamination fraction is not obviously 10%. The authors should state explicitly how the 10% contamination estimate is obtained from the Gaussian fit and whether the threshold was validated with the artificial-star injections. This point is central because the 536 LPV candidates and all derived completeness fractions depend on this threshold.","section":"Section 4, Figure 7"}],"minor_comments":[{"comment":"The abstract states that the authors 'found all the confirmed LPVs that Gaia DR3 detected in IC10 among our identified LPVs,' but Section 4.3 says that Rimoldini et al. identified 20 LPV candidates and the authors found all of them except one that lies outside the CCD4 area. The abstract should be qualified to state that all Gaia LPVs inside the survey footprint were recovered, or that the four spectroscopically/otherwise confirmed LPVs (Lebzelter et al.) were recovered.","section":"Abstract and Section 4.3"},{"comment":"The sentence 'Our results are consistent with the study by Boyer et al. (2015a,b)' is not supported by the preceding numbers, which show an order-of-magnitude discrepancy between recovered and total x-AGB variables. Please rephrase or provide a quantitative comparison.","section":"Section 5.2.1"},{"comment":"The phrase 'recovering approximately 73% of DUSTiNGS sources within the 2rh area in our catalog' is unclear about the denominator: it should state whether the 73% refers to all DUSTiNGS sources in 2rh, all in CCD4, or all in the full DUSTiNGS footprint, and it should give the matching radius used for cross-identification.","section":"Section 5.2.1"},{"comment":"The catalog description mentions that stars with an asterisk are marked due to quality of astrometric solution, but the text does not explain how the astrometric solution quality was assessed or how many stars are affected. A brief explanation would help users of the public catalog.","section":"Section 3.1 and Table 3"}],"recommendation":"major_revision","confidential_remarks":"The central contribution is a catalog and a set of recovery fractions that validate the detection method. The recovery-statistics inconsistency (Section 5.2.1 vs. the abstract, and Section 5.2.3 vs. Figure 21) is load-bearing but appears correctable by recomputation and rewording; it does not in itself invalidate the data products. I recommend major revision rather than rejection, and I would be willing to re-review the revised manuscript. The self-citations to the group's earlier methodology papers are appropriate, as the pipeline is being reused; however, the authors should ensure that the new recovery claims do not inadvertently overstate completeness relative to what the underlying counts show."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The catalog is genuinely useful: 536 LPV candidates in CCD4, 380 within two half-light radii, deeper than Gaia, with a public photometric catalog and light curves. The reduction follows the group's established INT survey pipeline, with artificial-star completeness tests, careful treatment of a bad-seeing epoch, and cross-matching against Spitzer/DUSTiNGS, HST, CFHT, and Gaia. That is real work and a real community resource.\n\nThe problem is the validation statistics. The abstract says the survey recovered 43% of Spitzer variables and 40% of the x-AGB stars. Section 5.2.1 reports 110 of 257 Spitzer variables recovered (which is 43%), but the same paragraph says 7 less-dusty AGB and 11 x-AGB variable candidates were recovered among the INT variables—only 18 objects. 11 of 235 is 4.7%, not 40%; 7 of 22 is 32%, not 68%. The HST numbers disagree too: the text says 22 variables (13 LPVs) matched, while the Figure 21 caption says 34 variables (17 LPVs). These are not cosmetic typos. The abstract's success claims—and the claim that the detection method is validated—rest on these counts. As written, the validation argument cannot be checked.\n\nA few smaller things: the mirror-Gaussian L-index threshold is standard, and the ~10% contamination estimate is acknowledged, but a robustness test around that assumption would help. The decision to reject the SFD98 map and adopt a constant reddening is argued from the CMD morphology; it is plausible but worth stating more quantitatively. Also, the abstract's \"found all the confirmed LPVs that Gaia DR3 detected\" softens to 19 of 20 in the text (one falls outside the CCD4 area); that is probably fine, but the wording should be precise.\n\nWho is this for: anyone working on AGB mass loss, dust production, or star formation histories in low-metallicity dwarfs, and people comparing optical versus mid-IR variability-selected samples. It deserves a serious referee. I would send it out with a request to recompute and report one consistent set of recovery statistics, then revisit the abstract. Once those numbers agree, the catalog stands on its own.","headline":"Useful new LPV catalog for IC10, but the recovery statistics that back the headline claims are internally inconsistent and need to be fixed before the completeness numbers can be trusted.","tokens_in":23812,"tokens_out":3491,"would_cite":true,"duration_ms":32400,"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":"Nine epochs of optical monitoring identify 536 long-period variable star candidates in IC 10, the nearest starburst dwarf galaxy, and recover every long-period variable that Gaia DR3 confirmed there.","keywords":["long-period variable stars","asymptotic giant branch stars","red supergiants","IC 10","dwarf irregular galaxies","starburst galaxies","stellar variability","optical monitoring survey"],"falsifier":"Rerun the exact same pipeline on the same photometry but with each star's measurement epochs randomly permuted in time: this preserves every noise property of the data while destroying any true variability signal, and the fraction of stars that still pass $L>1.05$, the 0.2-mag amplitude cut, and the epoch cut is a direct measurement of the false-positive rate, which should come out near the claimed 10%. If it comes out much higher, the 536-star catalog and the recovery fractions against Spitzer, HST, and Gaia are inflated. A complementary check is to take a sample of the 536 candidates and measure their periods with an independent, longer-baseline monitoring program: genuine LPVs should show coherent long-period pulsation in phase with the $i$-band amplitudes reported here.","tokens_in":22626,"feed_emoji":"🔭","tokens_out":19093,"duration_ms":160743,"temperature":0.7,"pith_summary":"IC 10, the nearest starburst dwarf galaxy in the Local Group, is a compact laboratory where recent bursts of star formation sit beside an old stellar population, and its cool evolved stars both record that history and supply the galaxy's dust. The paper claims to have produced the most extensive optical catalog of long-period variable (LPV) candidates in IC 10: 53,579 stars measured from nine epochs of Isaac Newton Telescope imaging, of which 536 qualify as LPV candidates and 380 lie within two half-light radii of the galaxy's center. The claim matters because these pulsating giants, asymptotic giant branch stars and red supergiants, are the brightest tracers of star formation over timescales from about 10 million to 10 billion years, and because they are the main producers of the dust that enriches the interstellar medium. The paper's validation is its core argument: roughly 73% of the Spitzer sources in the field are recovered, along with about 43% of the Spitzer-identified variables, about 40% of the extremely dusty AGB variables, about 70% of the Hubble variables, about 75% of the CFHT carbon stars, and every LPV that Gaia DR3 confirmed in IC 10.","feed_headline":"536 variable stars, including every LPV Gaia confirms in IC 10","feed_subtitle":"The pulsing giants are the galaxy's best star-formation clocks and its main dust factories.","key_machinery":"The load-bearing machinery is the Stetson variability index $L$, defined by $L = J \\times (K/0.798) \\times (\\sum_i w_i / w_{\\mathrm{all}})$, where the Welch-Stetson index $J$ accumulates weight-signed products of normalized magnitude residuals from paired $i$-band/$V$-band epochs, $K$ is a kurtosis-type shape statistic ($K=0.9$ for a sinusoid), and the weight ratio scales each star by how many epochs actually detected it. The decision threshold is set empirically: for each $i$-band magnitude bin the negative wing of the $L$ histogram is mirrored onto the positive side, a Gaussian is fitted to represent the non-variable population, and the threshold $L=1.05$ is chosen so that at least 90% of the real distribution lies above that Gaussian, implying roughly 10% contamination. Candidates must additionally pass an amplitude cut of 0.2 mag in $i$, a minimum of six measurements spaced at least 30 days apart (following Javadi et al. 2011a), and a non-negative extinction-corrected color $(V-i)_0 > 0$. This chain, index computation, mirrored-Gaussian threshold, amplitude and epoch cuts, and Gaia-based foreground removal, is what takes the 53,579 measured stars down to the 536 claimed LPV candidates.","core_discovery":"The paper's central claim is that targeted optical monitoring of IC 10, nine epochs between June 2015 and October 2017 in the Sloan $i$-band and Harris $V$-band reduced with PSF-fitting photometry and the Stetson variability indices $J$, $K$, and $L$, yields 536 LPV candidates, the most extensive optical catalog of evolved, pulsating, dust-producing stars in the nearest starburst galaxy. The sample is selected by the threshold $L > 1.05$ set through a mirrored-Gaussian fit to the $L$-index histogram, an $i$-band amplitude above 0.2 mag, at least six usable epochs, and removal of foreground stars identified from Gaia parallaxes and proper motions. The authors present completeness as the test of the method: the catalog recovers roughly 73% of the DUSTiNGS/Spitzer sources in the field, about 43% of the Spitzer variables, about 70% of the Hubble Catalog of Variables sources, about 75% of the CFHT carbon stars, and all of the LPVs that Gaia DR3 confirmed in IC 10. The paper also argues that the standard SFD98 reddening map fails for this galaxy, overestimating extinction and smearing the color-magnitude diagram, and that a constant reddening of $E(B-V)=0.78$ mag with a distance modulus of $\\mu=24.51$ mag from Sanna et al. (2008) is the correct extinction prescription, so the catalog magnitudes are corrected that way.","pith_inferences":["My inference, not the paper's: the true LPV population of IC 10 is probably larger than 536, because the survey itself misses roughly 60% of the extremely dusty AGB variables that Spitzer sees; a completeness-corrected count, rather than the raw number, should be used in any star-formation or dust-budget calculation built on this catalog.","My inference: the mirrored-Gaussian threshold procedure is a portable recipe for sparse-epoch surveys; applying it to fields with independently known variability content, or to Monte Carlo noise realizations, would calibrate the 10% contamination figure and test whether $L=1.05$ remains the right cut for other galaxies in the same monitoring program.","My inference: a period-luminosity follow-up on a longer time baseline would turn these candidates into distance anchors; because published distance moduli for IC 10 spread from roughly 22 to 25 mag, a PL-calibrated distance using these LPVs could help settle the galaxy's distance controversy."],"forward_implications":["The 536-star catalog becomes the population baseline for deriving IC 10's star formation history from the luminosity distribution of asymptotic giant branch stars and red supergiants, the stated next step of the survey series.","The recovery statistics quantify the optical blind spot in dusty starbursts: roughly 60% of the extremely dusty AGB variables seen by Spitzer are missed in the optical, most of them in the crowded, dust-rich center of the galaxy.","Because every Gaia-confirmed LPV in the galaxy appears in the candidate list, the catalog can be treated as magnitude-complete at the bright end, and the public data release lets other groups test variability and star-formation claims without new observing campaigns.","The demonstration that the SFD98 dust map overestimates extinction toward IC 10 shifts the recommended extinction treatment for this galaxy to a constant $E(B-V)=0.78$ mag, which affects any distance or luminosity estimate derived from its stars."],"supporting_citations":[{"why":"Supplies the $J$, $K$, and $L$ variability indices and the newtrial logic that the entire detection method is built on.","marker":"Stetson (1996)"},{"why":"Introduced the weighted $J$ index that pairs contemporaneous two-band measurements, the starting point of the variability statistic.","marker":"Welch & Stetson (1993)"},{"why":"Establishes the six-epoch minimum and the amplitude formula that the LPV candidate definition depends on.","marker":"Javadi et al. (2011a)"},{"why":"Provides the adopted distance modulus of 24.51 mag and constant reddening $E(B-V)=0.78$ mag that the extinction-corrected color-magnitude diagrams rest on.","marker":"Sanna et al. (2008)"},{"why":"The DUSTiNGS/Spitzer survey is the external baseline for the claimed 73% source recovery, 43% variable recovery, and 40% dusty-AGB recovery.","marker":"Boyer et al. (2015a,b)"},{"why":"The Gaia DR3 LPV catalog that lists four confirmed LPVs in IC 10; recovering all of them anchors the survey's completeness at the bright end.","marker":"Lebzelter et al. (2023)"},{"why":"The Gaia DR3 SOS catalog that lists 20 LPV candidates in IC 10, all but one of which (outside the field) are recovered in the INT catalog.","marker":"Rimoldini et al. (2023)"},{"why":"The Hubble Catalog of Variables is the baseline for the claimed ~70% recovery of HST variables in the field.","marker":"Bonanos et al. (2019)"},{"why":"The CFHT carbon star catalog is the baseline for the claimed ~75% carbon star recovery that places the LPVs on the color-magnitude diagram.","marker":"Demers et al. (2004)"},{"why":"The trilegal simulation provides the foreground star counts used to remove contamination and statistically correct the luminosity distribution.","marker":"Girardi et al. (2005)"}],"fun_headline_variants":["All Gaia-confirmed LPVs in IC 10 plus 500 more pulsating giants","Catalog of 536 pulsing giants in IC 10 recovers all Gaia LPVs","IC 10's pulsing giants: 536 stars, all Gaia LPVs recovered","Clocking starburst IC 10 with 536 pulsing red giants","536 LPVs in IC 10: every Gaia-confirmed variable recovered"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole candidate list rests on one statistical assumption: that mirroring the negative part of the $L$-index histogram and fitting a Gaussian to it correctly represents the non-variable population, so that the cut $L>1.05$ really isolates variables with about 10% contamination; if the photometric noise is non-Gaussian, or the mirroring distorts the shape of the non-variable distribution, then the numbers 536 and 380, and all the recovery fractions built on them, change.","fun_headline_variants_meta":{"raw":{"variants":["All Gaia-confirmed LPVs in IC 10 plus 500 more pulsating giants","Catalog of 536 pulsing giants in IC 10 recovers all Gaia LPVs","IC 10's pulsing giants: 536 stars, all Gaia LPVs recovered","Clocking starburst IC 10 with 536 pulsing red giants","536 LPVs in IC 10: every Gaia-confirmed variable recovered"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000947,"raw_usage":{"total_tokens":4187,"prompt_tokens":1231,"completion_tokens":2956,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":847,"completion_tokens_details":{"reasoning_tokens":2846}},"tokens_in":847,"tokens_out":2956,"duration_ms":17758,"temperature":1.0,"reasoning_tokens":2846,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:40:55.608634+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Rerun the exact same pipeline on the same photometry but with each star's measurement epochs randomly permuted in time: this preserves every noise property of the data while destroying any true variability signal, and the fraction of stars that still pass $L>1.05$, the 0.2-mag amplitude cut, and the epoch cut is a direct measurement of the false-positive rate, which should come out near the claimed 10%. If it comes out much higher, the 536-star catalog and the recovery fractions against Spitzer, HST, and Gaia are inflated. A complementary check is to take a sample of the 536 candidates and measure their periods with an independent, longer-baseline monitoring program: genuine LPVs should show coherent long-period pulsation in phase with the $i$-band amplitudes reported here.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduced the weighted $J$ index that pairs contemporaneous two-band measurements, the starting point of the variability statistic."},{"cited_title":"B., et al., 2008, ApJ, 688, L69","cited_arxiv_id":null,"evidence_quote":"Provides the adopted distance modulus of 24.51 mag and constant reddening $E(B-V)=0.78$ mag that the extinction-corrected color-magnitude diagrams rest on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The Gaia DR3 LPV catalog that lists four confirmed LPVs in IC 10; recovering all of them anchors the survey's completeness at the bright end."},{"cited_title":"Z., Yang, M., Sokolovsky, K","cited_arxiv_id":null,"evidence_quote":"The Hubble Catalog of Variables is the baseline for the claimed ~70% recovery of HST variables in the field."},{"cited_title":"& Ables, H., 1965, PASP, 77, 272 de Vaucouleurs, G., 1978, ApJ, 224, 710","cited_arxiv_id":null,"evidence_quote":"The CFHT carbon star catalog is the baseline for the claimed ~75% carbon star recovery that places the LPVs on the color-magnitude diagram."}],"review_version":1}