{"id":"85cf6f1b-9163-4fb3-be18-1211cf5e1b38","arxiv_id":"2411.17851","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Line splitting in Cepheids X Sgr and BG Cru is periodic on timescales of 12 and 3 days, ruling out pulsation-induced shocks and suggesting non-radial modes.","lead":"Using high-resolution spectra of classical Cepheids, the authors show that the line splitting seen in X Sagittarii and BG Crucis repeats with periods of about 12 and 3 days, not with the pulsation period. This rules out pulsation shocks as the cause and points to non-radial oscillations, with similar humps found in four more Cepheids.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"For X Sgr, the claimed 12.317 d line-splitting period is consistent with the combination frequency f_X - f0; the paper's Table 4 and Sec. 3.5 leave the independent-periodicity identification unresolved, so the abstract's periodicity-based shock-ruling-out argument is not yet secure.","rationale":"The paper is a competent observational study, and much of its evidence is independent of the X Sgr frequency ambiguity. The direct comparison of CCFs at the same pulsation phase in consecutive cycles (Fig. 3) shows that the splitting is not phase-locked to the 7.01 d pulsation. BG Cru's hump periodicity coincides with f_Z, a signal independently detected in RV, FWHM, and BIS (Sec. 3.3), and the period is not a simple combination with the first-overtone frequency. These facts prevent the concern from becoming fatal. However, the central abstract claim is explicitly phrased around 'the periodicity of line splitting' differing from the pulsation period. For X Sgr, the only direct measurement of that periodicity is the hump-tracing time series, and the paper itself flags in Sec. 3.5 that the 0.081 d^-1 signal could be the combination f_X - f0. Because combination frequencies can have amplitudes comparable to or larger than their parent modes (the paper cites Benko and Kovacs 2023), the observed 12.317 d power does not by itself establish an independent period. The numerical agreement between the FAMIAS 0.08123 d^-1 detection and f_X - f0 makes the combination hypothesis quantitatively plausible. A prewhitening test that removes f0, its harmonics, and f_X and checks for residual 0.0819 d^-1 power would settle the issue directly. If the residual power disappears, the X Sgr-specific wording should be revised to rely on Fig. 3 and on BG Cru; if it persists across independent instrument subsets, the current claim is strengthened. I therefore see no basis to change the reader's CONDITIONAL verdict, but the X Sgr frequency identification should be resolved before the abstract's periodicity-based conclusion is taken as established.","tokens_in":18659,"tokens_out":11842,"duration_ms":105489,"concrete_test":"Prewhiten the Coralie14 CCF time series of X Sgr with a simultaneous multi-frequency model containing f0, its harmonics, and f_X = 0.22382 d^-1 (amplitudes/phases fitted from the RV series), then search for residual power near 0.0819 d^-1 in both the 1D mean FAMIAS spectrum and the hump-position time series. If no significant residual peak remains, the 12.317 d period is a combination/beat artifact rather than an independent line-splitting periodicity; if it persists and is also seen when Hermes, Coralie07, and Coralie14 subsets are analyzed separately, the independent-periodicity interpretation is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing point is the identification of the X Sgr hump periodicity. Tracing the hump (Sec. 3.3) gives f_X,Sgr = 0.08190(3) d^-1, P = 12.317 d. The FAMIAS 1D mean spectrum (Table 4) detects 0.08123 d^-1, which is essentially f_X - f0 = 0.22382 - 0.14260 = 0.08122 d^-1; the hump-tracing value differs by ~0.0007 d^-1, about 2-3 frequency-resolution elements over the 10-year baseline. Section 3.5 explicitly states it is not clear which signal is the independent periodicity, and Table 4 lists f_X - f0 as a possible interpretation of the 0.081 d^-1 signal. The abstract's claim that the line-splitting periodicity differs from the pulsation period and hence rules out pulsation-induced shocks depends on treating this 12.317 d period as an independent physical periodicity. If it is instead a beat between the pulsation f0 and the 4.47 d signal f_X, the X Sgr 'periodicity differs' argument is weakened to the cycle-to-cycle phase comparison in Fig. 3, which is still evidence against phase-locked shocks. BG Cru is not affected: its hump periodicity matches the independently detected f_Z in RV/FWHM/BIS. Thus the X Sgr leg of the central claim needs an explicit combination-frequency test.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses high-resolution CCF time series from the VELOCE project to study line splitting in the classical Cepheids X Sgr and BG Cru. It develops three approaches: tracing the position of the hump in the CCF residual after a single-Gaussian fit, fitting triple-Gaussian components, and computing 2D/1D Fourier spectra with FAMIAS. The authors report a 12.317 d periodicity for the X Sgr hump and a 3.008 d periodicity for the BG Cru hump, both differing from the respective pulsation periods. They conclude that pulsation-induced shocks are ruled out and that non-radial modes are the most likely explanation, also reporting unresolved splitting humps in six additional Cepheids for an incidence rate of about 3% in the VELOCE sample.","tokens_in":18937,"tokens_out":10533,"duration_ms":88198,"significance":"If the conclusions hold, this work is significant for the interpretation of line-profile variability in classical Cepheids: it challenges the common assumption that line splitting in these stars is caused by pulsation-driven shocks and provides new evidence for additional low-amplitude phenomena, likely non-radial modes. The analysis benefits from high signal-to-noise CCFs, a long (multi-year) baseline, and a large sample of 258 Cepheids. A notable strength is the internal cross-checking of periodicities across different diagnostics (RV, FWHM, BIS, contrast, EW, hump tracing, triple-Gaussian fits, and FAMIAS), and for BG Cru the hump periodicity matches an independently detected signal fZ in multiple indicators. The identification of seven additional Cepheids with hump-like CCF distortions is a useful observational resource. However, the X Sgr frequency identification has an unresolved ambiguity that needs to be addressed before the specific 12.317 d periodicity can be quoted as the physical periodicity of the line splitting.","major_comments":[{"comment":"The claim that X Sgr's line-splitting hump has a distinct 12.317 d periodicity is not yet secure because the manuscript explicitly states it is unclear which signal is the independent periodicity. Table 4 shows that the 0.081 d^-1 signal detected in the FAMIAS 1D mean spectra is consistent with the combination frequency fX - f0 = 0.08122 d^-1, and the hump-tracing value 0.08190(3) d^-1 differs from this by about 2-3 frequency-resolution elements over the ~10-year baseline. Since the abstract and Sec. 4.3 use this periodicity to argue against pulsation-induced shocks and against rotation, the authors should perform an explicit combination-frequency test: fit a model containing f0 and its harmonics together with fX, fX-f0, and fX+f0 to the hump time series and to the RV residuals, then examine the residual periodogram at 0.0819 d^-1 and quantify the significance of any remaining signal. Until this test is done, the conclusions should be phrased with the ambiguity stated in Sec. 3.5 rather than asserting the 12.317 d period as the independent line-splitting periodicity.","section":"Sec. 3.5 and Table 4"}],"minor_comments":[{"comment":"The sample size is inconsistently reported as 258 in Sec. 3.1 and the abstract but as 285 in Sec. 4.2. The incidence rate of 3% should be computed with the correct denominator.","section":"Sec. 3.1 and Sec. 4.2"},{"comment":"The rotational velocity estimates contain numerical errors. For X Sgr with R=53 R_sun, a rotation period of 4.47 d gives v_eq ≈ 600 km/s (not 96 km/s), and the hump period of 12.317 d gives ≈ 218 km/s (not 35.8 km/s); for BG Cru with R=41 R_sun and P=3.008 d, v_eq ≈ 690 km/s (not 110 km/s). These corrections strengthen the argument against rotation, but the quoted values and the undefined symbol f1 should be corrected.","section":"Sec. 4.3"},{"comment":"The caption \"Frequency spectra for the time-series of RV of the hump in LPV\" appears to contain a typo; LPV should likely read CCF or hump.","section":"Fig. 10 caption"},{"comment":"The additional signals found from the triple-Gaussian component time series are all below the 5σ detection level. The text should clearly state that these results are tentative and only corroborate the other methods, which it does in part, but the phrasing \"revealed\" is too strong for 3σ detections.","section":"Sec. 3.4"},{"comment":"The visual classification of humps in six additional stars is subjective; a quantitative criterion (e.g., a threshold on the residual amplitude after Gaussian subtraction) would improve reproducibility.","section":"Sec. 2.3"},{"comment":"There is a typo: \"unkown\" should be \"unknown\".","section":"Sec. 4.2"}],"recommendation":"major_revision","confidential_remarks":"The paper is well written and the observational analysis is careful, but the frequency identification for X Sgr is genuinely ambiguous and is central to the quantitative claim in the abstract. The authors themselves acknowledge the unresolved choice between interpretations A and B in Sec. 3.5. A combination-frequency test is feasible and would settle whether 0.0819 d^-1 is independent; without it, the 12.317 d period should not be presented as the established line-splitting periodicity. The rest of the analysis, especially for BG Cru, is convincing. I recommend major revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis is a useful paper and worth reading carefully. The VELOCE dataset lets the authors do something that was not possible before: track the line-splitting hump in X Sgr and BG Cru across many cycles and measure its periodicity. The 12.3 d and 3.0 d periods are new, the demonstration that CCFs at the same pulsation phase differ between consecutive cycles (Fig. 3) is direct and convincing, and the candidate list of six additional stars with a 3% incidence rate is a concrete survey result. The paper is honest about its own ambiguities, which I appreciate.\n\nThe soft spots are real but not fatal. For X Sgr, the hump frequency fX_Sgr = 0.08190 d^-1 is suspiciously close to fX - f0 = 0.08122 d^-1, and the FAMIAS 1D spectrum (Table 4) lists that combination interpretation with a value that matches to about 2-3 resolution elements. The paper explicitly says it is not clear which signal is independent. The abstract's claim that the splitting period differs from the pulsation period, and hence rules out shocks, depends on the 12.3 d period being physically independent. If it is a combination, the argument weakens to the Fig. 3 phase comparison, which is still good evidence against phase-locked shocks but is not the same as an independent periodicity. BG Cru is not affected: the 3.0 d hump period matches the independently detected fZ in RV/FWHM/BIS.\n\nAlso, the six additional candidates come from visual screening without a quantitative false-positive analysis, so the 3% incidence figure should be treated as provisional. The triple-Gaussian fits only reach 3 sigma for most additional signals, which the authors acknowledge.\n\nIf I were refereeing this, I would ask for an explicit combination-frequency test for X Sgr—e.g., check whether fX_Sgr can be explained as fX - f0 with frequency resolution taken into account, or search for the hypothetical parent mode directly. I would also want a more rigorous detection criterion for the candidate humps, or at least a statement of how many stars were borderline and rejected. None of this would change the core observational result that line splitting in these stars is not phase-locked to pulsation, but the survey-wide conclusion needs firmer footing.\n\nWho is this for? Cepheid pulsation people and anyone working on non-radial modes in classical pulsators. The data release (radial velocities, shape indicators, and CCFs for X Sgr, BG Cru, and SZ Cas) is a reproducible contribution in itself.\n\nRecommendation: send it to a serious referee. It needs revision, not rejection.","headline":"A solid, honest VELOCE study of Cepheid line splitting whose central shock-ruling-out claim leans on an unresolved frequency identification for X Sgr, but whose cycle-to-cycle phase comparison stands on its own.","tokens_in":19595,"tokens_out":1014,"would_cite":true,"duration_ms":10567,"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":"In two classical Cepheids, X Sgr and BG Cru, spectral line splitting recurs on periods of 12.317 d and 3.008 d—not their pulsation periods—ruling out pulsation-induced shocks and pointing to non-radial modes.","keywords":["classical Cepheids","line splitting","cross-correlation function","non-radial modes","pulsation","VELOCE","X Sagittarii","BG Crucis"],"falsifier":"A dense, alias-free spectroscopic campaign on X Sgr tracking the hump position over several seasons could settle whether the $0.08190\\,\\mathrm{d}^{-1}$ signal is an independent frequency or a combination of the 7.0-day pulsation with another frequency; if the signal dissolves into combination frequencies, the independent-periodicity premise fails and the shock-ruling-out conclusion weakens, whereas a stable, independent peak would support the non-radial mode interpretation.","tokens_in":18355,"feed_emoji":"⭐","tokens_out":12858,"duration_ms":95732,"temperature":0.7,"pith_summary":"The paper aims to establish that line splitting in classical Cepheids is not caused by pulsation-driven shock waves, as previously assumed for X Sagittarii and BG Crucis. Using high signal-to-noise cross-correlation function (CCF) time series from the VELOCE project, it shows that the splitting hump in each star repeats on a periodicity clearly different from the dominant pulsation period—12.317 d versus 7.01 d for X Sgr, and 3.008 d versus about 3.34 d for BG Cru—and that CCFs at identical pulsation phases differ between consecutive cycles. The paper concludes that non-radial modes are the most likely explanation, and reports six additional Cepheids with unresolved splitting humps, putting the incidence at about 3% of the VELOCE sample. If correct, this removes shocks from the accepted picture of Cepheid atmospheres and adds a rare type of line-profile variability to look for in surveys.","feed_headline":"Splitting of Cepheid spectral lines rules out pulsation shocks","feed_subtitle":"Split lines in two Cepheids follow 12.3-day and 3.0-day cycles, not their pulsation periods—pointing to non-radial modes.","key_machinery":"The central object is the cross-correlation function (CCF) hump and the time series built from it. Each CCF is fitted with a single Gaussian to define the star's radial velocity, and the hump created by line splitting is located as the maximum of the residuals; the hump's velocity relative to the Gaussian center is then analyzed with standard Fourier techniques, converting a qualitative profile distortion into a measurable periodic signal. For X Sgr that signal is $0.08190\\,\\mathrm{d}^{-1}$ and for BG Cru $0.33243\\,\\mathrm{d}^{-1}$. Supporting machinery includes a triple-Gaussian decomposition of the split profile into blue, middle, and red components (following Mathias et al. 2006) and two-dimensional FAMIAS Fourier spectra, which confirm the same frequency in the hump position, in CCF shape indicators (FWHM, BIS, contrast, EW), and in the depths and centroids of the fitted components.","core_discovery":"The central discovery is that the recurring line-splitting hump in X Sgr and BG Cru is tied to periods that differ significantly from the stars' dominant pulsation periods: in X Sgr the hump repeats with $P = 12.317(4)$ d (frequency $0.08190\\,\\mathrm{d}^{-1}$) against a 7.01 d pulsation, and in BG Cru with $P = 3.00813(9)$ d against a first-overtone period near 3.34 d. Because pulsation-induced shocks would lock splitting to specific pulsation phases and repeat every cycle, the observed cycle-to-cycle differences at the same pulsation phase (shown for X Sgr in Fig. 3) rule out the shock interpretation. The paper concludes that non-radial modes are the most likely origin, notes that the periods are too short for rotation, and, by visually inspecting all 258 VELOCE Cepheids, finds humps in six further stars (LR TrA, V0411 Lac, V1334 Cyg, SZ Cas, V1019 Cas, ASAS J174603-3528.1), giving an incidence of about 3%. The hump stars tend to have broader CCFs, lower radial-velocity amplitudes, and lower CCF contrast than typical Cepheids.","pith_inferences":["A testable extension of the non-radial mode interpretation: mapping the hump's phase and amplitude across lines formed at different atmospheric depths would constrain the horizontal and vertical structure of the suspected mode, since the paper finds the splitting strongest in weak metallic lines and absent in Balmer lines.","The paper's 3% incidence rests on visual classification; an automated screen using FWHM, BIS, and contrast outliers across a larger Cepheid sample would turn this into a statistical measurement and test whether the association with broad CCFs is a detection bias or a physical correlation.","If the 12.317-day signal in X Sgr were shown to be a combination frequency rather than an independent periodicity, the shock-ruling-out argument would lose its main pillar for that star; the cycle-to-cycle CCF differences shown in Fig. 3 would remain as secondary evidence against phase-locked shocks.","If the 3.008-day signal in BG Cru is indeed a non-radial mode, then BG Cru hosts at least two independent additional modes beyond its radial first overtone, making it a promising testbed for mode-interaction and mode-identification studies that go beyond the present paper."],"forward_implications":["The shock-wave interpretation for line splitting in X Sgr and BG Cru is ruled out; the splitting recurs on periods that are not the pulsation period and CCFs at the same pulsation phase differ between cycles.","The splitting hump in each of the two stars carries a measurable periodicity that also appears in CCF shape indicators, meaning the phenomenon is detectable in standard radial-velocity and profile-shape diagnostics.","Among 258 VELOCE Cepheids, 8 show line splitting or humps (about 3%), and these stars preferentially have high average FWHM, low radial-velocity amplitude, and low CCF contrast.","The splitting is much more visible in weak metallic lines than in strong lines and is absent from Balmer lines, indicating that the phenomenon is confined to a narrow range of atmospheric depths.","For X Sgr, the 4.47-day signal seen in shape indicators is close to the combination of the pulsation frequency with the 12.317-day hump period, suggesting that the two periodicities trace the same underlying feature through different diagnostics."],"supporting_citations":[{"why":"Provides the VELOCE CCF time series, radial velocities, and instrument groupings that all of the periodicity and shape analyses in this work are based on.","marker":"Anderson et al. 2024"},{"why":"Advanced the shock-wave explanation for X Sgr line splitting and supplied the triple-Gaussian modeling approach that the paper adapts and then rules out.","marker":"Mathias et al. 2006"},{"why":"Reported line splitting in X Sgr, BG Cru, EV Sct, and V1334 Cyg and first proposed non-radial modes as the origin.","marker":"Kovtyukh et al. 2003"},{"why":"Published the BG Cru Coralie14 data and the CCF shape-indicator analysis that detected the fZ signal now identified with the hump periodicity.","marker":"Netzel et al. 2024"},{"why":"Detected the fX and fY signals in BRITE photometry of X Sgr, supporting the reality of the additional periodicities found in this paper.","marker":"Smolec et al. 2018"},{"why":"Reported the CCF line splitting in X Sgr, BG Cru, and LR TrA and showed that CCFs at the same pulsation phase differ between cycles.","marker":"Anderson 2013"},{"why":"Provides the interpretation of 0.61 Cepheid signals as harmonics of high-degree non-radial modes, used to contextualize X Sgr's and BG Cru's additional frequencies.","marker":"Dziembowski 2016"}],"fun_headline_variants":["Cepheid line splitting cycles defy pulsation periods, hinting at non-radial modes","Split lines in Cepheids cycle at odd periods, ruling out shock-driven splitting","Cepheid line splitting shows non-radial modes, not pulsation shocks","Cepheid humps beat at separate periods, pointing to non-radial modes","Line-splitting Cepheids reveal cycles that rule out shock waves"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument rests on the claim that the 12.317-day hump periodicity in X Sgr is an independent physical periodicity rather than a combination or beat with the 7.0-day pulsation frequency; the paper itself flags this ambiguity in its Section 3.5, so if the 12.317-day signal turned out to be a combination, the conclusion that the splitting period differs from the pulsation period would be weakened.","fun_headline_variants_meta":{"raw":{"variants":["Cepheid line splitting cycles defy pulsation periods, hinting at non-radial modes","Split lines in Cepheids cycle at odd periods, ruling out shock-driven splitting","Cepheid line splitting shows non-radial modes, not pulsation shocks","Cepheid humps beat at separate periods, pointing to non-radial modes","Line-splitting Cepheids reveal cycles that rule out shock waves"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000588,"raw_usage":{"total_tokens":2872,"prompt_tokens":1171,"completion_tokens":1701,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":787,"completion_tokens_details":{"reasoning_tokens":1596}},"tokens_in":787,"tokens_out":1701,"duration_ms":10251,"temperature":1.0,"reasoning_tokens":1596,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:46:28.428359+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A dense, alias-free spectroscopic campaign on X Sgr tracking the hump position over several seasons could settle whether the $0.08190\\,\\mathrm{d}^{-1}$ signal is an independent frequency or a combination of the 7.0-day pulsation with another frequency; if the signal dissolves into combination frequencies, the independent-periodicity premise fails and the shock-ruling-out conclusion weakens, whereas a stable, independent peak would support the non-radial mode interpretation.","supporting_citations":[{"cited_title":"VELOcities of CEpheids (VELOCE) I. High-precision radial velocities of Cepheids","cited_arxiv_id":"2404.12280","evidence_quote":"Provides the VELOCE CCF time series, radial velocities, and instrument groupings that all of the periodicity and shape analyses in this work are based on."},{"cited_title":"B., et al","cited_arxiv_id":null,"evidence_quote":"Advanced the shock-wave explanation for X Sgr line splitting and supplied the triple-Gaussian modeling approach that the paper adapts and then rules out."},{"cited_title":"V ., Andrievsky, S","cited_arxiv_id":null,"evidence_quote":"Reported line splitting in X Sgr, BG Cru, EV Sct, and V1334 Cyg and first proposed non-radial modes as the origin."},{"cited_title":"The VELOCE Modulation Zoo I. Spectroscopic detection of non-radial modes in the first-overtone Cepheids BG Crucis, QZ Normae, V0391 Normae, and V0411 Lacertae","cited_arxiv_id":"2403.13796","evidence_quote":"Published the BG Cru Coralie14 data and the CCF shape-indicator analysis that detected the fZ signal now identified with the hump periodicity."},{"cited_title":"R., Moffat, A","cited_arxiv_id":null,"evidence_quote":"Detected the fX and fY signals in BRITE photometry of X Sgr, supporting the reality of the additional periodicities found in this paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reported the CCF line splitting in X Sgr, BG Cru, and LR TrA and showed that CCFs at the same pulsation phase differ between cycles."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the interpretation of 0.61 Cepheid signals as harmonics of high-degree non-radial modes, used to contextualize X Sgr's and BG Cru's additional frequencies."}],"review_version":1}