REVIEW 3 major objections 4 minor 1 cited by
The Pristine survey. XXVI. Chemical abundances of subgiant stars of the extremelymetal-poor stream C-19
T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash
Pith's one-line read The subgiant stars of the extremely metal-poor stream C-19 share a single metallicity and an intrinsic magnesium spread, indicating that C-19 is a disrupted globular cluster.
desk verdict Careful new subgiant data for C-19, but the Mg spread that carries the globular cluster argument is weak; worth publishing as a data paper, not as a confirmed GC. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The argument rests on a two-dimensional Gaussian mixture model in radial-velocity and metallicity space, which separates the C-19 component from a fixed halo-contamination component, combined with a comparison of magnesium abundances across stream members. The mixture model, evaluated with a Markov-chain Monte Carlo sampler, identifies the likely contaminants and provides the posterior constraints on the mean and dispersion of velocity and metallicity. The magnesium comparison uses the standard deviation of $A(\mathrm{Mg})$ relative to the mean measurement uncertainty to establish the presence of an intrinsic chemical spread, which is the load-bearing evidence for the globular-cluster interpretation.
What would settle it
Measure the magnesium lines in the twelve confirmed C-19 subgiants at higher spectral resolution (R of at least 20,000) so that individual uncertainties drop below 0.1 dex; if the scatter in $A(\mathrm{Mg})$ then shrinks to match the measurement errors, the claimed intrinsic magnesium spread, and with it the globular-cluster interpretation, would be falsified.
Extended reading notes
Core claim
The central discovery is that the subgiant stars of C-19, observed with X-shooter at $G \approx 20$, are consistent with a single, extremely metal-poor stellar population: after removing three likely halo contaminants (S05, S12, and S13), the mean metallicity is $\langle [\mathrm{Fe/H}]\rangle = -3.1 \pm 0.1$ with a dispersion consistent with zero ($\sigma_{[\mathrm{Fe/H}]} < 0.35$ at 95% confidence). The magnesium abundances of the combined sample of subgiants and previously studied giants show a scatter of 0.44 dex against a mean measurement uncertainty of 0.25 dex, which the paper interprets as intrinsic abundance variation. Because such a magnesium spread is a hallmark of globular clusters rather than dwarf galaxies, and because the metallicity dispersion is unresolved, the paper concludes that its preferred interpretation is that C-19 is a disrupted globular cluster. The authors explicitly acknowledge that C-19 could instead be a dwarf galaxy that once hosted such a cluster, or that no cluster was involved, but they argue that in any case a globular cluster must have formed from gas as metal-poor as these stars.
Load-bearing premise
The analysis assumes that the three stars S05, S12, and S13 are halo contaminants and assigns the contamination component fixed velocity and metallicity parameters; if that contamination model is wrong, the inferred mean metallicity, dispersion, and magnesium scatter could change, weakening the globular-cluster conclusion.
Editorial extensions
If this is right
- Globular clusters can apparently form from gas at [Fe/H] ≈ -3.4, so any proposed metallicity floor for their formation must be abandoned or pushed down by about an order of magnitude.
- The C-19 stream's unresolved metallicity dispersion and intrinsic magnesium variation align it chemically with globular clusters rather than with dwarf galaxies, despite its large width and velocity dispersion.
- The twelve confirmed subgiant members have isochrone ages around 13 to 14 Gyr, making them among the oldest and most metal-poor stars known and suitable for further nucleosynthesis and age-dating studies.
- The three excluded stars (S05, S12, and S13) are likely halo contaminants, showing that radial-velocity and metallicity screening is essential when interpreting faint stream candidates selected by astrometric overdensities alone.
Reading between the lines
- A direct test of the globular-cluster hypothesis would be to measure sodium and aluminium in the C-19 subgiants; a Na-O or Mg-Al anticorrelation, as seen in bound clusters, would make the interpretation secure.
- If C-19 is a disrupted cluster embedded in a dwarf galaxy, the three metal-rich stars may trace that galaxy's field population; surveying a wider area around the stream could reveal the extent of that population and the progenitor's total mass.
- The ages close to the cosmic age suggest that the stream formed within the first billion years of the Universe; deeper photometry along the subgiant branch could sharpen the age measurement and constrain the earliest episodes of globular cluster formation.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper presents X-shooter spectroscopy of 15 subgiant candidates in the extremely metal-poor stellar stream C-19, measuring radial velocities and Fe and Mg abundances with the MyGIsFOS code and an independent reanalysis. A two-dimensional likelihood mixture model is used to separate C-19 members from halo contaminants; the authors identify three likely contaminants (S05, S12, S13) and infer for the remaining 12 stars a mean metallicity <[Fe/H]> = -3.1 ± 0.1, a mean radial velocity <v_r> = -192 ± 3 km/s, and a velocity dispersion sigma_vr = 5.9(+3.6,-5.9) km/s. Combining the nine member subgiants with six previously analyzed giants, they find a dispersion in A(Mg) of 0.44 dex against a mean uncertainty of 0.25 dex, which they interpret as evidence for an intrinsic Mg spread. On this basis, together with the unresolved metallicity dispersion, they argue that the preferred interpretation is that C-19 is a disrupted globular cluster, while acknowledging that a dwarf-galaxy progenitor containing a GC cannot be ruled out.
Significance. If the conclusions are robust, the paper would strengthen the case that globular clusters can form from gas at [Fe/H] ~ -3.4, an order of magnitude below the previously suggested metallicity floor for GC formation. The paper is careful in several respects: it provides an independent abundance reanalysis, quantifies systematic uncertainties in the atmospheric parameters, and explicitly acknowledges the sensitivity of the inferred properties to the removal of the three contaminants and the inconclusiveness of the metallicity dispersion alone. However, the central claim that the Mg spread supports a GC origin rests on a detection that is statistically weak in the new subgiant data alone and that may be affected by cross-sample systematics and uncertain membership of individual stars. The age result at the preferred distance is also in tension with the age of the Universe, which weakens the CMD-based membership assessment. These issues make the interpretation plausible but not established by the present data.
major comments (3)
- [Section 5, Figs. 9 and 10] The claim that the A(Mg) dispersion (0.44 dex) exceeds the mean uncertainty (0.25 dex) and thus indicates an intrinsic Mg spread is not statistically robust. The text states that the additional dispersion is detected at only 1.3σ for the subgiant sample alone, rising to 2σ only after 'arbitrarily remov[ing] the two stars with the larger error'. A 2σ effect obtained after a post-hoc removal is not a solid detection. Moreover, the combined 15-star sample mixes the new subgiant abundances with giant-star abundances from Martin et al. (2022) and Yuan et al. (2022), which were derived with different instruments, resolutions, and analysis pipelines. The statement that LTE abundances were used 'in order to be on the same scale' is not a substitute for quantifying zero-point offsets between the two samples. A systematic offset of ~0.4 dex in A(Mg) between giants and subgiants, or between the different literature analyses, would fully explain the observed dispersion. Since this Mg spread is the primary evidence for the GC interpretation, the conclusion needs either a rigorous combined likelihood that includes inter-sample zero-point terms or a demonstration that the spread persists within each homogeneous subsample.
- [Section 3, likelihood model and membership] The mixture model that identifies S05, S12, and S13 as contaminants fixes the contamination component to generic halo expectations (<v_r,cont> = -180 km/s, sigma_vr,cont = 100 km/s, <[Fe/H]>_cont = -1.5, sigma_[Fe/H],cont = 0.3). The inferred C-19 mean metallicity and dispersion, and hence the statement that the stream has a single metallicity, are sensitive to this choice, as the authors acknowledge. The paper does not provide a sensitivity analysis: for example, varying the contamination mean and dispersion within plausible ranges, or allowing the contamination parameters to be partially free with informative priors, would show whether the conclusion of only three contaminants and the resulting <[Fe/H]> = -3.1 ± 0.1 is stable. This is load-bearing because a mis-specified contamination component could either hide a real metallicity spread in C-19 or create an artificial narrow metallicity peak from a subset of the data.
- [Sections 4 and 5, age and membership] The paper derives isochrone ages that are older than the Universe at the preferred distance of 18 kpc (14.1 Ga vs 13.7 Ga), and only become consistent at 20.9 kpc. This internal tension is acknowledged but not resolved, yet it directly affects the membership assessment: S04 is flagged as a likely interloper partly because it is too young in the tau2 and tau3 projections, and S12/S13 are rejected as too young. If the distance is uncertain, the ages that drive these membership decisions are correspondingly uncertain. The paper should either adopt a self-consistent distance/age treatment (e.g., marginalize over distance in the age estimates) or explicitly test how the membership conclusions and the Mg-dispersion result change if S04 and S09 (the suspected SB2 binary) are excluded from the 15-star sample. The current text includes these stars in the Mg dispersion despite stating that S09's abundances should be considered with caution and that S04 may be an interloper.
minor comments (4)
- [Introduction] The sentence 'In this paper, we describe the analysis and the results of these observations' appears as a fragment after a discussion of the instrumental setup; it should be integrated into the preceding paragraph.
- [Section 2] The text says 'The signal-to-noise ratios (S/N) of all the NIR spectra were of too to be scientifically exploited'; this appears to be a typo, likely 'too low'.
- [Section 5] The paper uses 'metalicity' in the caption of Fig. 5; this should be 'metallicity'.
- [Figure 11] The caption of Fig. 11 states 'Orbits of the C-19 in Yuan et al. (2022)' but the figure also shows orbits for S12 and S13; the caption should clarify which curves correspond to which component.
Circularity Check
No significant circularity: the mixture-model membership cleaning is a sensitivity analysis with fixed contamination priors, and the Mg-spread evidence rests on external published abundances rather than on a fitted input.
full rationale
The paper's derivation chain does not reduce any prediction to its own inputs by construction. The membership model is a two-component likelihood in which the contamination parameters are fixed a priori to generic halo expectations (⟨vr,cont⟩=-180 km/s, σ=100 km/s, ⟨[Fe/H]⟩=-1.5, σ=0.3), so the classification of S05/S12/S13 as contaminants is not a free fit of the C-19 parameters. After removing these three stars, the mean metallicity, radial velocity, and dispersions are re-estimated from the remaining 12 spectra; the paper explicitly acknowledges that the posteriors 'are sensitive to our choice of removing the three likely contaminants' and presents the result as an inference, not as an independent prediction. This is a standard sensitivity analysis, not a definitional equivalence. The Mg-abundance dispersion that carries the globular-cluster interpretation is computed from 15 member stars (9 new subgiants plus 6 giants from Martin et al. 2022 and Yuan et al. 2022); the giant abundances are published external data, and the paper itself reports that the subgiant sample alone detects the extra dispersion at only 1.3σ (2σ after removing two high-error stars), which is a statistical weakness rather than circularity. The self-citations (Martin et al. 2022, Yuan et al. 2022, Errani et al. 2022, Viswanathan et al. 2024) are used as prior measurements and simulations, not as an imported uniqueness theorem or an unverified ansatz. The paper also candidly states that 'the possibility that no GC was associated with C-19 cannot be ruled out either,' further confirming that the central interpretation is presented as a preference, not as a forced consequence of the model.
Assumptions & free parameters
free parameters (10)
- Contamination fraction eta =
not quoted (posterior)
- Mean radial velocity <v_r> =
-192 +/- 3 km/s
- Velocity dispersion sigma_vr =
5.9^{+3.6}_{-5.9} km/s
- Mean metallicity <[Fe/H]> =
-3.1 +/- 0.1
- Metallicity dispersion sigma_[Fe/H] =
0.09^{+0.13}_{-0.09}
- Mg abundance dispersion =
not quoted; A(Mg) std=0.44 dex
- Assumed distance =
18 kpc (and 20.9 kpc alternative)
- Assumed stellar mass =
0.8 M_sun
- Assumed extinction A(V) =
0.048
- Contamination model parameters =
<v_r,cont>=-180 km/s, sigma=100 km/s, <[Fe/H]>=-1.5, sigma=0.3
assumptions (5)
- domain assumption LTE abundance analysis with ATLAS12 and MARCS model atmospheres is accurate for these stars
- domain assumption X-shooter wavelength calibration has a systematic uncertainty of 7.5 km/s which is added linearly to statistical errors
- ad hoc to paper The halo contamination component in the likelihood model is adequately represented by a 2D Gaussian with fixed parameters
- domain assumption The stars are at the assumed distance (18 kpc or 20.9 kpc) and have mass 0.8 M_sun
- domain assumption BASTI isochrones with [alpha/Fe]=0.4 and Y=0.247 are appropriate for age estimation
Cite this review
Pith. "Pith review of The Pristine survey. XXVI. Chemical abundances of subgiant stars of the extremelymetal-poor stream C-19." pith.science (2026). https://pith.science/paper/S7V3DWW5
@misc{pith2026241220776,
author = {Pith},
title = {Pith review of: The Pristine survey. XXVI. Chemical abundances of subgiant stars of the extremelymetal-poor stream C-19},
year = {2026},
howpublished = {\url{https://pith.science/paper/S7V3DWW5}},
note = {Machine review of arXiv:2412.20776}
}
abstract
Context: The C-19 stellar stream is the most metal-poor stream known to date. While its width and velocity dispersion indicate a dwarf galaxy origin, its metallicity spread and abundance patterns are more similar to those of globular clusters (GCs). If it is indeed of GC origin, its extremely low metallicity ([Fe/H]=-3.4, estimated from giant stars) implies that these stellar systems can form out of gas that is as extremely poor in metals as this. Previously, only giant stream stars were observed spectroscopically, although the majority of stream stars are unevolved stars. Aims: We pushed the spectroscopic observations to the subgiant branch stars ($G\approx 20$) in order to consolidate the chemical and dynamical properties of C-19. Methods: We used the high-efficiency spectrograph X-shooter fed by the ESO 8.2 m VLT telescope to observe 15 candidate subgiant C-19 members. The spectra were used to measure radial velocities and to determine chemical abundances using the \mygi\ code. Results; We developed a likelihood model that takes metallicity and radial velocities into account. We conclude that 12 stars are likely members of C-19, while 3 stars (S05, S12, and S13) are likely contaminants. When these 3 stars are excluded, our model implies a mean metallicity $\rm \langle [Fe/H]\rangle = -3.1\pm 0.1$, the mean radial velocity is $\langle v_r\rangle = -192\pm3$ kms$^{-1}$, and the velocity dispersion is $\sigma_{vr} = 5.9^{+3.6}_{-5.9}$ kms$^{-1}$. This all agrees within errors with previous studies. The A(Mg) of a sample of 15 C-19 members, including 6 giant stars, shows a standard deviation of 0.44 dex, and the mean uncertainty on Mg is 0.25 dex. Conclusions: Our preferred interpretation of the current data is that C-19 is a disrupted GC. We cannot completely rule out the possibility that the GC could have belonged to a dwarf galaxy that contained more metal-rich stars, however. This scenario would explain the radial velocity members at higher metallicity, as well as the width and velocity dispersion of the stream. In either case, a GC formed out of gas as poor in metals as these stars seems necessary to explain the existence of C-19. The possibility that no GC was associated with C-19 cannot be ruled out either.
Figures
Figures from the paper (7 more)
Forward citations
Cited by 1 Pith paper
-
The Pristine survey: XXVII. The extremely metal-poor stream C-19 stretches over more than 100 degrees
The extremely metal-poor stellar stream C-19 is shown to extend over more than 100 degrees, with newly confirmed members indicating higher velocity dispersion and mass than previously measured.
Reference graph
Works this paper leans on
-
[1]
& Plez, B
Alvarez, R. & Plez, B. 1998, A&A, 330, 1109
1998
-
[2]
2024, A&A, 687, L3
Balbinot, E., Dodd, E., Matsuno, T., et al. 2024, A&A, 687, L3
2024
-
[3]
A., Leaman, R., Gallart, C., et al
Beasley, M. A., Leaman, R., Gallart, C., et al. 2019, MNRAS, 487, 1986
work page 2019
-
[4]
L., Lallement, R., Ferron, S., Boonne, C., & Bodichon, R
Bertaux, J. L., Lallement, R., Ferron, S., Boonne, C., & Bodichon, R. 2014, A&A, 564, A46
work page 2014
-
[5]
2006, in Chemical Abundances and Mixing in Stars in the Milky Way and its Satellites, 232
Bonifacio, P., Zaggia, S., Sbordone, L., et al. 2006, in Chemical Abundances and Mixing in Stars in the Milky Way and its Satellites, 232
work page 2006
-
[6]
2024, A&A, 684, A37
Ceccarelli, E., Massari, D., Mucciarelli, A., et al. 2024, A&A, 684, A37
2024
-
[7]
M., Helmi, A., et al
Dodd, E., Callingham, T. M., Helmi, A., et al. 2023, A&A, 670, L2
2023
-
[8]
Errani, R., Navarro, J. F., Ibata, R., et al. 2022, MNRAS, 514, 3532 François, P., Caffau, E., Bonifacio, P., et al. 2018, A&A, 620, A187
work page 2022
Show all 48 references
-
[9]
2006, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol
Goldoni, P., Royer, F., François, P., et al. 2006, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 6269, Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, ed. I. S. McLean & M. Iye, 62692K
2006
-
[10]
G., Bonifacio, P., Bragaglia, A., et al
Gratton, R. G., Bonifacio, P., Bragaglia, A., et al. 2001, A&A, 369, 87
2001
-
[11]
J., El-Souri, M., Monaco, L., et al
Hansen, C. J., El-Souri, M., Monaco, L., et al. 2018, ApJ, 855, 83
2018
-
[12]
Harris, W. E. 1996, AJ, 112, 1487
1996
-
[13]
Hastings, W. K. 1970, Biometrika, 57, 97
1970
-
[14]
2021, A&A, 645, A106
Heiter, U., Lind, K., Bergemann, M., et al. 2021, A&A, 645, A106
2021
-
[15]
& White, S
Helmi, A. & White, S. D. M. 1999, MNRAS, 307, 495
1999
-
[16]
2021, ApJ, 914, 123
Ibata, R., Malhan, K., Martin, N., et al. 2021, ApJ, 914, 123
2021
-
[17]
A., Gilmore, G., & Irwin, M
Ibata, R. A., Gilmore, G., & Irwin, M. J. 1994, Nature, 370, 194
1994
-
[18]
J., et al
Kordopatis, G., Schultheis, M., McMillan, P. J., et al. 2023, A&A, 669, A104
2023
-
[19]
Kruijssen, J. M. D., Pfe ffer, J. L., Reina-Campos, M., Crain, R. A., & Bastian, N. 2019, MNRAS, 486, 3180
2019
-
[20]
Kurucz, R. L. 2005, Memorie della Societá Astronomica Italiana Supplementi, 8, 14
2005
-
[21]
S., Strader, J., & Brodie, J
Larsen, S. S., Strader, J., & Brodie, J. P. 2012, A&A, 544, L14
2012
-
[22]
2021, A&A, 656, A155
Lombardo, L., François, P., Bonifacio, P., et al. 2021, A&A, 656, A155
2021
-
[23]
Madau, P., Lupi, A., Diemand, J., Burkert, A., & Lin, D. N. C. 2020, ApJ, 890, 18
2020
-
[24]
A., Sharma, S., et al
Malhan, K., Ibata, R. A., Sharma, S., et al. 2022, ApJ, 926, 107
2022
-
[25]
A., et al
Malhan, K., Yuan, Z., Ibata, R. A., et al. 2021, ApJ, 920, 51
2021
-
[26]
1991, A&A, 252, 94
Mandushev, G., Staneva, A., & Spasova, N. 1991, A&A, 252, 94
1991
-
[27]
F., Venn, K
Martin, N. F., Venn, K. A., Aguado, D. S., et al. 2022, Nature, 601, 45
2022
-
[28]
2021, A&A Rev., 29, 5 Article number, page 9 of 12 A&A proofs: manuscript no
Matteucci, F. 2021, A&A Rev., 29, 5 Article number, page 9 of 12 A&A proofs: manuscript no. aa51517-24corr
2021
-
[29]
McConnachie, A. W. & Venn, K. A. 2020, AJ, 160, 124
2020
-
[30]
W., Rosenbluth, M
Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H., & Teller, E. 1953, J. Chem. Phys., 21, 1087
1953
-
[31]
2023, A&A, 669, A54
Minelli, A., Bellazzini, M., Mucciarelli, A., et al. 2023, A&A, 669, A54
2023
-
[32]
2017, A&A, 605, A46
Mucciarelli, A., Bellazzini, M., Ibata, R., et al. 2017, A&A, 605, A46
2017
-
[33]
2021, A&A, 653, A90
Mucciarelli, A., Bellazzini, M., & Massari, D. 2021, A&A, 653, A90
2021
-
[34]
P., Conroy, C., Bonaca, A., et al
Naidu, R. P., Conroy, C., Bonaca, A., et al. 2020, ApJ, 901, 48
2020
-
[35]
2017, A&A, 601, A112
Pancino, E., Romano, D., Tang, B., et al. 2017, A&A, 601, A112
2017
-
[36]
Paradiso, S., Colombo, L. P. L., Andersen, K. J., et al. 2023, A&A, 675, A12
2023
-
[37]
2021, ApJ, 908, 102 Planck Collaboration, Adam, R., Ade, P
Pietrinferni, A., Hidalgo, S., Cassisi, S., et al. 2021, ApJ, 908, 102 Planck Collaboration, Adam, R., Ade, P. A. R., et al. 2016, A&A, 594, A1
2021
-
[38]
2012, Turbospectrum: Code for spectral synthesis, Astrophysics Source Code Library, record ascl:1205.004
Plez, B. 2012, Turbospectrum: Code for spectral synthesis, Astrophysics Source Code Library, record ascl:1205.004
2012
-
[39]
2014, A&A, 564, A109
Sbordone, L., Caffau, E., Bonifacio, P., & Duffau, S. 2014, A&A, 564, A109
2014
-
[40]
Schlafly, E. F. & Finkbeiner, D. P. 2011, ApJ, 737, 103
2011
-
[41]
2024, A&A, 689, A201
Sestito, F., Vitali, S., Jofre, P., et al. 2024, A&A, 689, A201
2024
-
[42]
Simon, J. D. 2019, ARA&A, 57, 375
2019
-
[43]
C., Geha, M., Kirby, E
Vargas, L. C., Geha, M., Kirby, E. N., & Simon, J. D. 2013, ApJ, 767, 134
2013
-
[44]
2011, A&A, 536, A105
Vernet, J., Dekker, H., D’Odorico, S., et al. 2011, A&A, 536, A105
2011
-
[45]
2024, A& A submitted, arXiv e-prints, arXiv:2405.13124
Viswanathan, A., Yuan, Z., Ardern-Arentsen, A., et al. 2024, A& A submitted, arXiv e-prints, arXiv:2405.13124
2024 arXiv
-
[46]
Wilson, A. G. 1955, PASP, 67, 27
1955
-
[47]
C., & Huang, Y
Yuan, Z., Chang, J., Beers, T. C., & Huang, Y . 2020, ApJ, 898, L37
2020
-
[48]
F., Ibata, R
Yuan, Z., Martin, N. F., Ibata, R. A., et al. 2022, MNRAS, 514, 1664 Article number, page 10 of 12 P. Bonifacio et al.: Pristine XXIV - C19 SG Appendix A: Tables This appendix contains the tables with all the data cited in the paper. Table A.1. Identification and photometry of...
2022
Reviewed August 10, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.