Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Modeling the Gaia Color-Magnitude Diagram with Bayesian Neural Flows to Constrain Distance Estimates

T0 review · 4 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read The paper claims that a normalizing flow trained on noisy Gaia DR2 measurements can learn the Milky Way's color-magnitude diagram and use it as a Bayesian prior, yielding distance posteriors for 640 million stars with an average…

desk verdict The first normalizing-flow CMD prior for Gaia distance estimation is a genuine methodological step, but the catalog's accuracy rests on a single low-reddening cluster and a fixed extinction law, so treat the results as provisional until broader validation. read the letter →

arxiv 1908.08045 v1 pith:GO6W47GH submitted 2019-08-21 astro-ph.IM astro-ph.GAcs.LGstat.ML

classification astro-ph.IMastro-ph.GAcs.LGstat.ML
keywords normalizingflowscolor-magnitudediagramGaiaDR2photometricdistancesBayesianinferenceinterstellarduststellarpopulationsMilkyWaystructure
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to show that a fully data-driven model of the Gaia color-magnitude diagram — a normalizing flow trained on noisy parallaxes and photometry — can act as a Bayesian prior that converts weak parallax measurements into much sharper distance estimates. The authors claim the resulting catalog of 640 million photometric distance posteriors improves distance signal-to-noise by 48.6% on average over raw non-negative Gaia parallaxes, and that the gains are largest exactly where parallax is worst: low-signal-to-noise and negative-parallax stars. They argue the approach avoids the systematic errors of theoretical stellar models and the prior-dominance of the standard Milky Way distance prior, and they demonstrate it on a simulated dataset, the GD-1 stream, and the cluster M67. If the central claim is right, a large fraction of Gaia's catalog can be used for three-dimensional mapping of the Milky Way and for finding substructures like stellar streams beyond 1 kpc.

What carries the argument

The load-bearing object is a Masked Autoregressive Flow, constructed as repeated blocks of masked-autoencoder layers (MADE), BatchNorm, and Reverse layers, which represents the probability density of a star's dereddened absolute magnitudes $(g,\,\mathrm{bp}-\mathrm{rp},\,\mathrm{bp}-g)$. Because the MADE mask makes the flow's Jacobian triangular, the change-of-variables determinant is a cheap product, so the network can be trained by directly maximizing $\frac{1}{\sigma_P^2}\log P$ with $P$ the flow density and $\sigma_P^2$ the combined variance of the photometric and dust uncertainties. Around the flow, an iterative loop estimates distance without a prior: 32 parallax samples are drawn from a truncated normal, converted to distances, used to query the Bayestar dust map, and converted into dereddened magnitudes; each candidate is weighted by the flow probability, and the weighted sum becomes the next distance estimate. This loop runs five times in training and ten in evaluation, jointly settling distance, dust, and the CMD prior.

What would settle it

Take stars with independent distances measured by other means, cover a wide range of dust columns, and test whether the model's distance errors grow or correlate with the amount of reddening; a strong correlation would falsify the dust-and-extinction assumption at the base of the method.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that a normalizing flow can learn the joint density $P(g,\,\mathrm{bp}-\mathrm{rp},\,\mathrm{bp}-g)$ of dereddened absolute magnitudes directly from noisy Gaia DR2 sources, and that this learned density, when combined with iterative Bayestar dust estimation, produces a posterior over distance for every star. The paper reports that distance signal-to-noise improves on average by 48.6% relative to using the raw parallaxes for stars with positive parallax, the fraction of stars with signal-to-noise below 1 drops from 46.1% to 18.9%, and stars with negative parallaxes receive positive, finite distances. In a 30-million-star simulation the flow reconstructs the true color-magnitude diagram from heavily reddened, noisy data, and on real data it tightens the main sequence and giant branch while pulling GD-1 stream candidates to the stream's accepted distance of roughly 7–10 kpc instead of the prior-dominated values. The authors present this as evidence that a flexible, learned CMD prior beats both inverse-parallax distances and geometry-based distance priors in the noisy regime, without committing to theoretical stellar models.

Load-bearing premise

The whole chain assumes the dust map used to remove reddening is correct along every line of sight; if it is biased, the learned stellar colors and every distance estimate are biased with it.

Editorial extensions

If this is right

  • A catalog of 640,875,169 distance posteriors is produced, with only 18.9% of stars at signal-to-noise below 1 compared with 46.1% in the raw Gaia catalog.
  • Stars with negative parallaxes, 22.3% of the raw sample, receive positive distance posterior means instead of being unusable.
  • Distance estimates for GD-1 stream candidates improve from prior-dominated values to a median near 8 kpc, making kinematic substructure searches beyond 1 kpc feasible without assuming a shared isochrone.
  • The same flow architecture can exactly marginalize over missing photometric bands, so future versions can combine Gaia with other surveys without discarding sources missing some bands.
  • M67's distance from noisy-parallax stars, 0.845 kpc with a Gaussian spread of 0.006 kpc at SNR<50, is closer to the known 840 pc cluster distance and tighter than either inverse parallax or the standard geometric distance prior.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the 48.6% average holds across all stellar populations, the per-source gain likely varies with dust column and intrinsic color, so users should verify gains on their own subsamples before trusting the catalog for fine structure studies.
  • The method's remaining model dependence is concentrated in the dust map and a fixed $R_V=3.1$ extinction law; jointly learning the dust map and the CMD is the natural next step, and the iterative loop in this paper is already structured to admit it.
  • Because the autoregressive flow can marginalize over missing bands, a direct testable extension is to add near-infrared photometry to the same density model and check whether the distance posteriors of dust-obscured stars improve beyond the Gaia-only version.
  • A cautionary consequence: the catalog inherits the global 0.029 mas parallax zero-point correction and any residual parallax systematics, so absolute distances in the catalog should be validated against independent distance anchors rather than treated as purely photometric.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper presents a normalizing flow model that learns a three-dimensional color-magnitude diagram (CMD) from Gaia DR2 photometry and parallaxes, with iterative dereddening using the Bayestar dust map, and then uses this learned density as a prior to produce photometric distance posteriors for roughly 640 million stars. The authors claim an average 48% improvement in distance signal-to-noise relative to raw Gaia parallaxes and demonstrate applications to the GD-1 stream and the open cluster M67. A simulation with a simple analytic CMD shows that the model can approximately recover the true CMD from noisy, reddened data, and the real-data CMD shows visible main-sequence and giant-branch structure.

Significance. If validated, this work would provide a large public catalog of photometric distance posteriors and demonstrate a useful application of normalizing flows to stellar density estimation. The planned release of code and catalog is a strength, as is the explicit treatment of parallax and photometric uncertainties and the iterative incorporation of dust. However, the central accuracy claims currently rest on a small number of tests, and the self-training nature of the pipeline plus the untested fixed extinction law make the 48% SNR improvement and the accuracy improvements difficult to assess. The method is plausible and potentially valuable, but the evidence presented is not yet sufficient to support the catalog-scale claims.

major comments (4)
  1. [Section 3.2] The iterative dereddening scheme in Section 3.2 uses model-derived distances in Eq. (2) to estimate dust and deredden the photometry, and the same dereddened photometry is used to train the CMD prior that later produces distances for the catalog. This creates a self-training loop. The manuscript does not demonstrate that this loop converges to unbiased estimates or quantify the bias it may introduce. The simulation in Section 4.1 uses the same fixed extinction law and a simple dust map, so it cannot reveal problems with mis-specification, and the M67 validation in Section 4.3 is at low reddening (E(B-V) approximately 0.03), which does not exercise the dusty regime where the fixed extinction conversion is most risky. Please add a test with a deliberately mis-specified extinction law or a validation sample with independent distances in high-reddening regions.
  2. [Table 1 and abstract] The headline claim of a 48% average SNR improvement compares Gaia SNR defined as parallax over parallax uncertainty with model SNR defined as distance over distance uncertainty (Table 1). These are not directly comparable quantities, and the model's distance uncertainty comes from a posterior that includes the learned CMD prior, so a narrow posterior does not imply accuracy. Furthermore, the comparison excludes negative parallaxes for the Gaia column, while the model column has no negative distances by construction, which inflates the reported improvement. Please report accuracy metrics (e.g., bias and scatter against true distances in simulation or benchmark clusters) alongside precision metrics, and define the SNR improvement in a way that is consistent between the two quantities being compared.
  3. [Section 4.1] The simulation in Section 4.1 constructs a 'true' CMD by fitting a line to high-SNR Gaia data and adding Gaussian perturbations, so the recovery test is performed on an idealized linear relation rather than a realistic multi-population CMD. The comparison between Fig. 2a and Fig. 2b is visual only, with no quantitative metric of reconstruction quality, and the simulation does not validate the distance posteriors themselves (only the CMD density). Please add quantitative reconstruction metrics, such as the KL divergence between the true and learned CMDs, and a simulation with a more realistic CMD (e.g., multiple stellar populations, age and metallicity spreads) together with a comparison of inferred distances against true distances as a function of SNR and dust.
  4. [Section 4.3 and Table 2] The M67 test is the principal real-data accuracy validation, but it relies on a two-component GMM fit to the distance distribution, and the 'spread' in Table 2 is the width of the cluster component, which is not a rigorous measure of per-star distance accuracy. The improvement for the SNR<15 case (inverse-parallax center 0.77 kpc, model center 0.83 kpc, versus the adopted true distance around 0.84 kpc) is modest and based on a single cluster. Please provide a validation on a larger set of clusters with known distances spanning a range of reddening and distance, and report per-star accuracy metrics such as median absolute error or bias as a function of true SNR.
minor comments (6)
  1. [General] The manuscript contains several typographical and formatting issues: 'flexible' and 'T able 1' appear in the text, the URL 'norm flows.html' has an unescaped space, and the reference 'Bovy Jo et al.' is incorrectly capitalized. Please proofread carefully.
  2. [Section 5] The claim that the autoregressive flow allows exact marginalization is only true for a fixed variable ordering, as the text acknowledges; please state this caveat more explicitly in the main text so it is not read as a general exact marginalization over arbitrary photometric bands.
  3. [Section 2] Please specify the exact Gaia DR2 data release epoch (e.g., the 2018 April release) and the specific version or footprint of the Bayestar dust map used, as both affect reproducibility.
  4. [Figures 3 and 4] The color scales in Figures 3 and 4 are not fully described; please add explicit labels and color-bar units (e.g., probability density and log number density) so the visual 'tightening' claim can be evaluated quantitatively.
  5. [Equation (1)] Equation (1) is described as a 'minimal distance prior', but it is actually a truncated normal prior on parallax; please rephrase to avoid confusion between the parallax prior and the distance prior discussed in Section 4.2.
  6. [References] The Hogg (2018) reference is cited as an arXiv preprint; please update to the published version if one exists, and ensure all software citations (dustmap, extinction, PyTorch) include proper version information.

Circularity Check

1 steps flagged · score 5.0 of 10

Iterative dust loop makes the CMD prior depend on its own distance estimates, so the precision gain is partially self-referential; external simulation and M67 checks keep the core method from being fully circular.

  1. self definitional [Section 3.2, Eq. (2) and the loss-function paragraph]
    "These likelihood values are treated as weights, and the weights are used to calculate a new best-estimate for distance via a weighted sum: dbest = Σ_i d_i P(g_i,bp−rp,bp−g) / Σ_i P(g_i,bp−rp,bp−g). This dbest is then fed back into the loop and a new reddening is found using Bayestar. ... Once the final dbest is given, we calculate a best-estimate value for the dust. We use this to get final estimates for the dereddened G,bp−rp,bp−g."

    The normalizing flow P is trained on dereddened photometry (g, bp−rp, bp−g), and that dereddening is computed from dbest, which is itself a P-weighted average of distance samples (Eq. 2). Thus the training targets for P depend on P: changing P changes dbest, which changes the colors P is fit to. The final catalog distances are then produced with this same P, so the reported precision improvement is partly a self-consistent shrinkage rather than an independent measurement. The authors explicitly acknowledge a feedback loop and avoid it for g ('we do not use dbest to calculate the final g since this could create a feedback loop'), but the same loop remains through the colors used in training.

full rationale

The paper's central derivation is a Bayesian normalizing-flow prior over dereddened Gaia photometry, combined with parallax likelihoods. The most significant self-referential element is the iterative dust loop: dbest is defined via Eq. (2) as a P-weighted average, and then used to deredden the very photometry on which P is trained. This is a genuine feedback loop, and the paper itself notes that it can create unphysical artifacts. However, it is an iterative fixed-point procedure (similar to EM) rather than a pure tautology: raw parallax measurements, observed photometry, and the external Bayestar dust map all enter, and the simulation in Sec. 4.1 and the M67 test in Sec. 4.3 provide independent checks, although both still use the same iterative loop. The fixed extinction coefficients (AG = 2.71E(g−r), etc.) and the Bayestar dust assumption are load-bearing but are stated assumptions, not circular reductions. No load-bearing self-citation chain or ansatz-smuggling was found. The partial circularity is thus moderate, not total, so the score is 5.

Assumptions & free parameters 5 free parameters · 7 assumptions · 0 invented entities

The central claims rest on a learned CMD prior and an iterative dust correction. The key free parameters are the neural network weights and architecture choices, all fitted to the same Gaia DR2 data. The most consequential assumptions concern the fidelity of the Bayestar dust map, the extinction conversion coefficients, and the absence of residual parallax systematics.

free parameters (5)
  • Normalizing flow weights = Not reported (millions of parameters)
    The MAF network parameters are learned from Gaia DR2 training data and define the CMD prior, which is the core of the distance estimation.
  • Number of flow blocks = 35
    Selected via Bayesian hyperparameter search over 80 models; affects model flexibility and prior quality.
  • Hidden units per MADE layer = 500
    Same hyperparameter search; sets the capacity of each autoregressive layer.
  • Training dust iterations = 5
    Chosen for computational expense; evaluation uses 10 iterations, a training-evaluation mismatch.
  • Mini-batch size = 2048
    Chosen to balance artifact formation and model quality; smaller batches created artifacts.
assumptions (7)
  • domain assumption The Bayestar dust map produces accurate dust posteriors along each line of sight.
    Stated explicitly in Section 3 model assumptions; if wrong, dereddening is biased and all resulting distances are affected.
  • domain assumption The fixed extinction conversion coefficients (AG = 2.71 E(g-r), E(BP-RP) = 0.85 E(g-r), E(BP-G) = 0.39 E(g-r)) are accurate for Gaia bands.
    Assumed in Section 3.2; incorrect conversions would shift the CMD prior in color and absolute magnitude.
  • domain assumption The Gaia catalog has no parallax bias relative to any parameters beyond a single constant offset of 0.029 mas.
    Section 3 applies a global parallax offset; any residual spatially or color-dependent bias would propagate into the distance catalog.
  • domain assumption The expected color-magnitude relation is unchanging with respect to location in the Milky Way.
    The CMD model does not include alpha, delta as inputs, so spatial metallicity or age variations could bias the prior.
  • domain assumption The truncated normal parallax prior (Equation 1) is an appropriate minimal distance prior.
    Used during training to sample positive parallaxes; effectively imposes d > 0 without additional prior information.
  • domain assumption A flat prior over non-negative distances is appropriate during evaluation.
    The catalog is generated with a constant distance prior instead of the Bailer-Jones prior, chosen to avoid prior dominance in the halo.
  • domain assumption The contents of Gaia DR2, after cuts, are representative enough of the Milky Way's stellar populations.
    The paper notes that the CMD model will bias to disk metallicities because halo stars have noisier parallaxes, but still trains on all stars to lessen the effect.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Modeling the Gaia Color-Magnitude Diagram with Bayesian Neural Flows to Constrain Distance Estimates." pith.science (2026). https://pith.science/paper/GO6W47GH

@misc{pith2026190808045,
  author       = {Pith},
  title        = {Pith review of: Modeling the Gaia Color-Magnitude Diagram with Bayesian Neural Flows to Constrain Distance Estimates},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GO6W47GH}},
  note         = {Machine review of arXiv:1908.08045}
}
read the original abstract

We demonstrate an algorithm for learning a flexible color-magnitude diagram from noisy parallax and photometry measurements using a normalizing flow, a deep neural network capable of learning an arbitrary multi-dimensional probability distribution. We present a catalog of 640M photometric distance posteriors to nearby stars derived from this data-driven model using Gaia DR2 photometry and parallaxes. Dust estimation and dereddening is done iteratively inside the model and without prior distance information, using the Bayestar map. The signal-to-noise (precision) of distance measurements improves on average by more than 48% over the raw Gaia data, and we also demonstrate how the accuracy of distances have improved over other models, especially in the noisy-parallax regime. Applications are discussed, including significantly improved Milky Way disk separation and substructure detection. We conclude with a discussion of future work, which exploits the normalizing flow architecture to allow us to exactly marginalize over missing photometry, enabling the inclusion of many surveys without losing coverage.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Extreme Deconvolution Reimagined: Conditional Densities via Neural Networks and an Application in Quasar Classification

    astro-ph.IM 2024-12 conditional novelty 5.0 of 10

    For quasar-contaminant colors, CondXD produces noise-deconvolved conditional densities that visually match binned extreme deconvolution while training roughly ten times faster.

Reference graph

Works this paper leans on

32 extracted references · 2 canonical work pages · cited by 1 Pith paper

  1. [1]

    2019, MNRAS, stz1960, doi: 10.1093/mnras/stz1960

    Alsing, J., Charnock, T., Feeney, S., & Wandelt, B. 2019, MNRAS, stz1960, doi: 10.1093/mnras/stz1960

  2. [2]

    W., Leistedt, B., Price-Whelan, A

    Anderson, L., Hogg, D. W., Leistedt, B., Price-Whelan, A. M., & Bovy, J. 2018, AJ, 156, 145, doi: 10.3847/1538-3881/aad7bf

  3. [3]

    Bailer-Jones, C. A. L. 2015, PASP, 127, 994, doi: 10.1086/683116

  4. [4]

    2018, A&A, 156, 58, doi: 10.3847/1538-3881/aacb21

    Mantelet, G., & Andrae, R. 2018, A&A, 156, 58, doi: 10.3847/1538-3881/aacb21

  5. [5]

    2016, extinction v0.3.0, doi: 10.5281/zenodo.804967

    Barbary, K. 2016, extinction v0.3.0, doi: 10.5281/zenodo.804967. https://doi.org/10.5281/zenodo.804967

  6. [6]

    W., Price-Whelan, A

    Bonaca, A., Hogg, D. W., Price-Whelan, A. M., & Conroy, C. 2019, ApJ, 880, 38, doi: 10.3847/1538-4357/ab2873

  7. [7]

    2017, MNRAS, 470, 1360, doi: 10.1093/mnras/stx1277 Bovy Jo, Hogg, D

    Bovy, J. 2017, MNRAS, 470, 1360, doi: 10.1093/mnras/stx1277 Bovy Jo, Hogg, D. W., & Roweis, S. T. 2011, Annals of Applied Statistics, 5, 1657, doi: 10.1214/10-AOAS439

  8. [8]

    2019, arXiv e-prints, arXiv:1907.10621

    Brehmer, J., Kling, F., Espejo, I., & Cranmer, K. 2019, arXiv e-prints, arXiv:1907.10621

Show all 32 references
  1. [9]

    Brown, A. G. A., Vallenari, A., Prusti, T., et al. 2018, A&A, 616, A1, doi: 10.1051/0004-6361/201833051

  2. [10]

    2016, arXiv:1605.08803 [cs, stat] Modeling Color-Magnitude Diagrams with Neural Flows 15

    Dinh, L., Sohl-Dickstein, J., & Bengio, S. 2016, arXiv:1605.08803 [cs, stat] Modeling Color-Magnitude Diagrams with Neural Flows 15

  3. [11]

    2015, arXiv:1502.03509 [cs, stat]

    Germain, M., Gregor, K., Murray, I., & Larochelle, H. 2015, arXiv:1502.03509 [cs, stat]

  4. [12]

    2010, in Proceedings of the thirteenth international conference on artificial intelligence and statistics, 249–256

    Glorot, X., & Bengio, Y. 2010, in Proceedings of the thirteenth international conference on artificial intelligence and statistics, 249–256

  5. [13]

    2018, The Journal of Open Source Software, 3, 695, doi: 10.21105/joss.00695

    Green, G. 2018, The Journal of Open Source Software, 3, 695, doi: 10.21105/joss.00695

  6. [14]

    Finkbeiner, D. P. 2019, arXiv e-prints, arXiv:1905.02734

  7. [15]

    M., Schlafly, E

    Green, G. M., Schlafly, E. F., Finkbeiner, D., et al. 2018, MNRAS, 478, 651, doi: 10.1093/mnras/sty1008

  8. [16]

    J., Davies, G

    Hall, O. J., Davies, G. R., Elsworth, Y. P., et al. 2019, MNRAS, 486, 3569, doi: 10.1093/mnras/stz1092

  9. [17]

    Hawkins, K., Leistedt, B., Bovy, J., & Hogg, D. W. 2017, MNRAS, 471, 722, doi: 10.1093/mnras/stx1655

  10. [18]

    Hogg, D. W. 2018, ArXiv e-prints, 1804, arXiv:1804.07766

  11. [19]

    2017, ApJ, 844, 102, doi: 10.3847/1538-4357/aa75ca

    Huber, D., Zinn, J., Bojsen-Hansen, M., et al. 2017, ApJ, 844, 102, doi: 10.3847/1538-4357/aa75ca

  12. [20]

    E., Belokurov, V., Li, T

    Koposov, S. E., Belokurov, V., Li, T. S., et al. 2019, MNRAS, 485, 4726, doi: 10.1093/mnras/stz457

  13. [21]

    Leistedt, B., & Hogg, D. W. 2017, A&A, 154, 222, doi: 10.3847/1538-3881/aa91d5

  14. [22]

    2018, A&A, 616, A2, doi: 10.1051/0004-6361/201832727

    Lindegren, L., Hernndez, J., Bombrun, A., et al. 2018, A&A, 616, A2, doi: 10.1051/0004-6361/201832727

  15. [23]

    Malhan, K., & Ibata, R. A. 2018, MNRAS, 477, 4063, doi: 10.1093/mnras/sty912

  16. [24]

    A., & Martin, N

    Malhan, K., Ibata, R. A., & Martin, N. F. 2018, MNRAS, 481, 3442, doi: 10.1093/mnras/sty2474 O’Donnell, J. E. 1994, ApJ, 422, 158, doi: 10.1086/173713

  17. [25]

    2017, arXiv e-prints, arXiv:1705.07057

    Papamakarios, G., Pavlakou, T., & Murray, I. 2017, arXiv e-prints, arXiv:1705.07057

  18. [26]

    C., & Murray, I

    Papamakarios, G., Sterratt, D. C., & Murray, I. 2018, arXiv:1805.07226 [cs, stat]

  19. [27]

    2017, in NIPS Autodiff Workshop

    Paszke, A., Gross, S., Chintala, S., et al. 2017, in NIPS Autodiff Workshop

  20. [28]

    2011, Journal of Machine Learning Research, 12, 2825

    Pedregosa, F., Varoquaux, G., Gramfort, A., et al. 2011, Journal of Machine Learning Research, 12, 2825

  21. [29]

    M., & Bonaca, A

    Price-Whelan, A. M., & Bonaca, A. 2018, ApJ, 863, L20, doi: 10.3847/2041-8213/aad7b5

  22. [30]

    D., Evans, D

    Riello, M., Angeli, F. D., Evans, D. W., et al. 2018, A&A, 616, A3, doi: 10.1051/0004-6361/201832712

  23. [31]

    P., Tollerud, E

    Robitaille, T. P., Tollerud, E. J., Greenfield, P., et al. 2013, A&A, 558, A33, doi: 10.1051/0004-6361/201322068

  24. [32]

    2009, A&A, 503, 165, doi: 10.1051/0004-6361/200911918

    Yakut, K., Zima, W., Kalomeni, B., et al. 2009, A&A, 503, 165, doi: 10.1051/0004-6361/200911918

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.