Pith. sign in

REVIEW 3 major objections 5 minor 36 references

Gravitational Potential from small-scale clustering in action space: Application to Gaia DR2

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Maximizing small-scale clustering in action space recovers the Milky Way's gravitational potential and measures its dark-matter halo from Gaia DR2.

desk verdict The action-space clustering idea is interesting and the simulations show the score works, but Eq. 17 is not the derived likelihood — it's a heuristic, so the quoted errors are not to be trusted. read the letter →

arxiv 1908.02336 v3 pith:DEL7EGNC submitted 2019-08-06 astro-ph.GA astro-ph.COhep-ph

Tianyi. Yang , Supranta Sarma Boruah , Niayesh. Afshordi This is my paper · ORCID
classification astro-ph.GAastro-ph.COhep-ph
keywords gravitationalpotentialactionspacetwo-pointcorrelationfunctionGaiaDR2darkmatterhalostellarstreamslikelihoodinferenceMilkyWay
topics Dark Matter
open problems Dark Matter
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 claims that the gravitational potential of a galaxy can be read off from how strongly its stars cluster in action space: the correct potential is the one that maximizes small-scale clustering, because actions computed with the true potential are conserved integrals that keep tidally disrupted structures coherent. It derives a likelihood for this clustering from the two-point correlation function of stellar pairs in action space, without needing equilibrium, circular-motion, or stream-membership assumptions. Applied to Gaia DR2 stars at 9–11 kpc and 11.5–15 kpc, the method estimates the dark-matter halo's radial force fraction and density slope as (f_h, α) = (0.391 ± 0.009, 1.835 ± 0.092) and (0.351 ± 0.012, 1.687 ± 0.079). If right, this gives a new, assumption-light probe of dark matter in galaxies using full six-dimensional phase space data.

What carries the argument

The central object is the normalized action-space distance D = sqrt[(ΔJ_R/σ_JR)^2 + (ΔJ_φ/σ_Jφ)^2 + (ΔJ_z/σ_Jz)^2] between stellar pairs, computed from actions estimated by the Stäckel approximation. The load-bearing identity is the likelihood ln L = N_pairs ∫_{-∞}^{ln Dmax} P(ln D) ln[1+ξ(ln D)] d ln D, where P(ln D) is the observed pair-distance distribution and ξ(ln D) = P/P_uniform − 1 is the two-point correlation function; this follows from marginalizing over a Gaussian random field that models small-scale clustering on top of a uniform background (Wick's theorem plus a mean-field approximation). This identity ties the gravitational potential to a purely statistical observable — pair clustering in action space — and it is the quantity maximized to infer f_h and α.

What would settle it

Run the same likelihood pipeline on a large cosmological or Milky-Way-scale simulation with a known gravitational potential and a realistic mix of disk and halo stars; if the recovered f_h and α deviate from the true values by more than the reported 1%–4% systematic errors, or if the best-fit parameters shift significantly when the sample is split by kinematics or by D_max, the uniform-background-plus-Gaussian-field assumption fails. A cheaper check is to compute the same likelihood on mock Gaia-like catalogs built from a known potential with deliberately non-Gaussian action-space counts and see whether the likelihood peak moves.

Watch

Extended reading notes

Core claim

The paper's central claim is that the correct gravitational potential maximizes small-scale clustering in action space, and that the likelihood for a potential is an integral over the two-point correlation function of the pair-distance distribution. Under the model that stars are a Poisson sampling of a uniform action-space background plus a correlated Gaussian random field, the log-likelihood reduces to N_pairs times the expectation of ln(1+ξ(ln D)) over pairs, where ξ is measured relative to a uniform distribution and D is the distance between pairs normalized by the action dispersions. The paper verifies on simulated tidal streams that this likelihood recovers the input halo parameters, and on Gaia DR2 it yields the quoted constraints on the dark-matter halo's radial force fraction f_h and power-law slope α of its density profile. It also shows that the potential maximizing the likelihood indeed gives the most compact action-space distribution for both simulations and real data, and that the implied circular-velocity curve is consistent with earlier measurements though 5–10% lower than maser and globular-cluster estimates.

Load-bearing premise

The likelihood assumes the distribution of stars in action space is a uniform background plus a Gaussian random field whose correlation function depends only on the normalized pair distance D; if real action-space structure is driven by something else, such as a strongly non-uniform disk background or unrelaxed components, the log-likelihood in Eq. 17 is not the correct likelihood and maximizing it can bias the recovered halo parameters.

Editorial extensions

If this is right

  • The method measures the Milky Way's dark-matter halo parameters without assuming virial equilibrium, circular motion, or identifying streams beforehand.
  • The circular velocity curve derived from the best-fit potential is consistent with measurements from Eilers et al. (2019) but 5–10% lower than maser- and globular-cluster-based curves.
  • The likelihood statistic scales with the number of stellar pairs, so future Gaia data releases with more stars will tighten constraints.
  • The same approach can be applied to other galaxies or to constrain the disk potential if its parameters are varied along with the halo.

Reading between the lines

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

  • If the method holds up across independent data, it offers a way to weigh dark matter in dwarf galaxies or the Milky Way's outer halo, where equilibrium assumptions are weakest.
  • The likelihood's dependence on the chosen maximum pair separation D_max is a hidden free choice; a principled model-selection criterion for D_max could remove the largest subjective element in the analysis.
  • One could replace the Gaussian-field assumption with a more general marked point process or Cox process to see whether the halo constraints shift substantially, providing a direct test of the paper's most load-bearing assumption.
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

3 major / 5 minor

Summary. The paper proposes and applies a new method to constrain the Milky Way gravitational potential by maximizing small-scale clustering of stars in action space, avoiding equilibrium or circular-motion assumptions. The authors derive a log-likelihood expressed through the two-point correlation function ξ(ln D) of action-space pair distances (Eq. 17), validate it on simulated tidal streams with known potentials (§5.1, Appendix C), and apply it to two Gaia DR2 radial samples, obtaining (f_h, α) = (0.391±0.009, 1.835±0.092) at 9–11 kpc and (0.351±0.012, 1.687±0.079) at 11.5–15 kpc. They translate the constraints into a circular-velocity curve and compare it with other Milky Way mass measurements. The paper's central claim is that the correct potential is the one maximizing the likelihood derived from action-space clustering.

Significance. If the derivation and statistical interpretation are correct, the method is a genuinely new probe of the Galactic potential: it uses full 6D phase-space information, does not require equilibrium, circular orbits, or identification of streams, and the simulation tests in §5.1 and Appendix C demonstrate that the proposed statistic can locate an input potential with systematic offsets of about 1% in f_h and 4% in α. The Gaia DR2 application and the comparison with independent rotation-curve measurements are valuable demonstrations, and the paper is clearly written with useful consistency checks. However, the claim that Eq. (17) is the likelihood of the assumed model is not supported by the derivation (major comment 1), and the final error budget involves post hoc and self-calibrated choices; the quoted halo constraints should therefore be regarded as conditional on these caveats until the derivation and calibration are repaired.

major comments (3)
  1. [Appendix A, Eqs. (A2)–(A6) and Eq. (17)] Appendix A, Eqs. (A2)–(A6) and Eq. (17): the Wick expansion given in Eq. (A5) is not the Gaussian expectation of the model in Eq. (A2). Expanding ∏_k (1+χ_{a_k}) and applying Wick's theorem gives a monomer–dimer sum, E[∏_k(1+χ_{a_k})] = 1 + Σ_{i<j} ξ_{ij} + Σ_{ij,kl disjoint} ξ_{ij}ξ_{kl} + ..., with unpaired stars allowed. This is not equal to Σ_{perfect pairings} ∏_{pairs}(1+ξ_pair): for N=4 stars the exact expectation is 1+Σξ+Σξξ, whereas the paper's expression evaluates to 3+2Σξ+Σξξ. The subsequent replacement of the log of the sum by an average over pairings of Σ ln(1+ξ) in Eq. (A6) is an uncontrolled mean-field approximation. As a result, Eq. (17) is not the likelihood of the stated generative model but a heuristic clustering score. The simulations in §5.1 and Appendix C test the score itself, so they support its use as an empirical estimator, but they do not validate the likelihood interpretation or the Gaussian error bars quoted in Tables 2–4.
  2. [Section 5.2 and Section 6 (choice of ln D_max)] Section 5.2 and Section 6 (choice of ln D_max): The final choice ln D_max = −1.14 is made after inspecting the likelihood constraints on the same Gaia data, including the jump in f_h near ln D_max ≈ −1.5 in the 11.5–15 kpc sample; the authors then select the value with the smallest uncertainties. This is a post hoc selection on the target data, so the reported posteriors and 68% intervals do not include the selection effect, and the stability criterion proposed in Section 6 is calibrated on the same curves. A pre-specified selection rule, or a mock test that repeats the entire D_max-selection procedure, is needed before the quoted precision can be taken at face value.
  3. [Section 6, Appendix C, and Tables 2–4] Section 6, Appendix C, and Tables 2–4: the systematic-error calibration does not validate the load-bearing assumption of a uniform background plus Gaussian fluctuations against the non-Gaussian, non-uniform backgrounds that real galactic data contain. The background simulation in Appendix C is constructed by adding Gaussian scatter to the action-space positions of the same Gaia sample under the assumed model, so it cannot detect violations of that assumption; the 7–11% shift in α between the |z|>1 kpc and no-cut samples (Tables 2–4) and the disk-star contamination discussed in Section 6 show that such violations are potentially material. The authors should test the estimator on realistic (e.g., cosmological/hydrodynamical) simulations of a Milky Way-like galaxy, or otherwise show that the score's peak is unbiased for non-Gaussian backgrounds; until then, the systematic errors in Table 4 (which are also partly calibrated on the observed cut/no-cut differences via Appendix D) are not end-to-end validated.
minor comments (5)
  1. [Section 4, after Eq. (17)] Section 4, after Eq. (17): the statement that Npairs is the total number of pairs, i.e., half of the number of stars in the sample, is inconsistent with the standard definition N(N−1)/2; please clarify whether Npairs is defined in the pairing-average convention.
  2. [Section 5.2 and Table 2 caption] Section 5.2 and Table 2 caption: the maximum-likelihood values quoted in the text (f_h=0.376, α=1.974 for 9–11 kpc) differ from the caption values (f_h=0.375, α=1.967); these should be reconciled.
  3. [Table 1 caption] Table 1 caption: the text says galactocentric radius and velocity are normalized by the solar radius value, but velocity cannot be normalized by a radius; specify the normalization convention for each column.
  4. [Footnote 6] Footnote 6: the claim that the disk is 'a mixture of correlated structures and uniform background' is not obviously compatible with the model assumption of a Poisson-sampled uniform background plus Gaussian fluctuations; this point should be expanded or reworded.
  5. [Figure 1 and Eq. (16)] Figure 1 and Eq. (16): please state explicitly how D_max enters the estimate of P(ln D)/D^3 in the plotted curve, so the reader can distinguish the raw pair-count distribution from the normalized correlation function.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the method is benchmarked on external simulations and no central claim reduces to its inputs by construction.

full rationale

The paper's central claim is that the correct Galactic potential maximizes a likelihood built from the two-point correlation function in action space (Eq. 17). This claim is not circular: the statistic is not defined in terms of the fitted parameters, and the method is tested on mock stellar streams with known input potentials (Section 5.1 and Appendices B, C), so the maximum-likelihood recovery of (f_h, alpha) is an independent benchmark rather than a restatement of inputs. Equation (18) identifies lnL/N_pairs with the relative entropy of the pair-distance distribution; that identity is definitional, but the physical assertion that the true potential maximizes this statistic is validated against simulations, not assumed. The citation to Afshordi et al. (2009) is a historical reference for the action-space clustering idea and is not load-bearing because the subsequent simulations and Gaia analysis stand independently. The choice of lnDmax as a stability/least-uncertainty point is a nuisance calibration, not a fitted parameter renamed as a prediction; Appendix D's systematic-error combination is likewise an honest propagation of cut-induced discrepancies. The most serious issue in the paper is mathematical, not circular: the Wick-contraction step in Appendix A does not appear to yield the perfect-matching product in Eq. (A5), so Eq. (17) may be an ansatz rather than a derived likelihood. That would be a correctness/falsifiability problem, but it is not a circularity, because Eq. (17) is not equivalent to the model by construction and is tested on external simulations.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The method relies on a specific statistical generative model for the action-space distribution, a fixed baryonic potential, and several hand-tuned scales. No new physical particles or forces are introduced. The central claim is therefore conditional on these modeling choices.

free parameters (3)
  • ln D_max = -1.14
    Maximum pair separation in action space used in the likelihood integral; chosen post hoc where parameter constraints are least uncertain and stable across lnD_max in real data (Section 5.2, Fig. 4).
  • sigma_sys,max (f_h) = 0.02
    Unknown systematic error assigned to f_h in the late-time combination of cut and uncut samples in Appendix D.
  • sigma_sys,max (alpha) = 0.2
    Unknown systematic error assigned to alpha in Appendix D; chosen based on the difference between cut and uncut samples.
assumptions (6)
  • domain assumption The stellar distribution in action space is a uniform background plus a Gaussian random field with correlation function that depends only on normalized distance D.
    This is the central modeling assumption of the likelihood derivation in Appendix A; it allows the Poisson sampling equation (Eq. A2) and the likelihood formula (Eq. 17).
  • domain assumption Stars are Poisson samples of the underlying action-space density.
    Used in Appendix A to relate the observed discrete star counts to the continuous density field.
  • domain assumption Actions are computed with the Stäckel approximation via galpy and are conserved for these stars.
    Section 2.2.1; approximation errors are expected to be small but are only calibrated on simulations.
  • domain assumption The bulge and disk gravitational potentials are known and fixed to MWPotential2014 parameter values; only the dark matter halo is varied.
    Section 2.1; this assumption is load-bearing because any error in the fixed baryonic potential is absorbed into the inferred halo parameters, and the paper finds a systematic shift when changing the |z| cut, attributed partly to the disk model.
  • ad hoc to paper The prior ranges 0.5 < alpha < 2.5 and 0.25 < f_h < 0.55 are appropriate.
    Equation (20); the prior bounds affect the median estimates and credible intervals in the multi-modal posteriors.
  • ad hoc to paper The maximum pair separation scale Dmax can be chosen such that the likelihood is stable; the chosen value is valid for both radial bins.
    The paper's criterion for selecting lnDmax is empirical stability rather than a derivable quantity; the same lnDmax = -1.14 is applied to both samples.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Gravitational Potential from small-scale clustering in action space: Application to Gaia DR2." pith.science (2026). https://pith.science/paper/DEL7EGNC

@misc{pith2026190802336,
  author       = {Pith},
  title        = {Pith review of: Gravitational Potential from small-scale clustering in action space: Application to Gaia DR2},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DEL7EGNC}},
  note         = {Machine review of arXiv:1908.02336}
}
abstract

Most measurements of mass in Astronomy that use kinematics of stars or gas rely on assumptions of equilibrium that are often hard to verify. Instead, we develop a novel idea that uses the clustering in action space, as a probe of underlying gravitational potential: the correct potential should maximize small-scale clustering in the action space. We provide a first-principle derivation of likelihood using the two-point correlation function in action space, and test it against simulations of stellar streams. We then apply this method to the 2nd data release of Gaia, and use it to measure the radial force fraction $f_h$ and logarithmic slope $\alpha$ of dark matter halo profile. We investigate stars within 9-11 kpc and 11.5-15 kpc from Galactic centre, and find $(f_h,\alpha)= (0.391\pm 0.009, 1.835\pm 0.092) $ and $(0.351\pm 0.012,1.687\pm 0.079)$, respectively. We also confirm that the set of parameters that maximize the likelihood function do correspond to the most clustering in the action space. The best-fit circular velocity curve for Milky Way potential is consistent with past measurements (although it is $\sim$ 5-10\% lower than previous methods that use masers or globular clusters). Our work provides a clear demonstration of the full statistical power that lies in the full phase space information, relieving the need for {\it ad hoc} assumptions such as virial equilibrium, circular motion, or steam-finding algorithms.

Figures

Figures reproduced from arXiv: 1908.02336 by the authors.

Figure 1
Figure 1. ln h P(ln D) D3 i versus ln(D) calculated by using the Gaia DR2 real data from galactocentric radius 11.5-15 kpc with fh = 0.34, α = 1.66. Here D is the normalized distance of pairs of stars in the action space, while P(ln D) is its probability density over all pairs of Gaia DR 2 within our sample. The correlation function computed using Gaia-Enceladus data (Myeong et al. 2018b) is over-plotted on the same figure (o… view at source ↗
Figure 2
Figure 2. Top panel: likelihood test and error bar plot for case [fh = 0.35, α = 1.70]. Bottom panel: likelihood test and error bar plot for case [fh = 0.35, α = 2.0]. The maximum likelihood gives constraints on the parameters fh = 0.35, α = 1.63 for the first case and fh = 0.35, α = 1.95 for the second case. The initial set of parameter is indicated as black plus sign on the likelihood plot for either case. Error bars are de… view at source ↗
Figure 3
Figure 3. The galactocentric radius and tangential velocity distribution in cylindrical coordinates for the selected Gaia DR2 catalogue. Calculations are all conducted in cylindrical coordinate [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Top panel: likelihood test and error bar plot using sample stars from 9-11 kpc. Bottom panel: likelihood test and error bar plot using sample stars from 11.5-15 kpc. The maximum likelihood gives constraints on the parameters fh = 0.376, α = 1.974 for the first case and…
Figure 5
Figure 5. Figure 5: The posterior distribution of fh (upper panel) and α (lower panel) at three different values of Dmax calculated using stars with radial coverage from 11.5-15 kpc. Unlike simulation, the appearance of multiple peaks is obvious in the probability distribution and the pea…
Figure 6
Figure 6. Figure 6: Stellar distribution in the J˜R and J˜φ 2D projected plane varying with different choices of potential, where J˜R and J˜φ are defined as JR/σJR and Jφ/σJφ , i.e. radial and angular action variables normalized by their standard deviations over all stars in the sample. F…
Figure 7
Figure 7. Figure 7: Correlation function P(ln D) D3 as a function of the distance in the action space in natural logarithm scale. The purpose of this figure is to check how the two-point correlation function varies with different choices of potential and whether it is extremized around th…
Figure 8
Figure 8. Figure 8: Error bar plot using stars from 9 to 11 kpc (top) and 11.5 to 15 kpc (bottom) without selection cuts. Green square points shows the constraints to the parameters by directly finding the maximum from the likelihood plot, and at ln Dmax = -1.14, this gives fh = 0.392 and…
Figure 9
Figure 9. Figure 9: The total rotation curves calculated from different potential models. Results for this work are indicated as purple solid line (for the sample with selection cuts) and black solid line (or the sample without selection cuts). The shaded area indicates the 1-σ uncertaint…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 7 canonical work pages

  1. [1]

    Afshordi N., Mohayaee R., Bertschinger E., 2009, @doi [ ] 10.1103/PhysRevD.79.083526 , https://ui.adsabs.harvard.edu/abs/2009PhRvD..79h3526A 79, 083526

    work page doi:10.1103/physrevd.79.083526 2009

  2. [2]

    L., Bailer-Jones C

    Astraatmadja T. L., Bailer-Jones C. A. L., 2016, @doi [ ] 10.3847/0004-637X/832/2/137 , https://ui.adsabs.harvard.edu/abs/2016ApJ...832..137A 832, 137

    doi:10.3847/0004-637x/832/2/137 2016

  3. [3]

    Binney J., 2010, @doi [ ] 10.1111/j.1365-2966.2009.15845.x , https://ui.adsabs.harvard.edu/abs/2010MNRAS.401.2318B 401, 2318

    arXiv 2010

  4. [5]

    J., 2016, @doi [ ] 10.1093/mnras/stv2734 , https://ui.adsabs.harvard.edu/abs/2016MNRAS.456.1982B 456, 1982

    Binney J., McMillan P. J., 2016, @doi [ ] 10.1093/mnras/stv2734 , https://ui.adsabs.harvard.edu/abs/2016MNRAS.456.1982B 456, 1982

    doi:10.1093/mnras/stv2734 2016

  5. [6]

    F., Kere s D., 2017, @doi [ ] 10.3847/1538-4357/aa7d0c , https://ui.adsabs.harvard.edu/abs/2017ApJ...845..101B 845, 101

    Bonaca A., Conroy C., Wetzel A., Hopkins P. F., Kere s D., 2017, @doi [ ] 10.3847/1538-4357/aa7d0c , https://ui.adsabs.harvard.edu/abs/2017ApJ...845..101B 845, 101

    doi:10.3847/1538-4357/aa7d0c 2017

  6. [7]

    Bovy J., 2014, @doi [ ] 10.1088/0004-637X/795/1/95 , http://adsabs.harvard.edu/abs/2014ApJ...795...95B 795, 95

    doi:10.1088/0004-637x/795/1/95 2014

  7. [8]

    Bovy J., 2015, @doi [ ] 10.1088/0067-0049/216/2/29 , http://adsabs.harvard.edu/abs/2015ApJS..216...29B 216, 29

    doi:10.1088/0067-0049/216/2/29 2015

  8. [9]

    Bovy J., Rix H.-W., 2013, @doi [ ] 10.1088/0004-637X/779/2/115 , http://adsabs.harvard.edu/abs/2013ApJ...779..115B 779, 115

    doi:10.1088/0004-637x/779/2/115 2013

Show all 36 references
  1. [10]

    R., Hogg D

    Buckley M. R., Hogg D. W., Price-Whelan A. M., 2019, arXiv e-prints, https://ui.adsabs.harvard.edu/abs/2019arXiv190700987B p. arXiv:1907.00987

    2019 arXiv

  2. [11]

    S., Boylan-Kolchin M., 2017, @doi [ ] 10.1146/annurev-astro-091916-055313 , https://ui.adsabs.harvard.edu/abs/2017ARA&A..55..343B 55, 343

    Bullock J. S., Boylan-Kolchin M., 2017, @doi [ ] 10.1146/annurev-astro-091916-055313 , https://ui.adsabs.harvard.edu/abs/2017ARA&A..55..343B 55, 343

    2017 doi

  3. [12]

    arXiv:1804.09369

    Cropper M., et al., 2018, arXiv e-prints, https://ui.adsabs.harvard.edu/abs/2018arXiv180409369C p. arXiv:1804.09369

    2018 arXiv

  4. [13]

    Del Popolo A., Le Delliou M., 2017, @doi [Galaxies] 10.3390/galaxies5010017 , https://ui.adsabs.harvard.edu/abs/2017Galax...5...17D 5, 17

    2017 doi

  5. [14]

    W., Rix H.-W., Ness M

    Eilers A.-C., Hogg D. W., Rix H.-W., Ness M. K., 2019, @doi [ ] 10.3847/1538-4357/aaf648 , https://ui.adsabs.harvard.edu/abs/2019ApJ...871..120E 871, 120

    2019 doi

  6. [15]

    Gaia Collaboration et al., 2016, @doi [ ] 10.1051/0004-6361/201629272 , https://ui.adsabs.harvard.edu/abs/2016A&A...595A...1G 595, A1

    2016 doi

  7. [16]

    Gaia Collaboration et al., 2018, @doi [ ] 10.1051/0004-6361/201833051 , https://ui.adsabs.harvard.edu/abs/2018A&A...616A...1G 616, A1

    2018 doi

  8. [17]

    Gravity Collaboration et al., 2018, @doi [ ] 10.1051/0004-6361/201833718 , https://ui.adsabs.harvard.edu/abs/2018A&A...615L..15G 615, L15

    2018 doi

  9. [18]

    A., Tian H., Sales L

    Helmi A., Veljanoski J., Breddels M. A., Tian H., Sales L. V., 2017, @doi [ ] 10.1051/0004-6361/201629990 , https://ui.adsabs.harvard.edu/abs/2017A&A...598A..58H 598, A58

    2017 doi

  10. [19]

    A., 1951, @doi [Ann

    Kullback S., Leibler R. A., 1951, @doi [Ann. Math. Statist.] 10.1214/aoms/1177729694 , 22, 79

    1951

  11. [20]

    Magorrian J., 2014, @doi [ ] 10.1093/mnras/stt2031 , https://ui.adsabs.harvard.edu/abs/2014MNRAS.437.2230M 437, 2230

    2014 doi

  12. [21]

    J., 2017, @doi [ ] 10.1093/mnras/stw2759 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.465...76M 465, 76

    McMillan P. J., 2017, @doi [ ] 10.1093/mnras/stw2759 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.465...76M 465, 76

    2017 doi

  13. [22]

    C., White S., 2010, Galaxy Formation and Evolution

    Mo H., van den Bosch F. C., White S., 2010, Galaxy Formation and Evolution

    2010

  14. [23]

    C., Evans N

    Myeong G. C., Evans N. W., Belokurov V., Amorisco N. C., Koposov S. E., 2018a, @doi [ ] 10.1093/mnras/stx3262 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.475.1537M 475, 1537

    doi

  15. [24]

    C., Evans N

    Myeong G. C., Evans N. W., Belokurov V., Sand ers J. L., Koposov S. E., 2018b, @doi [ ] 10.3847/2041-8213/aad7f7 , https://ui.adsabs.harvard.edu/abs/2018ApJ...863L..28M 863, L28

    doi

  16. [25]

    F., Frenk C

    Navarro J. F., Frenk C. S., White S. D. M., 1996, @doi [ ] 10.1086/177173 , https://ui.adsabs.harvard.edu/abs/1996ApJ...462..563N 462, 563

    1996 doi

  17. [26]

    arXiv:1807.02519

    Necib L., Lisanti M., Belokurov V., 2018, arXiv e-prints, https://ui.adsabs.harvard.edu/abs/2018arXiv180702519N p. arXiv:1807.02519

    2018 arXiv

  18. [27]

    E., Walker M

    Pe \ n arrubia J., Koposov S. E., Walker M. G., 2012, @doi [ ] 10.1088/0004-637X/760/1/2 , https://ui.adsabs.harvard.edu/abs/2012ApJ...760....2P 760, 2

    2012 doi

  19. [28]

    Sanders J., 2012, @doi [ ] 10.1111/j.1365-2966.2012.21698.x , https://ui.adsabs.harvard.edu/abs/2012MNRAS.426..128S 426, 128

    2012

  20. [29]

    L., Binney J., 2013, @doi [ ] 10.1093/mnras/stt816 , https://ui.adsabs.harvard.edu/abs/2013MNRAS.433.1826S 433, 1826

    Sanders J. L., Binney J., 2013, @doi [ ] 10.1093/mnras/stt816 , https://ui.adsabs.harvard.edu/abs/2013MNRAS.433.1826S 433, 1826

    2013 doi

  21. [30]

    E., Helmi A., Hogg D

    Sanderson R. E., Helmi A., Hogg D. W., 2015, @doi [ ] 10.1088/0004-637X/801/2/98 , https://ui.adsabs.harvard.edu/abs/2015ApJ...801...98S 801, 98

    2015 doi

  22. [31]

    E., Hartke J., Helmi A., 2017, @doi [ ] 10.3847/1538-4357/aa5eb4 , http://adsabs.harvard.edu/abs/2017ApJ...836..234S 836, 234

    Sanderson R. E., Hartke J., Helmi A., 2017, @doi [ ] 10.3847/1538-4357/aa5eb4 , http://adsabs.harvard.edu/abs/2017ApJ...836..234S 836, 234

    2017 doi

  23. [32]

    Sch \"o nrich R., Aumer M., 2017, @doi [ ] 10.1093/mnras/stx2189 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.472.3979S 472, 3979

    2017 doi

  24. [33]

    Sch \"o nrich R., McMillan P., Eyer L., 2019, @doi [ ] 10.1093/mnras/stz1451 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.487.3568S 487, 3568

    2019 doi

  25. [34]

    Vasiliev E., 2019, @doi [ ] 10.1093/mnras/stz171 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.484.2832V 484, 2832

    2019 doi

  26. [35]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence output.state after.block = 'skip output.state before.all = 'skip after.sentence 'output.stat...

  27. [36]

    N., 2013, Journal of Improbable Astronomy, 1, 1

    Author A. N., 2013, Journal of Improbable Astronomy, 1, 1

    2013

  28. [37]

    Others S., 2012, Journal of Interesting Stuff, 17, 198

    2012

Pith tools

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