Dynamical mass distribution and velocity structure of the Galactic centre

A. Feldmeier-Krause; E. Gallego-Cano; G. van de Ven; N. Neumayer; T. Ver\v{s}i\v{c}

REVIEW 3 major objections 4 minor 80 references

Dynamical mass distribution and velocity structure of the Galactic centre

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper uses discrete chemo-dynamical Jeans models of 4,600 stars to pin Sgr A*'s mass at $(4.35 \pm 0.24) \times 10^6 \, M_{\odot}$ and finds the enclosed mass at 5-30 pc is substantially lower than previous dynamical estimates.

desk verdict A careful, honest dynamical modeling paper with genuinely new data at 5-30 pc and a lower mass profile that is plausible but rests on untested axisymmetric Jeans assumptions. read the letter →

arxiv 2506.06014 v1 pith:ZW3AUW6Y submitted 2025-06-06 astro-ph.GA

A. Feldmeier-Krause , T. Veršič , G. van de Ven , E. Gallego-Cano , N. Neumayer This is my paper

classification astro-ph.GA

keywords GalacticcentrenuclearstarclusterstellardiscdiscreteJeansmodellingsupermassiveblackholemassmetallicityenclosedprofiledynamics

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 measure the gravitational potential of the Milky Way's central ~200 pc using the discrete motions and chemistries of individual giant stars, rather than binned averages. Fitting axisymmetric Jeans models to 4,600 line-of-sight velocities (3,567 with proper motions), it recovers a supermassive black hole mass of $(4.35 \pm 0.24) \times 10^6 \, M_{\odot}$ for Sgr A*, consistent with stellar-orbit measurements, and derives an enclosed mass profile for the inner ~60 pc. The key result is that this profile is lower than earlier dynamical estimates throughout the 5-30 pc region, in the most extreme case by about 75 percent at 15 pc. The models also separate two stellar populations by metallicity: a dominant metal-rich component with strong rotation and mild tangential anisotropy, consistent with in situ formation from bar-driven gas inflow, and a metal-poor, radially anisotropic component that points to a different origin.

What carries the argument

The load-bearing tool is the discrete axisymmetric Jeans anisotropic MGE (JAM) model, in which the collisionless Boltzmann equation is solved for the first and second velocity moments under the assumptions of axisymmetry and a cylindrically aligned velocity ellipsoid. The gravitational potential is built from a multi-Gaussian expansion (MGE) of the stellar density map plus a point-like Sgr A*, with free parameters (black hole mass, mass-to-light ratio $\Upsilon$, velocity anisotropy $\beta_z$, rotation parameter $\kappa$, and background fraction $\epsilon$) sampled by MCMC. The novel part is the discrete likelihood: each star contributes a dynamical, spatial, and chemical probability, so contaminating bar stars and multiple stellar populations can be separated star-by-star instead of being baked into binned velocity statistics. The two-population version adds Gaussian metallicity distributions and a radial population fraction, allowing the metal-rich and metal-poor components to be assigned distinct $\beta_z$ and $\kappa$ values, and this machinery is what lets the paper attribute the lower 5-30 pc mass to a cleaner treatment of contamination.

What would settle it

Measure stellar orbits or accelerations in the 5-30 pc annulus with future astrometry, for example by combining high-precision proper motions over a decade-long baseline with line-of-sight velocities, and compare the enclosed mass implied by those orbits with the Jeans-model profile; any deviation larger than the quoted uncertainties would falsify the axisymmetric result.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims that discrete chemo-dynamical Jeans modelling of the Galactic centre yields a robust, data-driven answer to two open questions: the mass of Sgr A* and the enclosed mass of the nuclear region over the previously sparsely sampled 5-30 pc. The supermassive black hole mass is $(4.35 \pm 0.24) \times 10^6 \, M_{\odot}$, matching the stellar-orbit value, and the enclosed mass profile comes out lower than the literature, reaching a roughly 75 percent deficit at 15 pc relative to one widely used model. The authors argue this lower mass is credible because earlier studies had few stars in that annulus and may have been contaminated by high-velocity bar stars, which their model absorbs into an explicit background component. They also find that a dark matter component contributes at most a few percent to the enclosed mass within the nuclear star cluster and that a radially varying mass-to-light ratio is not required.

Load-bearing premise

The load-bearing premise is that the nuclear star cluster and nuclear stellar disc are axisymmetric with a velocity ellipsoid aligned to cylindrical polar coordinates; if an inner bar is present or the velocity ellipsoid is tilted, streaming motions could be misread as dispersion and bias the inferred mass profile.

Editorial extensions

If this is right

The supermassive black hole mass from discrete stellar dynamics now agrees with stellar-orbit fits, so the two independent methods no longer disagree at the Galactic centre.
Enclosed masses in the 5-30 pc region need to be revised downward, with the largest correction (about 75 percent at 15 pc) occurring where previous data were sparse.
Dark matter contributes little to the mass budget inside about 30 pc; a cored profile contributes below 0.5 percent of the enclosed mass at 5 pc in the 84th-percentile upper limit.
The nuclear star cluster is dominated (more than 90 percent) by a metal-rich population that is close to an isotropic rotator, supporting in situ formation from bar-transported gas, while the metal-poor population's radial anisotropy points to an accretion origin.
A constant mass-to-light ratio $\Upsilon \approx 0.7$ in the 4.5 $\mu$m band fits the data, so a radially varying stellar population is not required to explain the dynamics.

Reading between the lines

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

If the lower 5-30 pc mass profile survives, it would bring the dynamically inferred baryonic mass of the Galactic centre closer to the stellar-mass budget traced by the density map, easing a long-standing tension between dynamics and photometry in this region.
A testable extension would be to feed the same discrete likelihood framework with existing bulge control-field data to push the model beyond the current 33 pc limit and locate the predicted maximum of $V_{\rm LOS}/\sigma_{\rm LOS}$ in the nuclear stellar disc.
If the axisymmetry assumption is wrong and an inner bar exists, the derived mass deficit at 5-30 pc could partly be an artefact of misattributing streaming motions; a triaxial orbit-based model with the same discrete data would settle this.
Using elemental abundances such as $\alpha$-element ratios alongside [M/H] could sharpen the two-population separation; a distinct abundance pattern in the metal-poor component would strengthen the accretion origin beyond kinematics alone.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

The thing to know: this is a solid paper that gives the first discrete chemo-dynamical Jeans model of the Galactic centre with two stellar populations and a bar-background component, using 4,600 line-of-sight velocities (3,567 with proper motions) extending to l=33 pc. The headline result is an enclosed mass profile for 5-30 pc that is up to ~75% lower at 15 pc than earlier models, and a fitted SMBH mass of (4.35 +/- 0.24) x 10^6 solar masses that matches the stellar-orbit value.

What it does well: the data genuinely fill a gap, as previous work had only a few hundred stars in the 5-30 pc region and many of those were likely contaminated. The forward model handles contamination with a background component, and the authors run sensible robustness checks (free M_BH, radial M/L variation, cored and cusped DM). The one-population results are stable across these choices, and the M_BH agreement is a strong external check that the potential is not wildly off. The two-population decomposition is more speculative, but the Anderson-Darling tests showing that the low-[M/H] sample is distinct from the background give it some support. The paper is also candid: Section 6 explicitly notes the axisymmetric and velocity-ellipsoid-alignment assumptions and calls for triaxial orbit-based models.

The soft spots are real but proportionate. The axisymmetric Jeans assumption with a cylindrically aligned velocity ellipsoid is the main one; if the NSD has an inner bar or the ellipsoid is tilted, the rotation/dispersion decomposition can be biased, and the bias would tend to lower the inferred mass. The authors flag this but do not test it: there are no east-west residual maps or a direct comparison of the model's V_l field to the data, which would be a natural diagnostic. The claim that the low-[M/H] population's radial anisotropy (beta_z=+0.64) could absorb such a tilt is a plausible conjecture, but it is not a demonstrated bias. A more concrete issue is reproducibility: one of the input catalogues is unpublished (Xu et al., in prep.) and the code is not released, so the 5-30 pc result cannot be independently checked as it stands.

For a reader, this is a worthwhile paper for anyone working on Galactic centre dynamics or nuclear star clusters, and it deserves a serious referee. If I were the editor, I would send it out and ask for the combined data and code to be made available, and for a residual analysis or at least a stiffer caveat on the 5-30 pc mass. The core result is plausible, the models are standard for the field, and the paper is honest about its limits. I would accept it after moderate revision.

Referee Report

3 major / 4 minor

Summary. The paper constructs discrete chemo-dynamical axisymmetric Jeans models of the Galactic centre using 4,600 stars with line-of-sight velocities and metallicities, 3,567 of which also have proper motions. In one-population fits with a background component, the authors recover a Sgr A* mass of (4.35 +/- 0.24) x 10^6 M_sun, consistent with stellar-orbit measurements, and find that dark matter and radial mass-to-light ratio variations are negligible. In two-population fits, they separate a dominant high-[M/H] population that is mildly tangentially anisotropic and rotationally supported from a subdominant low-[M/H] population that is radially anisotropic and has weaker rotation. The enclosed mass profile derived for the inner ~60 pc is significantly lower in the 5-30 pc region than several previous dynamical estimates, by up to ~75% at ~15 pc. The paper attributes the difference to previously sparse data and possible contamination in earlier samples.

Significance. If the central result is correct, the paper fills an important observational gap in the 5-40 pc region of the Galactic centre, where the potential was previously constrained only by interpolation or sparse samples. The recovery of M_BH matching stellar-orbit values from discrete Jeans modelling is a valuable external validation. The chemically separated two-population analysis is also novel for the Galactic centre and provides a dynamical interpretation of the metal-rich and metal-poor components. The main result, however, rests on the axisymmetric Jeans assumption and the cylindrical alignment of the velocity ellipsoid, which the authors themselves identify as a limitation in Sect. 6. The paper is careful in its robustness tests of the gravitational potential (varying M_BH, DM, and M/L profiles), and it presents posterior distributions and model maps in the appendices. The external check against the GRAVITY stellar-orbit mass is a genuine strength, since many earlier dynamical models underestimated M_BH.

major comments (3)

[Sect. 3 and Sect. 6] The central claim of a lower enclosed mass at 5-30 pc depends critically on the assumptions stated at the start of Sect. 3: an axisymmetric potential and a velocity ellipsoid aligned with cylindrical coordinates (v_R v_z = 0). In a triaxial or barred potential, or with a tilted velocity ellipsoid, the Jeans equations solved by CJAM will misallocate signal between rotation (kappa) and dispersion (sigma), and the inferred mass at 5-30 pc could be biased. This is not a hypothetical concern: the paper itself notes in Sect. 6 that an inner bar cannot be excluded and that triaxial orbit-based models are needed to relax the ellipsoid-alignment assumption. To make the central claim robust, the authors should at least test the assumption internally, for example by showing residual maps of the best-fit model in V_LOS and proper motions as a function of Galactic longitude and latitude (e.g., east-west asymmetries), or by adding a term for v_R v_z in the Jeans equations. Without such a test, the 5-30 pc mass profile remains conditional on a symmetry assumption that the paper acknowledges may be violated.
[Sect. 3.7 and Sect. 5] The two-population kinematics and the membership probabilities are co-estimated in the same likelihood (Eq. 7). The low-[M/H] population is therefore selected using probabilities that depend on the same kinematic model that then defines its dynamical properties (beta_z^2 = +0.64, kappa^2 = -0.59). The Anderson-Darling tests reported in Sect. 6 compare samples assigned by the model (P_2 >= 0.5 vs. P_bg > 0.5), so they cannot independently demonstrate that the low-[M/H] stars form a distinct dynamical population. This circularity should be acknowledged explicitly, or the tests should be repeated with an independently defined sample (e.g., a metallicity cut) to verify that the inferred kinematic differences do not arise purely from the model's own classification.
[Sect. 3.4 and Sect. 4.4] The background component is modelled as a Gaussian with a fixed mean velocity of 0 km/s and a fixed dispersion of 130 km/s, adopted from Portail et al. (2017). The fitted background fraction (epsilon ~2.4% in the one-population model, ~1.7% in the two-population model) is small, but the treatment of background stars can directly affect the inferred velocity dispersion of the member populations and therefore the derived mass profile. The paper does not test the sensitivity of the enclosed mass to the assumed sigma_bg or to the Gaussian form of the background velocity distribution. A simple robustness test varying sigma_bg over a plausible range (e.g., 100-160 km/s) or using a non-Gaussian background would help establish that the 5-30 pc mass deficit is not an artifact of the background subtraction.

minor comments (4)

[Eq. (6)] The denominator in Eq. (6) contains an unmatched parenthesis; it should read sqrt(2*pi*((sigma_Z^k)^2 + (delta Z_i)^2)).
[Sect. 2.2] The text refers to 'Libralato et al. (2020)' in several places, but the reference list gives Libralato et al. (2020) with the correct year; however, in Fig. 1 the caption says 'Libralato et al. (2020)' while the text in Sect. 2.2 mentions 'Libralato et al. (2020)' and 'Libralato et al. (2021)' in the bibliography. Please ensure the citation year in the text and figure caption matches the reference list.
[Sect. 4.4] The sentence 'The rotation parameter kappa is also consistent among the different models' is somewhat misleading because Table 2 shows that kappa_0 has a broad posterior (e.g., -0.26 +/- 0.35), while kappa_inf is well constrained. It would be clearer to state that the outer rotation is robustly constrained while the inner rotation is not.
[Sect. 6] When comparing with the Sormani et al. (2022) profile, the text says the discrepancy is ~1.5 x 10^7 M_sun at ~17 pc and then notes the mass is '<=60% lower (highest discrepancy at 12 pc)' when a black hole mass is added. The exact comparison would be easier to follow if the authors stated which of their models is being compared (e.g., the two-population or the free-M_BH model) in the same sentence.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: forward Jeans modeling with external benchmarks.

full rationale

The central mass claims are derived from a forward axisymmetric Jeans model (CJAM, Watkins et al. 2013) in which the gravitational potential is constructed from Sgr A*, a fitted stellar MGE, optional dark matter, and a fitted mass-to-light ratio, with the kinematics entering only through the discrete likelihoods in Eqs. (3)-(8). The black hole mass is fitted with a uniform prior [2,6] x 10^6 Msun in Sect. 4.1 and then compared with the independent GRAVITY stellar-orbit value, so the agreement is an external consistency check rather than an input that guarantees the result. The lower enclosed mass at 5-30 pc is likewise a fitted model output compared with independent literature profiles (Sormani et al. 2020, 2022; Feldmeier-Krause et al. 2017b), so it is externally benchmarked and not a renaming of the input density. The two-population membership probabilities are co-estimated with the population kinematics, meaning the Anderson-Darling test on the model-selected samples in Sect. 6 is not an independent validation of the population split; however, this is a secondary interpretive check and does not enter the one-population mass profile that carries the paper's main quantitative conclusions. The axisymmetry and velocity-ellipsoid alignment assumptions are explicitly stated as assumptions (Sect. 3 and Sect. 6) and are potential biases, but they are not hidden circular inputs.

Assumptions & free parameters 24 free parameters · 9 assumptions · 2 invented entities

All model outputs are conditional on the axisymmetric Jeans framework, the red-giant light tracing stellar mass, the fixed background velocity distribution, and the adopted distance. The one-population mass profile has one external check, the fitted M_BH matching the stellar-orbit value, but the two-population properties are co-estimated with the membership that defines them. No new physical entity such as a particle or force is introduced; the two population components are data-driven constructs.

free parameters (24)

Background star fraction epsilon = 2.4 +/- 0.2% (one-pop), 1.71 +/- 0.16% (two-pop)
Fitted in every model as the fraction of contaminant stars from the bar background; it also absorbs part of any high-velocity outliers.
Inner anisotropy beta_0 = -0.15 +0.11 -0.19 (free-M run)
Inner value of the radially varying anisotropy profile; fitted with a modified Osipkov-Merritt profile.
Outer anisotropy beta_inf = -0.15 +0.08 -0.09 (free-M run)
Outer asymptotic anisotropy value; fitted jointly with beta_0.
Anisotropy transition radius R_beta = 58 arcsec +275 -41 (free-M run)
Transition radius of the anisotropy profile; weakly constrained in all runs.
Inner rotation kappa_0 = -0.26 +0.35 -0.33 (free-M run)
Inner value of the rotation parameter; consistent with zero in the central parsec.
Outer rotation kappa_inf = -1.10 +/- 0.05 (free-M run)
Outer rotation parameter; strongly constrained near -1.1.
Rotation transition radius R_kappa = 17 arcsec +8 -4 (free-M run)
Transition radius of the rotation profile between inner and outer values.
Constant mass-to-light ratio Upsilon = 0.72 +/- 0.03 (free-M run); 0.75 +/- 0.02 (two-pop)
Mass-to-light conversion in the 4.5 micron band applied to the MGE stellar density; common to both populations.
Sgr A* mass M_BH = 4.35 +0.24 -0.23 x 10^6 M_sun (free-M run); fixed to 4.3 elsewhere
Supermassive black hole mass fitted in one run and checked against the stellar-orbit value; fixed in all other runs.
DM scale density log(rho_S) = -0.57 +0.16 -0.37 (cusp); 1.62 +0.15 -0.56 (core)
NFW scale density fitted with fixed scale radius and fixed inner slope; poorly constrained, with a firm upper limit.
Inner M/L Upsilon_0 = 0.90 +0.28 -0.12
Central mass-to-light ratio in the radially varying M/L run.
Outer M/L Upsilon_inf = 0.68 +/- 0.04
Outer asymptotic mass-to-light ratio in the radially varying M/L run.
M/L transition radius R_Upsilon = 51 arcsec +280 -35
Transition radius for the M/L profile; very weakly constrained.
Inner high-M/H fraction h_0 = 0.98 +/- 0.01
Central fraction of the total stellar density assigned to the high-[M/H] population.
Outer high-M/H fraction h_inf = 0.91 +/- 0.01
Outer fraction of the total stellar density assigned to the high-[M/H] population.
Fraction transition radius R_h = 45 arcsec +32 -16
Transition radius for the population fraction profile.
High-M/H mean metallicity Z1_0 = 0.34 +/- 0.01 dex
Mean metallicity of the high-[M/H] Gaussian component.
High-M/H metallicity dispersion sigma1_Z = 0.30 +/- 0.01 dex
Metallicity dispersion of the high-[M/H] Gaussian component.
Low-M/H mean metallicity Z2_0 = -0.78 +/- 0.05 dex
Mean metallicity of the low-[M/H] Gaussian component.
Low-M/H metallicity dispersion sigma2_Z = 0.27 +/- 0.03 dex
Metallicity dispersion of the low-[M/H] Gaussian component.
Population 1 anisotropy beta_z^1 = -0.06 +0.03 -0.04
Radially constant anisotropy of the high-[M/H] population.
Population 2 anisotropy beta_z^2 = 0.64 +0.06 -0.08
Radially constant anisotropy of the low-[M/H] population.
Population 1 rotation kappa^1 = -1.07 +/- 0.04
Radially constant rotation parameter of the high-[M/H] population.
Population 2 rotation kappa^2 = -0.59 +0.16 -0.18
Radially constant rotation parameter of the low-[M/H] population.

assumptions (9)

domain assumption The stellar systems satisfy the collisionless Boltzmann equation and are in a steady state.
Invoked at the start of Sect. 3 as a basic assumption of the JAM/CJAM formalism; if the region is not in steady state, the inferred potential is biased.
domain assumption The stellar systems are axisymmetric and the velocity ellipsoid is aligned with cylindrical polar coordinates.
Stated as basic assumptions in Sect. 3; the model has no freedom for non-axisymmetric streaming, and the paper flags the possible inner bar in Sect. 6.
domain assumption The Ks-band red-giant surface density profile traces the total stellar mass distribution with a single mass-to-light ratio.
Used in Sect. 3.1 to convert the observed MGE (Table 1) into the stellar contribution to the potential; dark remnants or a varying initial mass function would break this.
domain assumption The system is observed edge-on, so the MGE is deprojected at inclination i=90 degrees.
Assumed in Sect. 3.1; at this inclination the deprojection is unique, but a different inclination would change the inferred intrinsic density.
domain assumption The bar background has a uniform surface density, a Gaussian velocity distribution with mean 0 and dispersion 130 km/s, and a fixed Gaussian metallicity with mean 0.07 dex and dispersion 0.3 dex.
Assumed in Sects. 3.4-3.6; if the bar contamination has a different velocity distribution, the inferred membership and the enclosed mass could be biased.
domain assumption Each modelled stellar population has a Gaussian [M/H] distribution, and the two population densities sum to the total MGE with a radial fraction h(r).
Assumed in Sects. 3.5-3.6; real metallicity distributions need not be Gaussian, and the fraction profile is a smooth monotonic function.
domain assumption A single galactocentric distance of 8.3 kpc is used for all stars to convert proper motions and projected radii.
Adopted in Sect. 2; a different distance would rescale all physical radii, velocities and masses.
domain assumption For the two-population runs, the SMBH mass is fixed to the literature value and dark matter is neglected.
Justified by the one-population tests in Sect. 4; the two-population results inherit any systematic error in that justification.
standard math The Jeans equations and MGE formalism of Cappellari (2008) and Watkins et al. (2013) are used without re-derivation.
The paper relies on the published CJAM code and equations; no formal verification is provided.

invented entities (2)

High-[M/H] stellar population
purpose: Dominant chemo-dynamical component used to separate the in-situ nuclear population from the metal-poor population and the bar background.
The population is defined by a fitted Gaussian in [M/H] and by membership probabilities from the same likelihood that estimates its kinematics; independent evidence would be chemical abundances or kinematics measured without assignment, which the paper does not provide.
Low-[M/H] stellar population
purpose: Sub-dominant component proposed to have a different, possibly external origin based on its radial anisotropy and weaker rotation.
Same co-estimation issue; the paper uses Anderson-Darling tests to show the selected low-[M/H] stars differ from the background, but the population itself is not independently observable outside the model.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Dynamical mass distribution and velocity structure of the Galactic centre." pith.science (2026). https://pith.science/paper/ZW3AUW6Y

@misc{pith2026250606014,
  author       = {Pith},
  title        = {Pith review of: Dynamical mass distribution and velocity structure of the Galactic centre},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZW3AUW6Y}},
  note         = {Machine review of arXiv:2506.06014}
}

abstract

The inner ~200 pc region of the Milky Way contains a nuclear stellar disc and a nuclear star cluster that are embedded in the larger Galactic bar. These stellar systems overlap spatially, which makes it challenging to separate stars that belong to the nuclear stellar systems, to deduce their internal dynamics, and to derive the central Galactic potential. Discrete stellar kinematics probe the mass distribution of a stellar system, and chemical tracers such as stellar metallicity can further separate multiple stellar populations that can have distinct kinematic properties. We took advantage of the information provided by discrete stellar kinematics and the metallicity in the Galactic centre using discrete chemo-dynamical modelling. We fitted axisymmetric Jeans models to discrete data of 4,600 stars. We fitted the stars as either one population plus a background component or as two populations plus a background that represents the bar. We tested the robustness of the inferred gravitational potential against a varying mass of the supermassive black hole, including dark matter, or a radially varying mass-to-light ratio. We obtained robust results on the fit with a single population and a background component. We obtained a supermassive black hole mass of (4.35$\pm 0.24) \times 10^6$ M$_\odot$, and we find that a dark matter component and radial variation in the mass-to-light ratio are negligible. We derived the enclosed mass profile of the inner ~60 pc and found a lower mass than reported in the literature in the region of ~5-30 pc. In our two-population fit, we found a high-[M/H] population that contributes more than 90% to the total stellar density. The properties of the high-[M/H] population are consistent with in situ formation after gas inflow from the Galactic disc via the bar. The distinct kinematic properties of the low-[M/H] population indicate a different origin. [abridged]

Figures

Figures reproduced from arXiv: 2506.06014 by the authors.

**Figure 1.** Spatial distribution of all stars we used for the model. The black crosses denote stars with VLOS and [M/H], red diamonds show stars with additional proper motions from Fritz et al. (2016), blue circles show stars from Libralato et al. (2020), and orange crosses show stars from Smith et al. (2025). 0 1 2 3 4 5 H-KS 16 14 12 10 8 K S 0 5 10 15 20 25 Density of stars with Vlos [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

**Figure 2.** Colour-magnitude diagram H − KS vs. KS of the stars with VLOS and [M/H]. The vertical green lines denote the colour cuts we used to remove likely foreground (H − KS ≤ 1.3 mag) and background stars (H − KS ≥3.5 mag). The colour represents the density of stars. ments of VLOS and [M/H] with a simple mean. As uncertainties, we used the propagated error of the mean (i.e. 0.5· √ (σ 2 1 + σ 2 2 )) or the standard deviation… view at source ↗

**Figure 3.** Upper panel: Surface density profile derived from the stellar density maps of Gallego-Cano et al. (2020, blue diamonds). The uncertainties are measured using the corresponding uncertainty map and are only shown for a subset of data points to improve visibility. We matched the profile to the centre value of Feldmeier-Krause et al. (2017b) to convert it into units of L⊙ pc−2 . The solid black line denotes the profil… view at source ↗

Figures from the paper (7 more)

**Figure 4.** Figure 4: Radial profile of the velocity anisotropy βz (top), the rotation κ (middle), and the mass-to-number density conversion Υ (bottom) as derived in Sect. 4.2 using a three-parameter function. The red lines denote the median of the posterior distribution, and the grey line…

**Figure 5.** Figure 5: Map of the stellar positions (top) and position-velocity plots along the Galactic longitude l for the PM along l (second panel), along b (third panel), and along the line of sight (fourth panel). The colourcoding is from the median realisation of the one-population mo…

**Figure 6.** Figure 6: Map of the velocity dispersion in three directions (σl , σb, and σLOS), VLOS, and VLOS/σLOS for the median model with a radially varying Υ [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]

**Figure 7.** Figure 7: Total enclosed mass as a function of spherical deprojected radius for the one-population fits with free M• (top left), radially varying Υ (top right), a DM cusp (bottom left), and for the two-population fits (bottom right). The vertical solid lines denote 1 Re of the N…

**Figure 8.** Figure 8: [M/H] histogram. The grey histogram denotes all stars, the vertical dotted lines show the 50th percentile of Z k , and the dashed lines represent the resulting distributions of Z k and σ k Z . Pink denotes the high- [M/H] population, cyan shows the low-[M/H] populati…

**Figure 9.** Figure 9: Map of the stellar positions (top), position-velocity plots along the Galactic longitude l for the PM along l (second panel), along b (third panel), along the line of sight (fourth panel), and along the stellar metallicity [M/H]. The colour-coding is from the 50th per…

**Figure 10.** Figure 10: Spatial distribution of stars, colour-coded from top to bottom by the velocity dispersion along l, along b, along the line of sight, VLOS, VLOS/σLOS, for the 50th percentile two-population model, with high-[M/H] stars (left), and with low-[M/H] stars (right). [M/H]. W…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

80 extracted references · 36 canonical work pages

[1]

2012, ApJ, 750, 111 Arca Sedda, M., Gualandris, A., Do, T., et al

Antonini, F., Capuzzo-Dolcetta, R., Mastrobuono-Battisti, A., & Merritt, D. 2012, ApJ, 750, 111 Arca Sedda, M., Gualandris, A., Do, T., et al. 2020, ApJ, 901, L29 Astropy Collaboration, Price-Whelan, A. M., Lim, P. L., et al. 2022, ApJ, 935, 167 Astropy Collaboration, Price-Whelan, A. M., Sip˝ocz, B. M., et al. 2018, AJ, 156, 123 Astropy Collaboration, Ro...

work page 2012
[2]

M., Schödel, R., et al

Boehle, A., Ghez, A. M., Schödel, R., et al. 2016, ApJ, 830, 17

2016
[3]

2002, MNRAS, 333, 400

Cappellari, M. 2002, MNRAS, 333, 400

2002
[4]

2008, MNRAS, 390, 71

Cappellari, M. 2008, MNRAS, 390, 71

2008
[5]

1993, ApJ, 415, 616

Capuzzo-Dolcetta, R. 1993, ApJ, 415, 616

work page 1993
[6]

2009, A&A, 508, 695

Carretta, E., Bragaglia, A., Gratton, R., D’Orazi, V ., & Lucatello, S. 2009, A&A, 508, 695

2009
[7]

& Saha, P

Chakrabarty, D. & Saha, P. 2001, AJ, 122, 232

work page 2001
[8]

K., Gerhard, O., et al

Chatzopoulos, S., Fritz, T. K., Gerhard, O., et al. 2015, MNRAS, 447, 948

2015

Show all 80 references

[9]

S., Patrick, L

Clark, J. S., Patrick, L. R., Najarro, F., Evans, C. J., & Lohr, M. 2021, A&A, 649, A43

2021
[10]

R., Debattista, V

Cole, D. R., Debattista, V . P., Erwin, P., Earp, S. W. F., & Roškar, R. 2014, MNRAS, 445, 3352

2014
[11]

S., Simpson, J

Cotera, A. S., Simpson, J. P., Erickson, E. F., et al. 1999, ApJ, 510, 747

1999
[12]

2004, PASJ, 56, 261

Deguchi, S., Imai, H., Fujii, T., et al. 2004, PASJ, 56, 261

2004
[13]

2020, ApJ, 901, L28

Do, T., David Martinez, G., Kerzendorf, W., et al. 2020, ApJ, 901, L28

2020
[14]

2015, ApJ, 809, 143

Do, T., Kerzendorf, W., Winsor, N., et al. 2015, ApJ, 809, 143

2015
[15]

D., Cotera, A., et al

Dong, H., Wang, Q. D., Cotera, A., et al. 2011, MNRAS, 417, 114

2011
[16]

S., Elston, R., Raines, S

Eikenberry, S. S., Elston, R., Raines, S. N., et al. 2004, in Society of Photo- Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 5492, Ground-based Instrumentation for Astronomy, ed. A. F. M. Moorwood & M. Iye, 1196–1207

2004
[17]

1994, A&A, 285, 723

Emsellem, E., Monnet, G., & Bacon, R. 1994, A&A, 285, 723

1994
[18]

2024, MNRAS, 528, 3613

Erwin, P. 2024, MNRAS, 528, 3613

2024
[19]

2014, A&A, 570, A2

Feldmeier, A., Neumayer, N., Seth, A., et al. 2014, A&A, 570, A2

2014
[20]

2022, MNRAS, 513, 5920

Feldmeier-Krause, A. 2022, MNRAS, 513, 5920

2022
[21]

2020, MNRAS, 494, 396

Feldmeier-Krause, A., Kerzendorf, W., Do, T., et al. 2020, MNRAS, 494, 396

2020
[22]

2015, A&A, 584, A2

Feldmeier-Krause, A., Neumayer, N., Schödel, R., et al. 2015, A&A, 584, A2

2015
[23]

2025, A&A, 696, A213

Feldmeier-Krause, A., Neumayer, N., Seth, A., et al. 2025, A&A, 696, A213

2025
[24]

W., Lang, D., & Goodman, J

Foreman-Mackey, D., Hogg, D. W., Lang, D., & Goodman, J. 2013, PASP, 125, 306

2013
[25]

K., Chatzopoulos, S., Gerhard, O., et al

Fritz, T. K., Chatzopoulos, S., Gerhard, O., et al. 2016, ApJ, 821, 44

2016
[26]

K., Patrick, L

Fritz, T. K., Patrick, L. R., Feldmeier-Krause, A., et al. 2021, A&A, 649, A83

2021
[27]

A., Bittner, A., Falcón-Barroso, J., et al

Gadotti, D. A., Bittner, A., Falcón-Barroso, J., et al. 2020, A&A, 643, A14

2020
[28]

A., Sánchez-Blázquez, P., Falcón-Barroso, J., et al

Gadotti, D. A., Sánchez-Blázquez, P., Falcón-Barroso, J., et al. 2019, MNRAS, 482, 506

2019
[29]

2018, A&A, 609, A26

Gallego-Cano, E., Schödel, R., Dong, H., et al. 2018, A&A, 609, A26

2018
[30]

2020, A&A, 634, A71

Gallego-Cano, E., Schödel, R., Nogueras-Lara, F., et al. 2020, A&A, 634, A71

2020
[31]

E., & Ott, T

Genzel, R., Pichon, C., Eckart, A., Gerhard, O. E., & Ott, T. 2000, MNRAS, 317, 348

2000
[32]

Genzel, R., Thatte, N., Krabbe, A., Kroker, H., & Tacconi-Garman, L. E. 1996, ApJ, 472, 153

1996
[33]

A., Rejkuba, M., Minniti, D., et al

Gonzalez, O. A., Rejkuba, M., Minniti, D., et al. 2011, A&A, 534, L14 GRA VITY Collaboration, Abuter, R., Aimar, N., et al. 2022, A&A, 657, L12

2011
[34]

W., Rieke, M

Haller, J. W., Rieke, M. J., Rieke, G. H., et al. 1996, ApJ, 456, 194

1996
[35]

R., Millman, K

Harris, C. R., Millman, K. J., van der Walt, S. J., et al. 2020, Nature, 585, 357

2020
[36]

W., Do, T., Lu, J

Hosek, M. W., Do, T., Lu, J. R., et al. 2022, ApJ, 939, 68

2022
[37]

Hunter, J. D. 2007, Computing In Science & Engineering, 9, 90

2007
[38]

2025, arXiv e-prints, arXiv:2501.14868

Hussein, A., Necib, L., Kaplinghat, M., et al. 2025, arXiv e-prints, arXiv:2501.14868

2025 arXiv
[39]

O., Wende-von Berg, S., Dreizler, S., et al

Husser, T. O., Wende-von Berg, S., Dreizler, S., et al. 2013, A&A, 553, A6

2013
[40]

2023, A&A, 674, A70

Ishchenko, M., Sobolenko, M., Kuvatova, D., Panamarev, T., & Berczik, P. 2023, A&A, 674, A70

2023
[41]

2022, ApJ, 939, 118

Kacharov, N., Alfaro-Cuello, M., Neumayer, N., et al. 2022, ApJ, 939, 118

2022
[42]

2020, MNRAS, 492, 966

Kamann, S., Dalessandro, E., Bastian, N., et al. 2020, MNRAS, 492, 966

2020
[43]

Kerzendorf, W. & Do, T. 2015, starkit: First real release, Zenodo Source Code Library, 2015

2015
[44]

2018, MNRAS, 473, 776

Kettlety, T., Hesling, J., Phillipps, S., et al. 2018, MNRAS, 473, 776

2018
[45]

F., Casali, M., et al

Kissler-Patig, M., Pirard, J. F., Casali, M., et al. 2008, A&A, 491, 941 Article number, page 14 of 20 A. Feldmeier-Krause et al.: Dynamical mass distribution and velocity structure of the Galactic centre

2008
[46]

Launhardt, R., Zylka, R., & Mezger, P. G. 2002, A&A, 384, 112

2002
[47]

P., & Bellini, A

Libralato, M., Fardal, M., Lennon, D., van der Marel, R. P., & Bellini, A. 2020, MNRAS, 497, 4733

2020
[48]

J., Bellini, A., et al

Libralato, M., Lennon, D. J., Bellini, A., et al. 2021, MNRAS, 500, 3213

2021
[49]

J., & Winnberg, A

Lindqvist, M., Habing, H. J., & Winnberg, A. 1992, A&A, 259, 118

1992
[50]

2019, MNRAS, 484, 1166

Magorrian, J. 2019, MNRAS, 484, 1166

2019
[51]

T., Sellgren, K., Becklin, E

McGinn, M. T., Sellgren, K., Becklin, E. E., & Hall, D. N. B. 1989, ApJ, 338, 824

1989
[52]

E., Schinnerer, E., van de Ven, G., et al

Meidt, S. E., Schinnerer, E., van de Ven, G., et al. 2014, ApJ, 788, 144

2014
[53]

1985, AJ, 90, 1027

Merritt, D. 1985, AJ, 90, 1027

1985
[54]

2010, Dark matter at the centres of galaxies (Cambridge University Press), 83–98

Merritt, D. 2010, Dark matter at the centres of galaxies (Cambridge University Press), 83–98

2010
[55]

F., Frenk, C

Navarro, J. F., Frenk, C. S., & White, S. D. M. 1996, ApJ, 462, 563

1996
[56]

2013, ApJ, 769, L28

Nishiyama, S., Yasui, K., Nagata, T., et al. 2013, ApJ, 769, L28

2013
[57]

T., Dong, H., et al

Nogueras-Lara, F., Gallego-Calvente, A. T., Dong, H., et al. 2018, A&A, 610, A83

2018
[58]

T., et al

Nogueras-Lara, F., Schödel, R., Gallego-Calvente, A. T., et al. 2019, A&A, 631, A20

2019
[59]

A., Meidt, S., Van de Ven, G., et al

Norris, M. A., Meidt, S., Van de Ven, G., et al. 2014, ApJ, 797, 55

2014
[60]

S., & Figer, D

Oh, S.-K., Kim, S.-S. S., & Figer, D. F. 2009, Journal of Korean Astronomical Society, 42, 17

2009
[61]

Osipkov, L. P. 1979, Pisma v Astronomicheskii Zhurnal, 5, 77

1979
[62]

2006, ApJ, 643, 1011

Paumard, T., Genzel, R., Martins, F., et al. 2006, ApJ, 643, 1011

2006
[63]

2021, ApJ, 921, 8

Pinna, F., Neumayer, N., Seth, A., et al. 2021, ApJ, 921, 8

2021
[64]

2013, arXiv e-prints, arXiv:1306.2945

Poleski, R. 2013, arXiv e-prints, arXiv:1306.2945

2013 arXiv
[65]

2017, MNRAS, 465, 1621

Portail, M., Gerhard, O., Wegg, C., & Ness, M. 2017, MNRAS, 465, 1621

2017
[66]

Reid, M. J. & Brunthaler, A. 2020, ApJ, 892, 39

2020
[67]

2003, in Society of Photo-Optical In- strumentation Engineers (SPIE) Conference Series, V ol

Rousset, G., Lacombe, F., Puget, P., et al. 2003, in Society of Photo-Optical In- strumentation Engineers (SPIE) Conference Series, V ol. 4839, Adaptive Op- tical System Technologies II, ed. P. L. Wizinowich & D. Bonaccini, 140–149

2003
[68]

L., Kawata, D., Matsunaga, N., et al

Sanders, J. L., Kawata, D., Matsunaga, N., et al. 2024, MNRAS, 530, 2972 Schödel, R., Feldmeier, A., Kunneriath, D., et al. 2014, A&A, 566, A47 Schödel, R., Merritt, D., & Eckart, A. 2009, A&A, 502, 91

2024
[69]

Scholz, F. W. & Stephens, M. A. 1987, Journal of the American Statistical As- sociation, 82, 918

1987
[70]

K., Nandakumar, G., et al

Schultheis, M., Fritz, T. K., Nandakumar, G., et al. 2021, A&A, 650, A191

2021
[71]

2013, The Messenger, 151, 21

Sharples, R., Bender, R., Agudo Berbel, A., et al. 2013, The Messenger, 151, 21

2013
[72]

C., Lucas, P

Smith, L. C., Lucas, P. W., Koposov, S. E., et al. 2025, MNRAS, 536, 3707

2025
[73]

C., Magorrian, J., Nogueras-Lara, F., et al

Sormani, M. C., Magorrian, J., Nogueras-Lara, F., et al. 2020, MNRAS, 499, 7

2020
[74]

C., Sanders, J

Sormani, M. C., Sanders, J. L., Fritz, T. K., et al. 2022, MNRAS, 512, 1857 Støstad, M., Do, T., Murray, N., et al. 2015, ApJ, 808, 106

2022
[75]

D., Ostriker, J

Tremaine, S. D., Ostriker, J. P., & Spitzer, L., J. 1975, ApJ, 196, 407

1975
[76]

E., et al

Trippe, S., Gillessen, S., Gerhard, O. E., et al. 2008, A&A, 492, 419

2008
[77]

2017, MNRAS, 464, 3720 van de Ven, G., van den Bosch, R

Tsatsi, A., Mastrobuono-Battisti, A., van de Ven, G., et al. 2017, MNRAS, 464, 3720 van de Ven, G., van den Bosch, R. C. E., Verolme, E. K., & de Zeeuw, P. T. 2006, A&A, 445, 513 van Donkelaar, F., Mayer, L., Capelo, P. R., et al. 2024, MNRAS, 529, 4104

2017
[78]

E., et al

Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020, Nature Methods, 17, 261 von Fellenberg, S. D., Gillessen, S., Stadler, J., et al. 2022, ApJ, 932, L6

2020
[79]

L., van de Ven, G., den Brok, M., & van den Bosch, R

Watkins, L. L., van de Ven, G., den Brok, M., & van den Bosch, R. C. E. 2013, MNRAS, 436, 2598

2013
[80]

2020, MNRAS, 496, 1579

Zhu, L., van de Ven, G., Leaman, R., et al. 2020, MNRAS, 496, 1579

2020

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

2012, ApJ, 750, 111 Arca Sedda, M., Gualandris, A., Do, T., et al

Antonini, F., Capuzzo-Dolcetta, R., Mastrobuono-Battisti, A., & Merritt, D. 2012, ApJ, 750, 111 Arca Sedda, M., Gualandris, A., Do, T., et al. 2020, ApJ, 901, L29 Astropy Collaboration, Price-Whelan, A. M., Lim, P. L., et al. 2022, ApJ, 935, 167 Astropy Collaboration, Price-Whelan, A. M., Sip˝ocz, B. M., et al. 2018, AJ, 156, 123 Astropy Collaboration, Ro...

work page 2012

[2] [2]

M., Schödel, R., et al

Boehle, A., Ghez, A. M., Schödel, R., et al. 2016, ApJ, 830, 17

2016

[3] [3]

2002, MNRAS, 333, 400

Cappellari, M. 2002, MNRAS, 333, 400

2002

[4] [4]

2008, MNRAS, 390, 71

Cappellari, M. 2008, MNRAS, 390, 71

2008

[5] [5]

1993, ApJ, 415, 616

Capuzzo-Dolcetta, R. 1993, ApJ, 415, 616

work page 1993

[6] [6]

2009, A&A, 508, 695

Carretta, E., Bragaglia, A., Gratton, R., D’Orazi, V ., & Lucatello, S. 2009, A&A, 508, 695

2009

[7] [7]

& Saha, P

Chakrabarty, D. & Saha, P. 2001, AJ, 122, 232

work page 2001

[8] [8]

K., Gerhard, O., et al

Chatzopoulos, S., Fritz, T. K., Gerhard, O., et al. 2015, MNRAS, 447, 948

2015