On the long-term evolution of razor-thin galactic discs: Balescu-Lenard prediction and perspectives

Christophe Pichon; Jean-Baptiste Fouvry; Mathieu Roule; Pierre-Henri Chavanis

REVIEW 3 major objections 5 minor 1 cited by

On the long-term evolution of razor-thin galactic discs: Balescu-Lenard prediction and perspectives

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

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

Pith's one-line read The Balescu–Lenard kinetic equation, applied to a cold razor-thin Mestel disc, predicts the measured ensemble-averaged relaxation rate in action space within 10 percent.

desk verdict First quantitative BL validation for cold discs—convincing but needs a convergence check on the half-mass disc. read the letter →

arxiv 2502.07342 v1 pith:D4P3T4UR submitted 2025-02-11 astro-ph.GA

Mathieu Roule , Jean-Baptiste Fouvry , Christophe Pichon , Pierre-Henri Chavanis This is my paper · ORCID

classification astro-ph.GA

keywords Balescu-LenardequationgalacticdiscskinetictheorysecularevolutionMesteldisccollectiveeffectsN-bodysimulationsresonantrelaxation

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

This paper sets out to show that the inhomogeneous Balescu–Lenard equation, the most complete kinetic theory available for isolated stellar systems, quantitatively predicts the long-term relaxation of a cold razor-thin galactic disc rather than merely its qualitative shape. The test is a tapered Mestel disc — a razor-thin model with a flat rotation curve, chosen to mimic disc galaxies like the Milky Way — and the comparison is between the theoretical rate of change of the distribution function in action space (the space of orbital labels, radial action and angular momentum) and the rate measured in 1,000 N-body simulations with $25\times10^6$ particles each. The integrated absolute relaxation rates agree within 10 percent, and the theoretical curves fall inside the scatter of the simulated realizations. If this stands, the slow collisional drift that eventually drives a stable disc to instability — the secular evolution that shapes galactic discs over cosmic time — becomes a quantity calculable from first principles, and the choice of softening kernel in simulations becomes a controlled element of the comparison rather than an incidental one.

What carries the argument

The load-bearing object is the inhomogeneous Balescu–Lenard equation, a diffusion equation in action space governing the mean distribution function $F(\mathbf{J}, t)$; its flux sums over resonant pairs of orbits $(\mathbf{k}, \mathbf{k}')$ selected by the resonance condition $\mathbf{k}\cdot\boldsymbol{\Omega}(\mathbf{J}) - \mathbf{k}'\cdot\boldsymbol{\Omega}(\mathbf{J}') = 0$. The efficiency of each resonance is set by the dressed coupling coefficients $U^d_{\mathbf{k}\mathbf{k}'}(\mathbf{J}, \mathbf{J}', \mathbf{k}\cdot\boldsymbol{\Omega})$, which are built from the susceptibility matrix $\mathbf{N}(\omega) = [\mathbf{I} - \mathbf{M}(\omega)]^{-1}$ of linear response theory and carry the disc's collective, swing-amplified response. The quantitative success reported here rests on computing these coefficients with 100 Clutton–Brock basis elements, 21 resonances, and the analytic continuation of the response matrix to real frequencies, and on evaluating the resonant integrals in adapted coordinates.

What would settle it

Recompute the figure-3 prediction with the convergence parameters deliberately varied, say 200 Clutton–Brock basis elements or more than 21 resonances: if the integrated absolute relaxation rate moves by more than about 10 percent, the reported agreement is a truncation artifact rather than a property of the kinetic equation. A second check is temporal: repeat the 1,000-realization measurement at a later time such as $t = 200\,t_{\rm dyn}$, closer to the instability onset, and see whether the Balescu–Lenard prediction still falls inside the realization scatter there. A third probe is modal: measure the power spectrum of $\ell = 2$ fluctuations in a long quiet-start run before the instability and look for damped-mode peaks at their inner Lindblad resonance radii, as the converged susceptibility predicts.

Watch

Extended reading notes

Core claim

The central claim is that the Balescu–Lenard equation (equation 3), evaluated with care, returns the ensemble-averaged relaxation rate $\partial F/\partial t$ of a cold razor-thin Mestel disc that matches 1,000 N-body realizations in both shape and amplitude: the integrated absolute rate agrees within 10 percent, and one-dimensional slices through action space stay inside the one-$\sigma$ scatter of the simulations (figures 3 and 4, at $t = 150\,t_{\rm dyn}$, before the disc becomes Vlasov-unstable). The agreement is presented as nontrivial on two counts: collective amplification speeds the relaxation by about three orders of magnitude relative to the undressed Landau equation, so the match tests the dressed coupling, not a trivial baseline; and the earlier Balescu–Lenard computation for this same disc (F+15) produced a sharp ridge that the present, converged calculation identifies as an artifact of an insufficiently converged linear susceptibility. The same machinery ties the long-term heating to the disc's weakly damped modes and swing amplification, and explains the groove, the resonant ridges, and the eventual dynamical phase transition toward instability seen in earlier N-body work, with the caveat that both the theory and the simulations here are restricted to the $\ell = 2$ bisymmetric harmonic.

Load-bearing premise

The paper's central claim rests on the numerical convergence of the dressed coupling coefficients $U^d$, computed here with 100 basis elements, 21 resonances, and an analytic continuation; the paper itself flags this premise as delicate — its figure 6 shows that an insufficiently converged susceptibility changes the predicted relaxation pattern qualitatively, and Section 4.2 concedes that improving the convergence of the linear predictions is necessary for more quantitative comparisons.

Editorial extensions

If this is right

The slow pre-instability phase of a cold disc — the secular drift that carves the groove and eventually destabilizes the disc — becomes predictable in advance by kinetic theory rather than diagnosable only after the fact in simulations.
Collective effects are not a small correction in this regime: the dressed Balescu–Lenard relaxation rate exceeds the undressed Landau rate by about three orders of magnitude, so quantitative theory for cold discs must include the dressing.
The sharp resonant ridge in the earlier Balescu–Lenard application to this disc (F+15) is identified as an artifact of a poorly converged susceptibility; converged coefficients give broad heating in action space, matching the new measurements.
Gravitational softening is a long-term bias, not just a short-range regularization: Plummer softening shifts the disc's modes and delays the relaxation and the phase transition, while Kuzmin softening leaves them nearly unchanged.
Close to marginal stability, the averaged evolution is not representative of individual realizations, whose localized action-space ridges vary in number, location, and strength from run to run.

Reading between the lines

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

A direct extension the paper leaves implicit: run the same 1,000-realization comparison at several later times ($t \gtrsim 200\,t_{\rm dyn}$) to map how the 10-percent agreement degrades as the disc approaches marginal stability — a regime where the authors themselves note the Balescu–Lenard equation diverges.
The softening-kernel result plausibly carries over to cosmological simulations of thin discs: quoted bar-formation or secular-relaxation times may inherit a kernel-dependent bias of order $\epsilon$ for Plummer-style softening, and this bias persists even as particle number grows.
The validation covers one disc family at one temperature ($Q = 1.5$), so whether the 10-percent accuracy is universal is open; repeating the comparison on exponential discs or at other Toomre $Q$ values would settle it.

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.

Desk Editor's Note

read the letter

Quick take: this is the first quantitative check of the inhomogeneous Balescu–Lenard equation for a cold razor-thin disc, and it holds up. The integrated relaxation rate agrees with 1000 N-body realizations to 10%, and the action-space slices sit inside the realization scatter. That's a real result, not a rounding error. The paper also ships its code (SecularResponse.jl and the N-body code mestel2d), which is more than most.

The main novelty is the convergence of the linear response. F+15 used 9 basis functions and an un-converged susceptibility; the appendix here shows that the same setup would wrongly predict instability. The improved treatment—Clutton-Brock basis, analytic continuation, resonant coordinates—makes the dressed couplings reliable enough to match N-body. The comparison is honest: the BL rate is computed from the initial DF, with no fitted parameters, so there's no circularity. The discussion of softening kernels (Plummer vs Kuzmin) and the demonstration that Plummer softening biases long-term evolution are useful and will save people from bad choices in future simulations.

Where I'd push back: the stress-test concern about convergence is real, though not fatal. The linear response parameters are inherited from PR+24, where convergence was checked for the Zang nu=4 disc, not for the half-mass Mestel disc studied here. That matters because the half-mass disc is close to marginal stability, and the response matrix is sensitive to truncation there. The paper itself shows how much the BL pattern changes when the basis is degraded (figure 6, top vs bottom), and Section 4.2 says 'improving the convergence of the linear predictions is necessary for more quantitative comparisons.' No convergence series in basis size or resonance count is shown for this disc. The saving grace is that the converged calculation matches N-body while the degraded one doesn't—that's indirect but decent evidence—but a referee should ask for a dedicated convergence test before this is fully settled.

Other soft spots are minor: the data are 'available on reasonable request,' no commit hash for the code, the quasi-static comparison (BL at initial DF versus measured 0–150 tdyn) is not spelled out explicitly, and the mode-fit errors in Appendix G are not quantified. None of these affect the central conclusion.

Bottom line: this paper deserves a serious referee and, with a convergence test added, would be solid. I'd bring it to reading group. I'd cite it if I were working in disc kinetics or N-body relaxation.

Referee Report

3 major / 5 minor

Summary. The paper tests the inhomogeneous Balescu–Lenard (BL) equation against N-body simulations for a half-mass, razor-thin Mestel disc. The central result is Fig. 3: the BL prediction of the initial relaxation rate in action space, computed with the linear-response code LinearResponse.jl, agrees with the ensemble-averaged rate from 1000 N-body realisations at the 10% level in an integrated sense, with action-space slices within the realisation scatter (Fig. 4). The remaining sections use the same tools to argue that collective effects amplify relaxation by roughly 10^3 relative to Landau theory, that weakly damped modes leave imprints in the susceptibility and shape the heating, that Plummer softening biases long-term evolution whereas Kuzmin softening does not, and that individual realisations show large stochasticity near marginal stability.

Significance. If the central comparison is accepted, this is the first quantitative validation of the inhomogeneous BL equation for a cold, self-gravitating disc, and a strong demonstration that a parameter-free kinetic prediction can capture an ensemble-averaged N-body measurement. The paper has real strengths: 1000 realisations, publicly available code, and an honest attempt to separate the ensemble average from single-realisation behaviour. However, the validation currently rests on two unproven technical assumptions: that the initial-DF BL rate can be compared with a 150 tdyn finite difference, and that the dressed couplings are converged for this specific near-marginal disc. Because the paper's own Fig. 6 shows strong sensitivity to linear-response truncation and Section 4.2 concedes convergence work remains, the 10% claim is not yet fully supported.

major comments (3)

[Section 3.3, Fig. 3, Appendix D] The BL rate in Fig. 3 is evaluated from the initial DF at t=0 (Appendix D), while the N-body rate is the finite difference [F(150)-F(0)]/150. The manuscript does not demonstrate that ∂F/∂t is approximately constant over this window; if the DF evolves appreciably, the comparison is not between the same quantity. Please quantify the predicted evolution of F over 0-150 tdyn (for example, by comparing the BL rate at t=0 and at a later time, or by showing that ∫dt ∂F/∂t is within the stated tolerance of [F(150)-F(0)]/150), or explicitly justify the quasi-static approximation.
[Appendix C.3, Fig. 6, Section 4.2] The central 10% agreement rests on the dressed couplings U^d of Eq. (C.5) being converged at 100 Clutton-Brock basis elements and 21 resonances. These parameters were validated in PR+24 for the unstable Zang nu=4 disc, not for the half-mass Mestel disc at Q=1.5 studied here; near marginal stability the response matrix is sensitive to truncation. The paper's own Fig. 6 shows that a coarser truncation yields a qualitatively different BL pattern, and Section 4.2 states that improving linear-response convergence is needed for 'more quantitative comparisons'. Please add a convergence study for this disc (for example, varying n_basis and resonance count and showing that the integrated rate and the Fig. 4 slices are stable), or qualify the 10% claim accordingly.
[Section 4, first paragraph] The text states that 'the overall amplitude of the flux is yet to be convincingly explained' immediately after Section 3.3 claims a 10% amplitude agreement. This apparent contradiction should be resolved: either the amplitude agreement is a quantitative validation, in which case the 'yet to be explained' statement should be removed or reformulated, or the agreement is not considered explanatory, in which case the central claim should be qualified.

minor comments (5)

[Section 3.3] The 10% criterion is attributed to Eq. (12) of Tep et al. (2022) but is not stated in the text; please define the integrated quantity and its normalisation so that the reader can reproduce the number.
[Appendix C.3] The sentence 'we used the same parameters as PR+24 (see tables F1 therein)' does not reproduce those tables; list the basis number, resonance range, and any other numerical parameters explicitly in this paper or in an appendix.
[Figure 6] The bottom panel is captioned 'numerically converged linear susceptibility', but no convergence test is shown for this disc; rephrase to 'computed with the standard truncation' or provide the convergence evidence in the text.
[Section 4.1] The estimate |∂tF^BL|/|∂tF^Landau| ≈ 10^3 is stated as following from |U^d/U|^2 ≈ 30^2; please specify whether this is evaluated at a representative resonance or from the integrated fluxes of Figs. 3 and 5, and add the numerical ratio.
[Data distribution] The averaged N-body data underlying Fig. 3 are only available 'through reasonable request'; consider publishing the binned rates as supplementary material so that the 10% criterion can be checked by readers.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the Balescu-Lenard prediction contains no parameters fitted to the N-body data and is benchmarked against independent linear-response tests.

full rationale

The central claim, that the inhomogeneous Balescu-Lenard equation reproduces the ensemble-averaged relaxation rate of the Mestel disc, is not circular. The BL rate (Eq. 3) is computed from the initial DF of Eq. (A.2) via the dressed coupling coefficients of Eq. (C.5), with no free parameter adjusted to the N-body measurements. The comparison in Fig. 3 uses the same initial DF as the simulations, which is the appropriate predictive test rather than an input-output inversion. The numerical machinery is supported independently: LinearResponse.jl is validated against the unstable Zang ν=4 disc and against the mode frequencies of Sellwood & Evans (2001) and De Rijcke et al. (2019a) in Fig. G.1, and Appendix D reports an internal resonance-convergence check. The paper's own admission in Sec. 4.2 that 'improving the convergence of the linear predictions is necessary for more quantitative comparisons' and the sensitivity to truncation shown in Fig. 6 are legitimate numerical-convergence caveats; they do not amount to a constructional equivalence between the predicted and measured quantities. The self-citations to PR+24 and Fouvry & Prunet (2022) are citations to shared numerical tools, not to a uniqueness theorem or an unverified ansatz, and they do not make the BL prediction logically dependent on the N-body result being predicted.

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

No parameters were fitted to the N-body data. Disc parameters (xi=0.5, q=11.4, tapers) are inherited from S12 and F+15; numerical settings (100 basis elements, 21 resonances, Kuzmin epsilon=0.16, timestep) are convergence or numerical choices, not free parameters tuned to make the BL prediction match. No new particles, forces, or conserved quantities are introduced. The weakly damped modes discussed in Section 4.2 are standard linear-theory objects, not new entities. The axioms listed are the background assumptions that the prediction and comparison rely on, the most fragile being the convergence of the linear response and the quasi-static comparison.

assumptions (6)

domain assumption The inhomogeneous Balescu-Lenard equation (Eq. 3) describes the ensemble-averaged long-term evolution of isolated self-gravitating stellar systems.
Treated as the master equation from Heyvaerts (2010) and Chavanis (2012); the paper validates, rather than derives, it for cold discs.
domain assumption The half-mass Mestel disc with parameters xi=0.5, q=11.4, tapers Rin=1, Rout=11.5, nu=4, mu=5 approximates a galactic disc, and its angle-action coordinates are integrable.
Model taken from S12 and F+15; the integrability of axisymmetric razor-thin discs is standard (Lynden-Bell & Kalnajs 1972).
domain assumption Linear response theory, computed with 100 Clutton-Brock basis elements, 21 resonances, and the analytic continuation of Fouvry & Prunet (2022), yields converged dressed coupling coefficients U^d for the BL prediction.
Section 4.2 and Appendix E state that F+15's coefficients were not converged and that the present calculation is tied to the accuracy of the linear susceptibility; convergence is checked internally, not against an external benchmark for this disc.
domain assumption The restriction to ell=2 harmonics and to the ILR, OLR, and corotation resonances with |kr|<=1 captures the dominant relaxation.
Appendix D states that increasing the number of radial resonances had negligible impact; the ell=2 restriction follows the setup of S12 and F+15.
domain assumption The Kuzmin softened N-body simulations with epsilon=0.16 (and grid/time-step settings) faithfully represent the collisionless dynamics of the unsoftened disc over 150 tdyn.
Appendix B.3 and figure G.1 support the small bias of the Kuzmin kernel, but the equivalence is assumed, not proven.
ad hoc to paper The BL prediction evaluated with the initial DF can be compared directly with the N-body finite-difference rate over 0 to 150 tdyn (quasi-static approximation).
Section 3.3 does not state this explicitly; the disc is slowly evolving until about 200 tdyn, so the approximation is plausible but unquantified.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the long-term evolution of razor-thin galactic discs: Balescu-Lenard prediction and perspectives." pith.science (2026). https://pith.science/paper/D4P3T4UR

@misc{pith2026250207342,
  author       = {Pith},
  title        = {Pith review of: On the long-term evolution of razor-thin galactic discs: Balescu-Lenard prediction and perspectives},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/D4P3T4UR}},
  note         = {Machine review of arXiv:2502.07342}
}

read the original abstract

In the last five decades, numerical simulations have provided invaluable insights into the evolution of galactic discs over cosmic times. As a complementary approach, developments in kinetic theory now also offer a theoretical framework to understand statistically their long-term evolution. The current state-of-the-art kinetic theory of isolated stellar systems is the inhomogeneous Balescu-Lenard equation. It can describe the long-term evolution of a self-gravitating razor-thin disc under the effect of resonant interactions between collectively amplified noise-driven fluctuations. In this work, confronting theoretical predictions to numerical simulations, we quantitatively show that kinetic theory indeed captures the average long-term evolution of cold stellar discs. Leveraging the versatility of kinetic methods, we then offer some new perspectives on this problem, namely (i) the crucial impact of collective effects in accelerating the relaxation; (ii) the role of (weakly) damped modes in shaping the disc's orbital heating; (iii) the bias introduced by gravitational softening on long timescales; (iv) the resurgence of strong stochasticity near marginal stability. These elements call for an appropriate choice of softening kernel when simulating the long-term evolution of razor thin discs and for an extension of kinetic theory beyond the average evolution. Notwithstanding, kinetic theory captures quantitatively the ensemble-averaged long-term response of such discs.

Figures

Figures reproduced from arXiv: 2502.07342 by the authors.

**Figure 1.** Time evolution of the bisymmetric fluctuations (equation B.6) in 12 different N–body realisations of the Mestel disc, with N = 25×106 particles each, using a running average over 30 dynamical times. The disc is initially stable and slowly relaxes towards an unstable state. Once unstable, the evolution is dominated by an exponentially growing mode before it saturates. The disc’s configuration is illustrated in figure… view at source ↗

**Figure 2.** Illustration of the typical evolution of a Mestel disc in one N– body simulation, via the surface density (up to R= 8R0) at two different times. During the collisional phase (left), the disc remains reasonably axisymmetric, displaying recurrent weak transient spirals. After the phase transition (right), the disc develops strong bisymmetric fluctuations, which ultimately saturate (figure 1), up to the late formati… view at source ↗

**Figure 4.** Slices of the local relaxation rate, ∂F/∂t, from figure 3 for fixed Jr (top) or fixed L (bottom), as predicted from BL (dashed) and measured in N–body simulations (full lines). Here, ∂F/∂t in the top (resp. bottom) panel has been averaged over an interval of width δJr =±0.02 (resp. δL= 0.06) and subsequently smoothed with a running average of width δL= 0.03 (resp. δJr = 0.01). For the N–body, we measured ∂F/∂t in e… view at source ↗

Figures from the paper (5 more)

**Figure 3.** Figure 3: Top: Local relaxation rate, ∂F/∂t, in action space as predicted by BL (equation 3) for the Mestel disc (section 2.1), computed in the centre of the action bins. Red regions correspond to an increase in the number of particles while blue contours correspond to a depleti…

**Figure 5.** Figure 5: Relaxation rate, ∂F/∂t, as predicted by the Landau equation (see appendix F), using the same convention as in figure 3. The Landau equation predicts a more isotropic diffusion. In addition, the relaxation time predicted by BL is three orders of magnitude shorter than t…

**Figure 6.** Figure 6: is composed of two panels, each of them with two sub-plots. In both panel, the top plot is the BL relaxation rate in action space, using the same convention as in figure 3. The bottom plot represents the disc’s susceptibility, through the determinant of the susceptibi…

**Figure 8.** Figure 8 [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]

**Figure 9.** Figure 9: Relaxation rate, ∂F/∂t, measured in twelve independent N– body realisations, using the same convention as in figure 3. The number, location, amplitude and birth time of the action ridges vary strongly from one realisation to the other. This is a measure of a long-term…

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

Galactokinetics II: Spiral structure
astro-ph.GA 2025-07 conditional novelty 7.0 of 10

A single linear response kernel for galactic disks reproduces Lindblad-Kalnajs waves, swing amplification, groove instabilities, and Lin-Shu-Kalnajs modes as limiting cases, with smooth connections between them.

Reference graph

Works this paper leans on

75 extracted references · 73 canonical work pages · cited by 1 Pith paper

[1]

Aarseth , S. J. 1963, , 126, 223

work page 1963
[2]

Benetti , F. P. C. & Marcos , B. 2017, , 95, 022111

work page 2017
[3]

Bezanson, J., Edelman, A., Karpinski, S., & Shah, V. B. 2017, SIAM Review, 59, 65

2017
[4]

& Tremaine , S

Binney , J. & Tremaine , S. 2008, Galactic Dynamics: Second Edition (Princeton Univ. Press)

work page 2008
[5]

Bouchet , F. 2020, J. Stat. Phys., 181, 515

work page 2020
[6]

2008, , 78, 040102

Campa , A., Chavanis , P.-H., Giansanti , A., & Morelli , G. 2008, , 78, 040102

work page 2008
[7]

2004, , 352, 427

Chakrabarty, D. 2004, , 352, 427

work page 2004
[8]

1943, ApJ, 97, 255

Chandrasekhar , S. 1943, ApJ, 97, 255

work page 1943

Show all 75 references

[9]

2012, Physica A, 391, 3680

Chavanis , P.-H. 2012, Physica A, 391, 3680

2012
[10]

2013, , 556, A93

Chavanis , P.-H. 2013, , 556, A93

2013
[11]

2024, Eur

Chavanis , P.-H. 2024, Eur. Phys. J. Plus, 139, 51

2024
[12]

1972, , 16, 101

Clutton-Brock , M. 1972, , 16, 101

1972
[13]

1979, ApJ, 234, 1036

Cohn , H. 1979, ApJ, 234, 1036

1979
[14]

2019 a , MNRAS, 485, 150

De Rijcke , S., Fouvry , J.-B., & Dehnen , W. 2019 a , MNRAS, 485, 150

2019
[15]

2019 b , MNRAS, 484, 3198

De Rijcke , S., Fouvry , J.-B., & Pichon , C. 2019 b , MNRAS, 484, 3198

2019
[16]

Debattista , V. P. & Sellwood , J. A. 2000, ApJ, 543, 704

2000
[17]

2001, MNRAS, 324, 273

Dehnen , W. 2001, MNRAS, 324, 273

2001
[18]

Earn , D. J. D. & Sellwood , J. A. 1995, ApJ, 451, 533

1995
[19]

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

Eilers , A.-C., Hogg , D. W., Rix , H.-W., & Ness , M. K. 2019, ApJ, 871, 120

2019
[20]

Evans , N. W. & Read , J. C. A. 1998, MNRAS, 300, 106

1998
[21]

& Bouchet , F

Feliachi , O. & Bouchet , F. 2022, J. Stat. Phys., 186, 22

2022
[22]

& Fouvry , J.-B

Feliachi , O. & Fouvry , J.-B. 2024, , 110, 024108

2024
[23]

& Bar-Or , B

Fouvry , J.-B. & Bar-Or , B. 2018, MNRAS, 481, 4566

2018
[24]

2021, MNRAS, 508, 2210

Fouvry , J.-B., Hamilton , C., Rozier , S., & Pichon , C. 2021, MNRAS, 508, 2210

2021
[25]

2017, MNRAS, 471, 2642

Fouvry , J.-B., Pichon , C., Chavanis , P.-H., & Monk , L. 2017, MNRAS, 471, 2642

2017
[26]

2015, , 584, A129

Fouvry , J.-B., Pichon , C., Magorrian , J., & Chavanis , P.-H. 2015, , 584, A129

2015
[27]

& Prunet , S

Fouvry , J.-B. & Prunet , S. 2022, MNRAS, 509, 2443

2022
[28]

R., Qin , H., & Geyko , V

Fu , Y., Angus , J. R., Qin , H., & Geyko , V. I. 2024, arXiv, 2410.12079

2024 arXiv
[29]

2016, A&A, 595, A2

Gaia Collaboration et al. 2016, A&A, 595, A2

2016
[30]

P., Mather , J

Gardner , J. P., Mather , J. C., et al. 2006, Space Sci. Rev., 123, 485

2006
[31]

1998, MNRAS, 298, 1239

Giersz , M. 1998, MNRAS, 298, 1239

1998
[32]

& Lynden-Bell , D

Goldreich , P. & Lynden-Bell , D. 1965, MNRAS, 130, 125

1965
[33]

V., Poggio , E., et al

Grion Filho , D., Johnston , K. V., Poggio , E., et al. 2021, MNRAS, 507, 2825

2021
[34]

2024, MNRAS, 528, 5286

Hamilton , C. 2024, MNRAS, 528, 5286

2024
[35]

& Heinemann , T

Hamilton , C. & Heinemann , T. 2020, arXiv, 2011.14812

2020 arXiv
[36]

& Heinemann , T

Hamilton , C. & Heinemann , T. 2023, MNRAS, 525, 4161

2023
[37]

C., Breen , P

Heggie , D. C., Breen , P. G., & Varri , A. L. 2020, MNRAS, 492, 6019

2020
[38]

1971, , 14, 151

H \'e non , M. 1971, , 14, 151

1971
[39]

& Ostriker , J

Hernquist , L. & Ostriker , J. P. 1992, ApJ, 386, 375

1992
[40]

2010, MNRAS, 407, 355

Heyvaerts , J. 2010, MNRAS, 407, 355

2010
[41]

Hunt , J. A. S., Bub , M. W., Bovy , J., et al. 2019, MNRAS, 490, 1026

2019
[42]

C., Petersen , M

Johnson , A. C., Petersen , M. S., Johnston , K. V., & Weinberg , M. D. 2023, MNRAS, 521, 1757

2023
[43]

Julian , W. H. & Toomre , A. 1966, ApJ, 146, 810

1966
[44]

Kalnajs , A. J. 1976, ApJ, 205, 745

1976
[45]

2024, , 968, L15

Kuhn , V., Guo , Y., Martin , A., et al. 2024, , 968, L15

2024
[46]

& Kalnajs , A

Lynden-Bell , D. & Kalnajs , A. J. 1972, MNRAS, 157, 1

1972
[47]

2007, MNRAS, 381, 1663

Magorrian , J. 2007, MNRAS, 381, 1663

2007
[48]

1996, AJ, 111, 2462

Merritt , D. 1996, AJ, 111, 2462

1996
[49]

1963, MNRAS, 126, 553

Mestel, L. 1963, MNRAS, 126, 553

1963
[50]

Miller , R. H. 1971, , 14, 73

1971
[51]

S., Roule , M., Fouvry , J.-B., Pichon , C., & Tep , K

Petersen , M. S., Roule , M., Fouvry , J.-B., Pichon , C., & Tep , K. 2024, MNRAS, 530, 4378

2024
[52]

2023, KITP (Santa Barbara), doi:10.26081/K6BD50

Pichon, C. 2023, KITP (Santa Barbara), doi:10.26081/K6BD50

2023 doi
[53]

Polyachenko , E. V. 2013, Astro. Lett., 39, 72

2013
[54]

S., et al

Reddish , J., Kraljic , K., Petersen , M. S., et al. 2022, , 512, 160

2022
[55]

& Oberman , C

Rogister , A. & Oberman , C. 1968, J. Plasma Phys., 2, 33

1968
[56]

2022, , 106, 044118

Roule , M., Fouvry , J.-B., Pichon , C., & Chavanis , P.-H. 2022, , 106, 044118

2022
[57]

B., Breen , P

Rozier , S., Fouvry , J. B., Breen , P. G., et al. 2019, MNRAS, 487, 711

2019
[58]

& Laurikainen , E

Salo , H. & Laurikainen , E. 2000, MNRAS, 319, 393

2000
[59]

Sellwood , J. A. 2012, ApJ, 751, 44

2012
[60]

Sellwood , J. A. 2014, Rev. Mod. Phys., 86, 1

2014
[61]

Sellwood , J. A. 2020, MNRAS, 492, 3103

2020
[62]

Sellwood , J. A. 2024, MNRAS, 529, 3035

2024
[63]

Sellwood , J. A. & Binney , J. J. 2002, MNRAS, 336, 785

2002
[64]

Sellwood , J. A. & Evans , N. W. 2001, ApJ, 546, 176

2001
[65]

Sellwood , J. A. & Kahn , F. D. 1991, MNRAS, 250, 278

1991
[66]

2022, MNRAS, 514, 875

Tep , K., Fouvry , J.-B., & Pichon , C. 2022, MNRAS, 514, 875

2022
[67]

Tep , K., Pichon , C., & Petersen , M. S. 2024, arXiv, 2412.15033

2024 arXiv
[68]

1998, , 330, 1180

Theis , C. 1998, , 330, 1180

1998
[69]

1981, in Structure and Evolution of Normal Galaxies, 111

Toomre , A. 1981, in Structure and Evolution of Normal Galaxies, 111

1981
[70]

H., Coronado , J., & Rix , H.-W

Trick , W. H., Coronado , J., & Rix , H.-W. 2019, MNRAS, 484, 3291

2019
[71]

Weinberg , M. D. 1993, ApJ, 410, 543

1993
[72]

Weinberg , M. D. 1994, ApJ, 421, 481

1994
[73]

Weinberg , M. D. 2001, MNRAS, 328, 311

2001
[74]

Weinberg , M. D. 2015, arXiv, 1508.06855

2015 arXiv
[75]

Zang , T. A. 1976, PhD thesis, Massachusetts Institute of Technology

1976

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.

[1] [1]

Aarseth , S. J. 1963, , 126, 223

work page 1963

[2] [2]

Benetti , F. P. C. & Marcos , B. 2017, , 95, 022111

work page 2017

[3] [3]

Bezanson, J., Edelman, A., Karpinski, S., & Shah, V. B. 2017, SIAM Review, 59, 65

2017

[4] [4]

& Tremaine , S

Binney , J. & Tremaine , S. 2008, Galactic Dynamics: Second Edition (Princeton Univ. Press)

work page 2008

[5] [5]

Bouchet , F. 2020, J. Stat. Phys., 181, 515

work page 2020

[6] [6]

2008, , 78, 040102

Campa , A., Chavanis , P.-H., Giansanti , A., & Morelli , G. 2008, , 78, 040102

work page 2008

[7] [7]

2004, , 352, 427

Chakrabarty, D. 2004, , 352, 427

work page 2004

[8] [8]

1943, ApJ, 97, 255

Chandrasekhar , S. 1943, ApJ, 97, 255

work page 1943