On the Birkhoff conjecture for Kepler billiards

Irene De Blasi; Stefano Baranzini; Susanna Terracini; Vivina L. Barutello

REVIEW 3 major objections 4 minor 37 references

On the Birkhoff conjecture for Kepler billiards

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read This paper establishes that high-energy analytic Kepler billiards are integrable only on elliptic tables with the attraction center at a focus.

desk verdict Genuinely new and likely correct, but the high-energy asymptotic estimates that carry the non-integrability theorem are compressed and need referee scrutiny. read the letter →

arxiv 2507.08446 v1 pith:RZHAJE55 submitted 2025-07-11 math.DS

Stefano Baranzini , Vivina L. Barutello , Irene De Blasi , Susanna Terracini This is my paper · ORCID

classification math.DS MSC 37C8334C2837J5137J46

keywords KeplerbilliardsBirkhoff-Poritskyconjectureanalyticintegrabilitysymbolicdynamicstopologicalentropyfocalpointsofthesecondkindhigh-energyregimeshadowingmethod

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 asks whether a mechanical billiard with an inverse-distance gravitational force can be analytically integrable. It argues that, at high energies, the answer is essentially no for strictly convex analytic tables, unless the table is an ellipse and the attracting center sits at a focus. The proof constructs symbolic dynamics—a coding of trajectories by a full shift on two symbols—by shadowing nearly degenerate triangular billiard paths through the center. If correct, the results give a partial affirmative answer to the Keplerian analogue of the Birkhoff-Poritsky conjecture, up to one exceptional center position in non-elliptic domains.

What carries the argument

The argument is carried by the high-energy asymptotics of the Jacobi length: a direct Keplerian arc between two boundary points has length $\sqrt{h}\lvert p_0-p_1\rvert + O(\mu/\sqrt{h})$, and an indirect arc through the center has length $\sqrt{h}(\lvert p_0\rvert + \lvert p_1\rvert) + O(\mu/\sqrt{h})$ plus a logarithmic term, with variants for antipodal pairs. These expansions turn the multi-valued generating function of the Kepler billiard into $\sqrt{h}$ times a perimeter function $\Psi(\xi,\eta)=\lvert\gamma(\xi)\rvert+\lvert\gamma(\eta)\rvert+\lvert\gamma(\eta)-\gamma(\xi)\rvert$, whose critical points are punctured Birkhoff triangles. The persistence of these critical points under the high-energy perturbation is controlled by topological degree, and shadowing periodic words in the resulting symbolic system is extended to all bi-infinite sequences by a diagonal argument.

What would settle it

Find a real-analytic strictly convex non-elliptic domain admitting two distinct focal points of the second kind; the polynomial $P(t)$ derived in Lemma 4.5 constrains such points to one orthogonal chord and forces the domain to be an ellipse, so exhibiting such a pair would refute Theorem 4.1 and the rigidity half of Theorem 1.1. Equivalently, exhibit an analytic first integral for a high-energy Kepler billiard in a non-elliptic analytic table with a non-focal center.

Watch

Extended reading notes

Core claim

The central discovery is a rigidity theorem: for a strictly convex bounded planar domain with real-analytic boundary, at sufficiently high energy the Kepler billiard map cannot be analytically integrable unless the boundary is an ellipse and the center of attraction is one of its foci. For non-elliptic domains, the theorem allows at most one exceptional center position, and for ellipses the only exceptions are the two foci. This is proved by showing that, for every non-exceptional center, the high-energy first-return map contains a subsystem semi-conjugate to a Bernoulli shift, which implies positive topological entropy and the absence of analytic first integrals. A companion geometric rigidity result says that real-analytic non-elliptic domains admit at most one focal point of the second kind, while ellipses are the only domains admitting two, namely their classical foci.

Load-bearing premise

The paper relies on high-energy asymptotic expansions of the Jacobi length holding uniformly with controlled remainders and with the Keplerian arcs staying inside the strictly convex domain, so that the true generating function can be replaced by $\sqrt{h}$ times a perimeter function without losing critical points.

Editorial extensions

If this is right

If the central claim is correct, any strictly convex real-analytic non-elliptic table at high energy has positive topological entropy for the Kepler billiard map for all but possibly one center position.
At those centers, the first-return map admits no analytic first integral, so the system is not analytically integrable in the sense used here.
The only analytically integrable high-energy Kepler billiards among analytic convex tables are elliptic tables with the center at a focus, matching Panov's known integrable examples.
The rigidity theorem implies that a table with two focal points of the second kind must be an ellipse, giving a sharp geometric obstruction to the symbolic-dynamics construction.
The paper exhibits an infinite-dimensional family of non-elliptic analytic tables with a focal point of the second kind, so the exceptional case in Theorem 1.1 is real and not vacuous.

Reading between the lines

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

The same triangle-shadowing mechanism may extend to finite energies if quantitative bounds on the Jacobi-length remainders can be obtained; the paper only claims the high-energy regime.
The numerical simulations reported for a focal point of the second kind suggest that chaotic behavior persists even where the proof does not apply, so the exceptional one-center case may be a limitation of the method rather than true integrability.
The geometric part of the result—at most two focal points of the second kind, with two forcing an ellipse—stands independently of Keplerian dynamics and can be tested directly on the Birkhoff billiard map.
A natural testable extension is to push the shadowing construction to the zero-energy case, which the paper notes is conjugate to an ordinary Birkhoff billiard on the double cover of the domain.

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

This paper is worth taking seriously. The main result—high-energy analytic non-integrability for non-elliptic Kepler billiards except possibly one center position—is a real step beyond the earlier first-kind construction in [3] and beyond Panov's integrability example. The three symbolic-dynamics constructions, especially the new ones for second-kind and non-focal centers via punctured Birkhoff triangles, are clever and, as far as I can tell, sound. The focal-point rigidity section is genuinely elegant: the Hessian computation, the chord argument, and the use of invariant graphs to force ellipticity are convincing and largely independent of the high-energy asymptotics. The construction of an infinite-dimensional family of non-elliptic tables with a focal point of the second kind is a nice bonus and makes Theorem 1.1's exception sharp.

The stress-test note names the right soft spot. The whole non-integrability theorem rests on Lemmas 2.2 and 2.3 giving uniform asymptotic control of the Jacobi length for every endpoint pair used in the concatenations. Lemma 2.2 is mostly a reference to [3] with a sketch. Lemma 2.3, proved in the appendix, gives C^1 convergence but leaves the remainder at O(1), not o(1), and the uniformity across the switching locus det(p0,p1)=0 is not fully written out. The degree-persistence argument in Lemmas 3.9 and 3.10 needs the remainder to be uniformly smaller than sqrt(h) times the minimal modulus of the limiting gradient on the isolating neighborhoods; that quantitative condition is not checked. Containment inside the domain is also asserted by uniform convergence, but for direct arcs it is stated only when |p0-p1| >= epsilon, and the concatenations in Lemma 3.10 use short transfer arcs that need the same control.

I do not see an actual contradiction, and the authors may well have the estimates in hand. But this is the load-bearing step, and the paper currently treats it too casually. Lemma 4.9's unique-continuation step is quick, though plausible. These are addressable gaps rather than fatal flaws.

My take: the central claim is likely correct, the rigidity half is solid, and the paper deserves a serious referee. I would send it out and ask the referee to push hard on Lemmas 2.2 and 2.3, and to make the isolating neighborhoods in Lemmas 3.9 and 3.10 explicit. If that section tightens up, this becomes a strong paper.

Referee Report

3 major / 4 minor

Summary. The paper studies analytic integrability of Kepler billiards in strictly convex real-analytic planar domains at high energies. The main result (Theorem 1.1) states that, apart from at most one position of the attraction center (two for an ellipse, namely its foci), the Kepler billiard is not analytically integrable; the mechanism is the construction, for all centers of the first kind or non-focal centers, of a subsystem of the first-return map that is semi-conjugate to a Bernoulli shift (Theorem 3.4 and Corollary 3.11), yielding positive topological entropy and no analytic first integral. The shadowing is based on high-energy asymptotics of the Jacobi length of direct/indirect/clockwise/anticlockwise Keplerian arcs, with limiting perimeter functionals Psi, Psi_a, Psi_c; critical points are propagated by topological degree. The second half of the paper classifies focal points of the second kind, showing that there are at most two, with equality only for ellipses at their foci, and constructs an infinite-dimensional family of non-elliptic domains with one focal point, supported by numerics.

Significance. If the central arguments are completed, this is a substantial contribution to the Keplerian analogue of the Birkhoff-Poritsky conjecture. It introduces a clean three-case shadowing strategy, uses topological-degree persistence in an essential way, and separates a rigidity problem about focal points that is interesting in its own right. The paper is honest about the dependence on [3] for the first-kind case and about the limitation to high energies; the exceptional focal case is explicitly identified rather than hidden. The main weaknesses are that two load-bearing analytic steps are compressed (uniform asymptotic remainder estimates and the passage from an invariant-graph identity to ellipsoidal rigidity) and that the claimed construction of non-elliptic focal domains is not actually proved. The result is, in my assessment, likely correct in broad outline, but the manuscript needs a substantial revision before it can be accepted.

major comments (3)

[Section 2, Lemma 2.3 and Appendix A] The proof of Lemma 2.3 does not fully establish the uniform first-order asymptotic used in Lemma 3.10. The statement requires g_a and g_c to be C^1 and uniformly bounded in p0,p1, and the derivation in Appendix A obtains the leading term of \nabla_{p1} L(z_a) from an expansion y = y0 + sqrt(h') g with y0 piecewise defined at <w0,w1>=0. The passage "by regularity of the involved functions" from this expansion to uniform C^1 bounds for the remainder is not justified near that switching locus, where the chosen square-root branch w1 = +/- sqrt(p1) changes; and the uniform control for h' -> 0 is only sketched. Since the degree-persistence argument in Lemma 3.10 requires the remainder in (15) to be uniformly O(1) (i.e., o(sqrt(h))) on the whole product of the neighborhoods I_m, I_M, I_P, I_Q and for every periodic word s, this is a load-bearing point. Please provide a complete proof or a precise reference for the uniform C^1 estimate across det(p0,p1)=0.
[Section 4.1, Proposition 4.10] The construction of non-elliptic domains with a focal point of the second kind is not proved. After defining gamma as a level set of f(x)=|x-c|+d(x,tau), the proof asserts that "clearly" phi_c is constant because all chords orthogonal to T have the same length; this does not follow, since phi_c is defined by two consecutive reflections on gamma and involves the three distances |gamma(xi)-c|, |gamma(xi)-gamma(xi')| and |gamma(xi')-c|, not the chord lengths of tau. The intermediate claim that gamma has a unique periodic trajectory through c is also insufficient. This gap affects the advertised infinite-dimensional family and the sharpness discussion of Theorem 1.1, so it should either be fixed with a real proof or the corresponding claims should be weakened.
[Section 4, Lemma 4.9 and proof of Theorem 4.1] The rigidity step at the end of Lemma 4.9 and in the proof of Theorem 4.1 is too compressed. From the identity delta_+(S^1) = delta^c_-(S^1) the manuscript concludes that partial Omega is an ellipse with foci 0 and c, saying that the family of invariant lines consists of the two pencils through the foci and draws triangles of equal perimeter. This is a nontrivial characterization and no proof is supplied. Because Theorem 4.1 is what limits the exceptional center positions in Theorem 1.1, this step is load-bearing; it should be stated as a lemma with a complete argument or an explicit reference.

minor comments (4)

[Abstract and Section 3] The notation "focal points of the second kind" is used in the abstract before its definition in Definitions 3.1 and 3.3; consider adding a forward reference.
[Section 3.2.2, Lemma 3.10] The derivative formula (15) is stated only for k=1,2; the analogous formula for the third coordinate u_{i,3} is needed for the critical-point equation and should be included.
[Throughout] There are several typos ("critival" in Appendix B, "anergy" in Corollary 1.5, "world" in Section 3.3), and the reference to Figure 15 in Section 4.1 does not mention the numerical integration method; these should be corrected.
[Section 3.1, Lemma 3.5] The function B in the proof is introduced without specifying its domain and regularity; a brief justification (or a local version via the implicit function theorem) would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No definitional circularity or fitted-input prediction is present; the paper's new symbolic-dynamics and rigidity arguments are derived in-paper, and reliance on [3] is a disclosed prior-theorem dependency rather than a reduction to its own inputs.

full rationale

The high-energy symbolic-dynamics mechanism for non-focal centers is self-contained: Lemmas 2.2 and 2.3 supply asymptotic expansions of Jacobi lengths, and Lemmas 3.9 and 3.10 prove persistence of critical points of the total generating function by comparing the gradient of the true action with sqrt(h) times the gradient of the limiting perimeter functional and applying topological-degree homotopy arguments. No parameter is fitted to integrability data, and no quantity that is meant to be predicted is built into the assumptions. The focal-point counting in Section 4 is an independent geometric and rigidity argument, leading to the at-most-one / exactly-two classification. The main self-citations are Lemma 3.2 (importing the first-kind case from [3]), the indirect-arc C^2 estimate in Lemma 2.2, and the diagonal extension in the proof of Theorem 3.4. These are prior published results by overlapping authors, but they have stated assumptions that do not include the present integrability conclusion and are not equivalent by construction to the claimed theorems; the paper explicitly discloses that the first construction was already carried out in [3]. A failure of the uniform remainder estimates in Lemmas 2.2/2.3 would be a correctness gap, not a circularity. Under the hard rules, such theorem citations count as real evidence and do not raise the circularity score.

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

No new physical entities are postulated. 'Punctured Birkhoff billiard' and 'focal point of the second kind' are mathematical definitions used inside the proof, not independently evidenced objects. No free parameters are fitted to data; the numerical simulation uses illustrative physical parameters.

assumptions (6)

standard math Panov's theorem: elliptic Kepler billiards with center at a focus are integrable at all energy levels
Used in Theorem 1.1 to identify the two exceptional positions for ellipses; stated in the Introduction.
standard math Lemma 3.2 from [3]: a center of the first kind yields symbolic dynamics at high energy
Invoked in Theorem 3.4 and Corollary 3.11; this is prior published work by overlapping authors.
standard math Positive topological entropy of a subsystem implies the absence of nonconstant analytic first integrals
Used in Corollary 3.11; cited to [17], [24], and [3, Theorem 4.5].
standard math Hopf index theorem and topological degree product and homotopy invariance
Central to Proposition 3.7 and Lemmas 3.9 and 3.10 for persistence of critical points under high-energy perturbation.
domain assumption Analyticity of the Birkhoff billiard map and unique continuation for analytic curves
Used in Lemma 4.9 to glue local stable and unstable manifold coincidences into global equality of invariant graphs; relies on the real-analytic boundary.
standard math Jacobian formula for the Birkhoff map at a 2-periodic orbit
Used in Lemma 4.8; quoted from [22, Theorem 4.2, Part V].

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the Birkhoff conjecture for Kepler billiards." pith.science (2026). https://pith.science/paper/RZHAJE55

@misc{pith2026250708446,
  author       = {Pith},
  title        = {Pith review of: On the Birkhoff conjecture for Kepler billiards},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RZHAJE55}},
  note         = {Machine review of arXiv:2507.08446}
}

read the original abstract

We investigate the integrability of Kepler billiards-mechanical billiard systems in which a particle moves under the influence of a Keplerian potential and reflects elastically at the boundary of a strictly convex planar domain. Our main result establishes that, except possibly for one location of the gravitational center, analytic integrability at high energies occurs only when the domain is an ellipse and the center is placed at one of its foci. This provides a partial affirmative answer to a Keplerian analogue of the classical Birkhoff-Poritsky Conjecture. Our approach is based on the construction of symbolic dynamics arising from chaotic subsystems that emerge in the high-energy regime. Depending on the geometric configuration of the boundary and the location of the attraction center, we construct three types of symbolic dynamics by shadowing chains of punctured Birkhoff-type trajectories. These constructions yield subsystems conjugated to Bernoulli shifts, implying positive topological entropy and precluding analytic integrability. We further analyze the notion of focal points of the second kind, showing that in real-analytic, non-elliptic domains there can be at most one such point, while ellipses are the only domains admitting two-coinciding with their classical foci. Finally, we demonstrate the existence of an infinite-dimensional family of non-elliptic domains possessing a focal point of the second kind, and conclude with numerical simulations illustrating chaotic behavior in such cases.

Figures

Figures reproduced from arXiv: 2507.08446 by the authors.

**Figure 1.** Kepler billiard with gravitational center at c and positive energy. raises fundamental questions concerning integrability, existence, and stability of periodic orbits, ergodic properties, and the structure of the corresponding billiard map, which encodes the sequence of impact points and directions. Let Ω ⊂ R 2 be a bounded domain with smooth boundary ∂Ω. At a point of impact p ∈ ∂Ω, the particle has a unit inward-p… view at source ↗

**Figure 2.** Elements to shadow in the three different constructions of symbolic dynamics. Corollary 1.5. Under the same assumptions of Theorem 1.4, assume moreover that Ω possesses a central or a cyclic rotational symmetry. Then, if ∂Ω is not an ellipse, for any choice of the gravitational center there is h > 0 such that, for any anergy h ≥ h the associated Kepler billiard map has a subsystem displaying a symbolic dynamics. The… view at source ↗

**Figure 3.** Direct and indirect arcs connecting two points in R 2 \ {0}. p0, p1 ∈ ∂Ω can be connected by two solutions, possibly regularized, of the problem (2)    z ′′(t) = ∇V (z(t)), t ∈ [0, T], 1 2 |z ′ (t)| 2 − V (z(t)) = h, t ∈ [0, T], z(0) = p0, z(T) = p1, for a suitable T > 0. While in the classical Birkhoff case segments connecting points of ∂Ω are contained in Ω thanks to its strict convexity, this is not the ca… view at source ↗

Figures from the paper (15 more)

**Figure 4.** Figure 4: In the first line, the anticlockwise, za, and clockwise, zc, arcs connecting two (possibly antipodal) points. Depending on the difference in the arguments between p0 and p1, they correspond either to direct or indirect arcs. In the second line, a family of anticlockwi…

**Figure 5.** Figure 5: Shadowing of a Birkhoff triangle (in light grey): a vertex is at c, the other two on ∂Ω and the bounces are elastic. Definition 3.1. A point c ∈ Ω is said to be of the first kind if the distance function |γ(·) − c| has at least one pair of strict local extremal (minima…

**Figure 6.** Figure 6: Possible concatenations of triangular trajectories. Left: a concatenation of arcs realizing a 2-periodic word with periodicity modulus T T′ . Right: a concatenation of arcs realizing a periodicity modulus of a 3-periodic world with s0 = s1 = T ′ and s2 = T. As one can…

**Figure 7.** Figure 7: Degenerate case: the minimum and maximum of |γ(·)| correspond to antipodal points, and there is a second pair of antipodal homothetic directions corresponding to inflection points for |γ(·)|. The piecewise curve with arrows is an example of transfer concatenation from …

**Figure 8.** Figure 8: If 0 is focal, then all triangles with a vertex at the origin and one reflection on the boundary satisfy the elastic reflection condition also at the second bouncing point. These triangles have all the same perimeter, the value of φ. Corollary 4.2. Assume that ∂Ω is n…

**Figure 9.** Figure 9: Left, the vector field ˆδ introduced in the proof of Lemma 4.6. Right, following the proof of Lemma 4.6, every double-bouncing triangle has a vertex on ∂Ω + and the second on ∂Ω −. non-trivial curve in the phase cylinder, δ, which is invariant and piece-wise analytic. …

**Figure 10.** Figure 10: An illustration of the proof of Lemma 4.8. The curves δ± are contained in the stable/unstable set of p1 and p2. 0 π 2 π 3 π 2 2 π π 2 π ξ α [PITH_FULL_IMAGE:figures/full_fig_p020_10.png]

**Figure 11.** Figure 11: Phase portrait of the second iterate T 2 , choosing Ω as in Section 4.1. In red, the invariant curves δ±(S 1 ), see Lemmas 4.7 and 4.9. Proof of Theorem 4.1. From Lemma 4.5 we know that Ω has at most two focal points of the second kind. If there are exactly two, say 0…

**Figure 12.** Figure 12: Left, the curves T and α as in Eq. (20) when a0 = 1, a5 = 1/5, with ak = 0 when k ̸= 0, 5. Right, the curves T and α as in Eq. (20) when a0 = 1, a7 = 1/7, with ak = 0 when k ̸= 0, 7. Proposition 4.10. For any analytic curve T as in (20) bounding a convex constant-widt…

**Figure 13.** Figure 13: Left, the curves T, α and γ as in Eqs. (20) and (21) when a0 = 1, a3 = 1/3, with ak = 0 when k ̸= 0, 3. The point c is located at (3, 0) and ˜ℓ = 6. Center, a point x belongs to γ if and only if the sum of length of the dashed segments is constantly equal to ˜ℓ (see E…

**Figure 14.** Figure 14: In the same setting of [PITH_FULL_IMAGE:figures/full_fig_p023_14.png]

**Figure 15.** Figure 15: First return map associated to the Kepler billiard when the boundary is constructed as in [PITH_FULL_IMAGE:figures/full_fig_p023_15.png]

**Figure 16.** Figure 16: The anticlockwise and clockwise arcs connecting two (possibly antipodal) points. Depending on the difference in the arguments between p0 and p1, they correspond either to direct or indirect arcs. If w(τ ) is the regularized solution corresponding to za, one has that …

**Figure 17.** Figure 17: The open set U (left) is diffeomorphic to an open cylinder whose boundary components correspond to two copies of ∆ (right). The vector field ∇Ψ extends continuously to the closure of such cylinder, Σ, and points inward along its boundary. Thus, ∂ξΨ = 0 if and only if …

**Figure 18.** Figure 18: An illustration of the proof of Lemma 3.8. The dotted curve in the left picture is the graph of η = η(ξ) such that γ(ξ) and γ(η(ξ)) are antipodal on ∂Ω. Its image trough ∇Ψ and ∇Ψ ∗ is the same. For this reason its intersections with both ∇Ψ ◦ c and ∇Ψ ∗ ◦ c coincide.…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 36 canonical work pages

[3]

V. L. Barutello, I. De Blasi, and S. Terracini. Chaotic dynamics in refraction galactic billiards.Nonlinearity, 36(8):4209, 2023

work page 2023
[1]

Avila, J

A. Avila, J. De Simoi, and V. Kaloshin. An integrable deformation of an ellipse of small eccentricity is an ellipse. Annals of Mathematics, pages 527–558, 2016

work page 2016
[2]

V. L. Barutello, A. M. Cherubini, and I. De Blasi. Exploration of billiards with keplerian potential.Non- linearity, 38(5):055004, 2025

work page 2025
[4]

R. H. Battin.An introduction to the mathematics and methods of astrodynamics. AIAA Education Series. American Institute of Aeronautics and Astronautics (AIAA), Reston, VA, revised edition, 1999. With a foreword by J. S. Przemieniecki

work page 1999
[5]

M. Bialy. Convex billiards and a theorem by E. Hopf.Math. Z., 214(1):147–154, 1993

work page 1993
[6]

Bialy and A

M. Bialy and A. E. Mironov. The Birkhoff-Poritsky conjecture for centrally-symmetric billiard tables.Ann. Math. (2), 196(1):389–413, 2022

work page 2022
[7]

G. D. Birkhoff. On the periodic motions of dynamical systems.Acta Mathematica, 50 (1):359–379, 1927

work page 1927
[8]

S. V. Bolotin. Degenerate billiards in celestial mechanics.Regul. Chaotic Dyn., 22(1):27–53, 2017

work page 2017

Show all 37 references

[9]

S. V. Bolotin and R. S. Mackay. Periodic and chaotic trajectories of the second species for the n-centre problem. Celestial Mechanics and Dynamical Astronomy, 77(1):49–75, 2000

2000
[10]

Boltzmann

L. Boltzmann. Lösung eines mechanichen problems.Wiener Berichte, 58:1035–1044, 1868
[11]

R. L. Devaney.An introduction to chaotic dynamical systems. Studies in Nonlinearity. Westview Press, Boulder, CO, 2003. Reprint of the second (1989) edition

1989
[12]

Y. N. Fedorov. An ellipsoidal billiard with a quadratic potential.Functional Analysis and Its Applications, 35(3):199–208, 2001

2001
[13]

G. Felder. Poncelet property and quasi-periodicity of the integrable boltzmann system.Lett. Math. Phys., 111(1):12, Feb. 2021

2021
[14]

Gallavotti and I

G. Gallavotti and I. Jauslin. A theorem on ellipses, an integrable system and a theorem of boltzmann. ArXiv preprint arXiv:2008.01955, 2020

2008 arXiv
[15]

Gasiorek and M

S. Gasiorek and M. Radnović. Periodic trajectories and topology of the integrable boltzmann system. Contemporary Mathematics, 807:111–130, 2024

2024
[16]

Guillemin and A

V. Guillemin and A. Pollack.Differential Topology. AMS Chelsea Publishing: An Imprint of the American Mathematical Society, 1974

1974
[17]

Hasselblatt and A

B. Hasselblatt and A. Katok.A first course in dynamics. Cambridge University Press, New York, 2003

2003
[18]

Huang, V

G. Huang, V. Kaloshin, and A. Sorrentino. Nearly circular domains which are integrable close to the boundary are ellipses.Geom. Funct. Anal., 28(2):334–392, 2018

2018
[19]

C. G. J. Jacobi. Zur theorie der variations-rechnung und der differential-gleichungen.Journal für die reine und angewandte Mathematik, 1837. ON THE BIRKHOFF CONJECTURE FOR KEPLER BILLIARDS 30
[20]

Kaloshin and A

V. Kaloshin and A. Sorrentino. On the integrability of Birkhoff billiards. Philos. Trans. Roy. Soc. A, 376(2131):20170419, 16, 2018

2018
[21]

Kaloshin and A

V. Kaloshin and A. Sorrentino. On the local Birkhoff conjecture for convex billiards.Ann. of Math. (2), 188(1):315–380, 2018

2018
[22]

Katok and J.-M

A. Katok and J.-M. Strelcyn.Invariant manifolds, entropy and billiards. Smooth maps with singularities, volume 1222. Springer, 2006

2006
[23]

O. Knill. On nonconvex caustics of convex billiards.Elem. Math., 53(3):89–106, 1998

1998
[24]

V. Kozlov. Integrability and nonintegrability in Hamiltonian mechanics.Uspekhi Mat. Nauk, 38(1 (229)):3– 67, 240, 1983

1983
[25]

V. Kozlov. Some integrable extensions of jacobi’s problem of geodesics on an ellipsoid.Journal of Applied Mathematics and Mechanics, 59(1):1–7, 1995

1995
[26]

Krawcewicz and J

W. Krawcewicz and J. Wu.Theory of degrees with applications to bifurcations and differential equations. Series of Monographs and Advanced Texts. Canadian Mathematical Society, John Wiley & Sons, New York, 1997

1997
[27]

Levi-Civita

T. Levi-Civita. Sur la résolution qualitative du problème restreint des trois corps.Acta Math., 30(1):305– 327, 1906

1906
[28]

Montgomery

R. Montgomery. Minimizers for the Kepler problem.Qual. Theory Dyn. Syst., 19(1):31, 2020

2020
[29]

A. A. Panov. Elliptical billiard table with newtonian potential.Mathematical Notes, 55(3):334–334, 1994

1994
[30]

Poritsky

H. Poritsky. The billiard ball problem on a table with a convex boundary - An illustrative dynamical problem. I. The invariant integral.Ann. Math. (2), 51:446–470, 1950

1950
[31]

S. E. Pustovoitov. Topological analysis of a billiard in elliptic ring in a potential field.Fundamentalnaya i Prikladnaya Matematika, 22(6):201–225, 2019

2019
[32]

S. E. Pustovoitov. Topological analysis of a billiard bounded by confocal quadrics in a potential field. Sbornik: Mathematics, 212(2):211, 2021

2021
[33]

Tabachnikov.Geometry and billiards, volume 30 ofStudent Mathematical Library

S. Tabachnikov.Geometry and billiards, volume 30 ofStudent Mathematical Library. American Mathemat- ical Society, Providence, RI, 2005

2005
[34]

Takeuchi and L

A. Takeuchi and L. Zhao. Conformal transformations and integrable mechanical billiards.Advances in Mathematics, 436:109411, 2024

2024
[35]

Takeuchi and L

A. Takeuchi and L. Zhao. Projective integrable mechanical billiards.Nonlinearity, 37(1), 2024

2024
[36]

A. Wintner. The Analytical Foundations of Celestial Mechanics. Princeton Mathematical Series, vol.5. Princeton University Press, Princeton, N.J., 1941

1941
[37]

L. Zhao. Projective dynamics and an integrable Boltzmann billiard model.Communications in Contempo- rary Mathematics, 24(10):2150085, 2022. University of Turin, Department of Mathematics, via Carlo Alberto 10, Turin, Italy

2022

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [3]

V. L. Barutello, I. De Blasi, and S. Terracini. Chaotic dynamics in refraction galactic billiards.Nonlinearity, 36(8):4209, 2023

work page 2023

[2] [1]

Avila, J

A. Avila, J. De Simoi, and V. Kaloshin. An integrable deformation of an ellipse of small eccentricity is an ellipse. Annals of Mathematics, pages 527–558, 2016

work page 2016

[3] [2]

V. L. Barutello, A. M. Cherubini, and I. De Blasi. Exploration of billiards with keplerian potential.Non- linearity, 38(5):055004, 2025

work page 2025

[4] [4]

R. H. Battin.An introduction to the mathematics and methods of astrodynamics. AIAA Education Series. American Institute of Aeronautics and Astronautics (AIAA), Reston, VA, revised edition, 1999. With a foreword by J. S. Przemieniecki

work page 1999

[5] [5]

M. Bialy. Convex billiards and a theorem by E. Hopf.Math. Z., 214(1):147–154, 1993

work page 1993

[6] [6]

Bialy and A

M. Bialy and A. E. Mironov. The Birkhoff-Poritsky conjecture for centrally-symmetric billiard tables.Ann. Math. (2), 196(1):389–413, 2022

work page 2022

[7] [7]

G. D. Birkhoff. On the periodic motions of dynamical systems.Acta Mathematica, 50 (1):359–379, 1927

work page 1927

[8] [8]

S. V. Bolotin. Degenerate billiards in celestial mechanics.Regul. Chaotic Dyn., 22(1):27–53, 2017

work page 2017