An odd-period pedal-antipedal area invariant in elliptic billiards
ORCID: 0009-0001-9320-500X · Zenodo record and DOI
A closed billiard orbit can move around an ellipse while every side remains tangent to the same smaller confocal ellipse. The two polygons considered here are built from the original orbit: its perpendicular feet from the center, and intersections of lines perpendicular to its vertex radii. For least odd periods, their signed areas can change as the orbit moves, but their product stays fixed.
Original manuscript abstract
We prove invariant k403,a observed by Reznik, Garcia and Koiller: for every primitive odd-period elliptic billiard family with a strict confocal elliptical caustic, the signed areas of the original orbit's origin pedal and origin antipedal have constant product. All admitted coprime star windings are included, and every real antipedal intersection is finite. A complete pole analysis distinguishes coincident from opposite complex endpoints. Odd-period arithmetic then places the antipedal area and the credited pedal trace in complementary one-dimensional elliptic-function spaces. A direct argument covers circles; an exact four-periodic example rules out dropping the odd restriction.
Read the paper
Open the original PDF · Download the PDF · BibTeX
The unchanged manuscript is dated 13 October 2026. The Zenodo record was created that day in UTC and identifies version 1.0; the manuscript does not print a version number.
The two constructions and the invariant
Put the ellipse's center O at the origin. For an ordered orbit P0, …, PN−1, let qj be the perpendicular foot from O to the line through Pj and Pj+1. These feet form the origin-pedal polygon. To form the origin-antipedal polygon, intersect consecutive lines
Lj = {X : Pj · X = Pj · Pj}, Rj = Lj ∩ Lj+1.
Each Lj passes through the original vertex and is perpendicular to its radius. Both constructions use the original orbit. Write AO for the signed shoelace area of the feet and AO* for that of the intersections. Signed area keeps the vertex order and also applies to stars and self-crossing polygons.
Theorem. Fix a noncircular ellipse with semiaxes a > b > 0 and a strictly nested, nondegenerate confocal elliptical caustic. In every billiard family of least odd period N ≥ 3,
AO AO* = constant along the family.
Every admitted coprime star winding is included, and all consecutive real antipedal intersections are finite. The statement does not require either derived signed area to be nonzero. Reversing the orbit changes the sign of both areas, preserving their product.
Why odd period matters
The proof uses classical Jacobi elliptic coordinates for the billiard. It expresses the actual antipedal area as a sum over consecutive orbit vertices, then follows every possible complex pole of that expression. Coincident finite endpoints give removable singularities; opposite finite endpoints may leave simple poles. At a pole vertex, the two incident terms cancel their double-pole parts.
For a least odd period and a coprime winding, the remaining poles fall into the same two classes on a smaller complex torus. The area is consequently a constant multiple of a Jacobi-function trace. The credited earlier origin-pedal argument gives a complementary translate of that trace. Their half-period shift makes the product constant. The proof never divides by either signed area.
The all-period claim rests on this analytic argument. Exact examples and numerical diagnostics in the archive check constructions and limiting behavior; finite computations do not replace the proof, and high-precision diagnostics are not interval certificates.
Circles, repetitions and the even-period boundary
The theorem concerns the least period. Repeating an admissible cycle r times multiplies each signed area by r, so its product scales by r2. Repetition does not change the primitive-period hypothesis. Hyperbolic and collapsed caustics are outside the stated domain.
A circle is handled directly. For a circle of radius R and a primitive forward star with step θ = 2πτ/N in (0, π), the derived radii are R cos(θ/2) and R/cos(θ/2). Their signed area product is
N2 R4 sin2(θ) / 4.
The odd restriction for noncircular ellipses cannot simply be removed. With a2 = 5 and b2 = 3, a strict four-periodic caustic has squared semiaxes 25/8 and 9/8. An axis diamond and a rectangle in that same reflected family have products 225/4 and 289/4. They provide an exact counterexample to an unrestricted even-period extension.
The observation and the prior mathematics
Reznik, Garcia and Koiller recorded this original-orbit product as k403,a in their 2020 invariant table and in the published 2021 Table 5. The paper proves that observation in the domain stated above. It does not claim to have discovered the observed invariant.
Elliptic-billiard parametrization and complex methods are classical, as are cyclic Jacobi identities and Landen transformations. The origin-pedal trace was already proved in my earlier k203,b campaign argument and is restated here. The additional step is the construction-specific antipedal pole analysis, including opposite endpoints. No historical-firstness certification is claimed.
Original files and citation
The unchanged files appear in their native Zenodo order:
- odd-origin-area-product-verification.zip (170,660 bytes) · Zenodo original ZIP
- odd_origin_area_product.pdf (70,407 bytes) · Zenodo original PDF
Alec Kriebel. An odd-period pedal-antipedal area invariant in elliptic billiards. Zenodo version 1.0, 13 October 2026. doi:10.5281/zenodo.23334001. Download BibTeX.
The authored manuscript and verification materials use CC BY 4.0. Third-party papers and raw extracts are excluded from that grant, and mathematical inputs retain their stated credit.
Status and collaboration
AI tools were used extensively in solving, verification, literature research, coding, writing and internal adversarial review. This is an unrefereed preprint without external human peer review or proof-assistant formalization.
Independent mathematical verification and substantive work toward a journal submission would be useful contributions. Collaboration and any authorship would reflect actual scholarly contributions.
Credited literature
- D. Reznik, R. Garcia and J. Koiller, Fifty new invariants of N-periodics in the elliptic billiard, Arnold Mathematical Journal 7 (2021), 341–355, Table 5. Publisher DOI. The earlier table appears in arXiv:2004.12497v11.
- H. Stachel, On the motion of billiards in ellipses, European Journal of Mathematics 8 (2022), 1602–1622. Publisher DOI.
- A. Akopyan, R. Schwartz and S. Tabachnikov, Billiards in ellipses revisited, European Journal of Mathematics 8 (2022), 1313–1327. Publisher DOI.
- NIST, Digital Library of Mathematical Functions, §22.4, §22.8 and §22.9.
- A. Khare, A. Lakshminarayan and U. Sukhatme, Cyclic identities involving Jacobi elliptic functions. II, arXiv:math-ph/0207019v2; A. Khare and U. Sukhatme, Generalized Landen transformation formulas for Jacobi elliptic functions, arXiv:math-ph/0312074v1.
- A. Kriebel, The origin-pedal area product in an elliptic billiard, k203,b campaign proof, 1 October 2026. Archived proof.