Left-handed geodesic flow of spheres of revolution

Pierre Dehornoy, Ana Rechtman · 2025

Read Left-handed geodesic flow of spheres of revolution by Pierre Dehornoy, Ana Rechtman (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-19979.txt.

Introduction

Pierre Dehornoy and Ana Rechtman study the motion of geodesics on rounded surfaces in this two thousand twenty-five math paper. A geodesic is the straightest possible path on a curved surface. The flow of all geodesics can be called left-handed when any two steady patterns link with a negative twist. The authors work on spheres of revolution, surfaces spun from a curve. They require the surface's curvature to be pinched within set bounds.

They prove that with the tightest pinching, the geodesic flow is left-handed. For any looser pinching, they build a surface where the flow fails to be left-handed. The split supports a conjecture that the bound they use is the best possible. The paper serves dynamicists who study flows on three-dimensional spaces.

Catalog record

CREATORPierre Dehornoy, Ana Rechtman
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE52 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-19979

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Left-handed geodesic flow of spheres of revolution Pierre Dehornoy, Ana Rechtman Abstract . A flow on a 3-manifold is left-handed if any two ergodic invariant measures have negative linking number. We prove that on a 2-sphere of revolution whose curvature is 1 / 4 1/4 -pinched the geodesic flow is left-handed. Conversely, for every δ ≤ 1 / 4 \delta\leq 1/4 , we construct a 2-sphere whose curvature is δ \delta -pinched and whose geodesic flow is not left-handed. This gives credit to a conjecture of Florio and Hryniewicz that 1 / 4 1/4 is the optimal pinching constant for left-handedness among arbitrary positively curved 2-spheres. 1. Introduction The aim of this note is to study the condition of left-handedness for geodesic flows of spheres of revolution, or in other words genus zero surfaces of revolution. Left-handedness relies on the notion of linking of pairs of invariant measures, and corresponds to constant sign of these linking numbers. Since the flows we consider here are integrable, they are left-handed if, roughly speaking, every pair of periodic orbits has negative linking number. Left-handedness, introduced by É. Ghys [ 5 ] , implies in particular that every finite collection of periodic orbits bounds a Birkhoff section. We proved that, in negative curvature, left-handedness of the geodesic flows occurs for triangular orbifolds only [ 2 ] . Here, we relate left-handedness to the pinching of the curvature: the quotient \delta(S)=\frac{K_{min}}{K_{max}}\leq 1 of the extreme values of the Gaussian curvature of the surface S . We say that the curvature is \delta -pinched. A. Florio and U. Hryniewicz proved that, in general, if the Gaussian curvature of a 2-sphere satisfies that \delta>0.7225 , then the geodesic flow on its unit tangent bundle is left-handed [ 4 ] 1 1 1 Florio and Hryniewicz adopt a convention different than ours for linking numbers of lifts of geodesics to the unitary tangent bundle, which leads to a change of sign. Hence their flows are right-handed when ours are left-handed. . They conjectured that the good bound for \delta should be 1/4 . Our main result is that this is the optimal bound among spheres of revolution. Theorem 1.1 . Let S be a sphere of revolution in \mathbb{R}^{3} with the induced metric. If \delta(S)>1/4 , the geodesic flow on \mathrm{T}^{1}S is left-handed. Conversely, for every \delta\leq 1/4 , there exists a \delta -pinched sphere of revolution in \mathbb{R}^{3} whose geodesic flow is not left-handed (nor right-handed). The proof relies on a direct computation of the linking numbers between any two closed geodesics and on the fact that the geodesic flow admits an invariant function given by Clairaut’s identity, see Section 4.1 . A “figure-eight” geodesic on a surface of revolution is a closed geodesic that turns twice around the axis of revolution and intersects the equator at one point with multiplicity two. Lifts of figure-eight closed geodesic and the equator have zero linking number (see Proposition 5.2 ). We say that the surface has no asymptotic figure-eight geodesic if there exists a>0 such that for every closed geodesic \gamma and two points p_{1},p_{2} in the intersection of \gamma with the equator that are consecutive along \gamma , the distance between p_{1} and p_{2} is at least a . A figure-eight yields an asymptotic figure-eight, but the converse is not true in general. The key point for proving Theorem 1.1 is the following: Proposition 1.2 . Let S be a sphere of revolution in \mathbb{R}^{3} , then the geodesic flow on \mathrm{T}^{1}S is left-handed if and only if there is no asymptotic figure-eight geodesic on S . Applying this criterion to ellipsoids of revolution yields in particular Proposition 1.3 . Let \mathbb{E}_{b} be the ellipsoid of revolution with axes (1,1,b) . The geodesic flow of \mathrm{T}^{1}\mathbb{E}_{b} is left-handed if and only if b<2 . Since the pinching \delta(\mathbb{E}_{b}) is equal to \frac{1}{b^{4}} , this proposition implies the second part of Theorem 1.1 for \delta\leq\frac{1}{16} . For \delta in (\frac{1}{16},\frac{1}{4}] , we note that the (asymptotic) figure-eight in a pinched ellipsoid lies in a small neighbourhood of the equator and in particular does not use all the freedom brought by the pinching. Our proof of the second part of the theorem is obtained by smoothly gluing a neighborhood of the equator on \mathbb{E}_{1/\sqrt{\delta}} to a piece of round sphere. 2. Linking numbers and left-handed flows As mentioned before, a flow is left-handed if all invariant measures have negative linking number. In this section, we review some notions and results that allow to make this definition precise. 2.1. Linking number of links Let M be a 3-dimensional compact manifold that is a rational homology sphere. Then the integral homologygroup \mathrm{H}_{1}(M;\mathbb{Z}) is finite, in particular it has an exponent e which is a positive integer. Given an oriented link L_{1} , its multiple eL_{1} bounds a surface S_{1} . Given another link L_{2} disjoint from L_{1} , their linking number , denoted by \mathrm{Lk}(L_{1},L_{2}) , is the algebraic intersection between S_{1} and L_{2} , divided by e . Thanks to the nullity of [L_{2}] , it does not depend on the choice of the surface S_{1} . By construction, the linking number is symmetric, bilinear and lies in \frac{1}{e}\mathbb{Z} . An equivalent construction is obtained as follows. The normal blow-up of L_{1} is the manifold M_{L_{1}} obtained by replacing every point p of L_{1} by the unit normal bundle to L_{1} , denoted by N^{1}_{L_{1}}(p) . The manifold M_{L_{1}} has a toric boundary N^{1}_{L_{1}} . Because M is a homology sphere, \mathrm{H}_{1}(M_{L_{1}};\mathbb{R}) is homeomorphic to \mathbb{R} . Denote by m_{1} a meridian of L_{1} , and by [m_{1}] its class in \mathrm{H}_{1}(M_{L_{1}};\mathbb{R}) . A multi-curve L_{2} has a homology class [L_{2}] which is a multiple of [m_{1}] . The linking number \mathrm{Lk}(L_{1},L_{2}) is then the unique rational number so that [L_{2}]=\mathrm{Lk}(L_{1},L_{2})[m_{1}]. 2.2. Linking number of invariant measures Assume now that M carries a flow (\varphi^{t})_{t\in\mathbb{R}} of class C^{1} . Given an ergodic \varphi^{t} -invariant probability measure \mu and a \mu -regular point x , one can pick a sequence (t_{n})\to+\infty so that \varphi^{t_{n}}(x) tends to x . Concatenating the orbit segment \varphi^{[0,t_{n}]}(x) with the shortest geodesic arc connecting its endpoints, one obtains a closed curve, denoted by k_{\varphi}(x,t_{n}) . Playing the same game with a second ergodic measure \nu and a \nu -regular point x^{\prime} not on the orbit of x , one would like to define the linking of \mu and \nu as \lim_{t_{n},t^{\prime}_{n}\to\infty}\frac{1}{t_{n}t^{\prime}_{n}}\mathrm{Lk}(k_{\varphi}(x,t_{n}),k_{\varphi}(x^{\prime},t^{\prime}_{n})) . There are difficulties in order to make the above definition rigorous. However the flows we consider in this note are integrable, and for this type of flows we shall later prove Lemma 2.1 (linking of ergodic invariant measures) . Let S be a sphere of revolution in \mathbb{R}^{3} endowed with the induced metric and let \varphi^{t} be its geodesic flow. Let \mu,\nu be two \varphi^{t} -invariant ergodic measures that do not charge the same periodic orbit, and let x,x^{\prime} be regular points for \mu and \nu respectively, then the limit \lim_{t_{n},t^{\prime}_{n}\to\infty}\frac{1}{t_{n}t^{\prime}_{n}}\mathrm{Lk}(k_{\varphi}(x,t_{n}),k_{\varphi}(x^{\prime},t^{\prime}_{n})) exists and does not depend on x,x^{\prime},t_{n},t^{\prime}_{n} . We then define the (dynamical) linking \mathrm{Lk}^{\varphi}(\mu,\nu) of the invariant measures as the above limit. Remark 2.2 . Lemma 2.1 is also valid for non integrable

Read the complete text