← All problems

Berger conjecture on closed geodesics

Let (M,g)(M,g) be a simply connected Riemannian manifold. If every geodesic of (M,g)(M,g) is closed, then all prime closed geodesics of (M,g)(M,g) have the same length; equivalently, there exists L>0L>0 such that every prime closed geodesic has length LL.

References

Progress summary

Refreshed
Open · progress reported

The conjecture is proved for spheres in dimensions two and at least four, while the full statement for simply connected manifolds remains open; a new unrefereed hypersurface result does not change that.

Berger's conjecture says that on a simply connected Riemannian manifold whose geodesics are all closed, all prime geodesics have the same length. The general manifold case remains unresolved.

Known results

  • The conjecture holds on the 22-sphere (Gromoll–Grove).
  • It holds on every topological sphere of dimension n≥4n \ge 4 (Radeschi–Wilking; later free-loop-space work, 2017).
  • In dimension 33, Olsen proved structural results for the geodesic-flow isotropy set, but not the full conjecture.
  • Except possibly in dimension 33, closed-geodesic metrics on spheres are Zoll (Grove–Gromoll and Radeschi–Wilking).

October 2026 hypersurface result

Christian Lange's paper claims rigidity, regularity, and embeddedness results for Besse families of minimal hypersurfaces that parallel Berger-type conclusions. This is progress for a related hypersurface problem, not a solution of Berger's conjecture for manifolds.

Current status (as of October 2026): The conjecture is settled for spheres in dimensions 22 and n≥4n \ge 4, but remains open for general simply connected Besse manifolds, with the latest hypersurface result unverified and non-resolving.

Sources

Solutions 0

No solutions have been posted yet.