5.2 (Agol) — Large injectivity radius in hyperbolic homology manifolds

Do there exist fibered hyperbolic 33-manifolds that are homology S2×S1S^2\times S^1 and have arbitrarily large injectivity radius? Are there hyperbolic homology spheres of arbitrarily large injectivity radius?

References

Progress summary

Refreshed
Open

The scan found no solution: neither the fibered case nor the homology-sphere case has been settled.

Agol’s problem asks whether there are fibered hyperbolic 33-manifolds with homology S2×S1S^2\times S^1 and arbitrarily large injectivity radius, and whether hyperbolic homology 33-spheres can have the same property.

Known results

  • A 2013 paper constructs integer homology 33-spheres whose injectivity radius exceeds any fixed RR at all but an arbitrarily small proportion of points, but not everywhere.
  • The same paper constructs closed hyperbolic mapping tori with inj⁡(M)>R\operatorname{inj}(M)>R for every RR, without establishing the homology S2×S1S^2\times S^1 condition.
  • A 2005 result gives towers of hyperbolic rational homology 33-spheres with injectivity radius tending to infinity, conditional on the Generalized Riemann Hypothesis and an additional GL2\mathrm{GL}_2 conjecture.

2022 confirmation of openness

A 2022 source explicitly states that homology spheres with arbitrarily large injectivity radius remain unknown. The supplied scan found no subsequent proof, counterexample, verification, or claimed solution for either part.

Current status (as of September 2026): Both questions remain open; known constructions give only weakened density results, conditional rational-homology results, or mapping tori without the required homology condition.

Sources

Solutions 0

No solutions have been posted yet.