MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am thinking about the following questions about the cutlocus of a point in a Riemannian manifold or of a hypersurface in the Euclidean space:

1) If all the points of the (nonvoid) cutlocus of a point $p$ in a Riemannian manifold are also conjugate points, then the manifold is rotationally symmetric around the point $p$ and the cutlocus is a single point?

2) If all the points of the (nonvoid) cutlocus of a smooth, connected hypersurface embedded in ${\mathbb R}^{n+1}$ are also conjugate points, what such a hypersurface looks like?
One guess is that it should be a level set of the distance function from the cutlocus itself. In the case $n=1$, that is, curves in the plane, it is a circle.

It can be seen that in this case, by Sard theorem, the cutlocus has Hausdorff $H^n$-measure zero.

share|cite|improve this question
up vote 4 down vote accepted

It is known that for a simply-connected compact Riemannian symmetric space, the cut locus of a point coincides with its first conjugate locus, see e.g. the book by Cheeger and Ebin, "Comparison theorems in Riemannian geometry", Theorem 5.13. So the answer to (1) is no.
Examples of compact symmetric spaces include spheres, projective spaces (real, complex, quaternionic, octonionic), Grassmannian manifolds, Lie groups with bi-invariant metrics etc. So you see that this class of Riemannian manifolds is not small.
Further, the cut locus of a point in a symmetric space can be explictly described, and it is in general not a point. In fact, it is a highly non-trivial result, proved by Berger (Green in the $2$-dimensional case), that a manifold with the property that the cut locus of any point is reduced to a single point ("wiedersehens manifold") has to be isometric to a round sphere.

share|cite|improve this answer
Thanks, nice answer, I was really too optimistic... possibly, also $S^2\times S^2$ is a counterexample. Anyway, do you know a reference where I can find the description of the possible cut loci of the symmetric spaces? They should have dimension less than $n-1$ ($n$ is the dimension of the manifold). I am now asking myself if the property $Cut_p=Conj_p$ holds for $every$ point $p$ of the manifold, what can be said about this latter? – Carlo Mantegazza Jan 24 '13 at 23:27
I think this is covered in Sakai's book on Riemannian geometry, but I don have a copy with me now. In fact, the description of the conjugate loci in symmetric spaces is very easy using the system of restricted roots. I think your final question is open and very hard, see section 6 of [this paper] ( – Claudio Gorodski Jan 25 '13 at 13:08
Very interesting reference, nice paper, thanks! – Carlo Mantegazza Jan 25 '13 at 13:26

For question 1, the answer is NO.

Take any rotationally symmetric manifold with fixed point $p$. Perturb it smoothly it in the directions tangent to the spheres centered at $p$. You still get one point as the cutlocus of $p$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.