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

Let $X$ be a topological manifold of dimension $n$, equipped with a compatible CAT(0) metric. Are sufficiently small metric spheres in $X$ homeomorphic to metric spheres in Euclidean space $\mathbb{E}^n$?

[In "Ideal boundary of CAT(0) spaces" (1998) by Myung-Jin Jeon, this was unclear to the author; see the bottom of Page 104. That paper examined geodesic completeness for CAT(0) manifolds.]

EDIT: It is well-known that sufficiently small metric balls in any CAT(0) space are contractible. Using the manifold property, I believe that my question reduces to a very basic one:

If $U$ is a contractible open subset of $\mathbb{R}^n$, then is $U$ homeomorphic to $\mathbb{R}^n$?

share|cite|improve this question
The answer to the question in the edit is no. The Whitehead manifold embeds in $S^3$. See – Richard Kent Aug 21 '11 at 3:24
The question in the edit has a negative answer in all dimensions $>2$; read In dimensions $>3$ the point is that there are many compact contractible manifolds, and the double of a compact contractible manifold along the boundary is a homotopy sphere, hence it is homeomorphic to $S^n$, and removing a point gives $\mathbb R^n$. – Igor Belegradek Aug 21 '11 at 13:35
up vote 15 down vote accepted

The answer is no. By the double suspension theorem of Cannon and Edwards, if $X^n$ is a homology $n$-sphere, then the double suspension $S^2X$ is a sphere. In particular, the cone $CSX$ on $SX$ will be a simply connected manifold homeomorphic to $R^{n+2}$, since it is obtained from the double suspension by deleting a point. If there is a simplicial complex structure on $SX$ which is flag (e.g. by barycentric subdivision), then the cone $CSX$ will have a CAT(0) metric. The point is that one may make each of the simplices of the triangulation of $SX$ into a right-angled $n+1$ spherical simplex, then take the metric cone on this to get a quadrant of $R^{n+2}$, and glue these quadrants together in the manner prescribed by the simplicial complex. The metric ball about the cone point of $CSX$ will not be homeomorphic to a ball, since its boundary will be $SX$ which is not a manifold. Check out the papers of Davis-Januskiewicz for constructions of CAT(0) complexes.

share|cite|improve this answer
Okay, sorry, I got confused by the different uses of $X$... for my original question, a counter-example is $X := CS\Sigma$ where $\Sigma$ is a closed $n$-manifold with the homology type, but not the homotopy type, of a sphere. By Cannon--Edwards, $X \approx \mathbb{R}^{n+2}$, and by Davis--Januskiewicz, $X$ is CAT(0). Thanks again! – Qayum Khan Sep 5 '11 at 1:40

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.