Let $f:M\to M$ be a partially hyperbolic diffeomorphism. That is, there exists a continuous splitting $TM=E^u\oplus E^c\oplus E^s$ into unstable, center and stable bundles. It is well known that there exist foliations $\mathcal{W}^u$ and $\mathcal{W}^s$ tangent to $E^u$ and $E^s$, respectively.

Let's assume $f$ is dynamically coherent. That is, there exist $\mathcal{W}^{cu}$ and $\mathcal{W}^{cs}$ tangent to $E^u\oplus E^c$ and $E^s\oplus E^c$, respectively, which are subfoliated by $\mathcal{W}^u$, $\mathcal{W}^s$ and $\mathcal{W}^{c}=\mathcal{W}^{cu}\cap \mathcal{W}^{cs}$.

Let's consider a center-stable manifold $W^{cs}(x)$.

Question: When does the following hold: $W^{cs}(x)=\bigcup_{y\in W^c(x)}W^s(y)$? Is it always true, or do we need some extra assumption for it?

Another choice is to ask for $W^{cs}(x)=\bigcup_{y\in W^s(x)}W^c(y)$. Just the previous one looks more natural.


No, it is in general not true. An example is the one recently constructed by Rodriguez Hertz-Rodriguez Hertz-Ures. In that example (non-dynamically coherent in dimension 3) the union of stable manifolds through a center manifold is strictly contained in the center-stable manifold (its boundary is a strong stable manifold). You can see a drawing in this slides of M.A.Rodriguez Hertz. Also, you can look at these notes I wrote (page 21) were this possibility is also discussed.

The question of wether such example can exist for a transitive partially hyperbolic diffeomorphism is open (see the paper by Bonatti-Wilkinson for more discussion).

If your question was more devoted to understanding if the union of stable manifolds is contained in a center-stable one, see the paper of Burago-Ivanov in JMD (Proposition 3.1). For results in higher dimensions, see the papers by Gogolev, Carrasco and Bonhet.

  • $\begingroup$ Their example is an interesting one (although not dynamically coherent, that's why they produce it). The mechanism is quite clear and answers my question. $\endgroup$
    – Pengfei
    Sep 10 '13 at 5:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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