### March 20, 2005

Research supported in part by National Science Foundation.
<ph f="cmbx">Hyperbolic Invariant Sets With Positive Measures</ph>

### Zhihong Xia

Department of Mathematics, Northwestern University, Evanston, Illinois 60208 E-mail address : xia@math.northwestern.edu
• Abstract. In this note we prove some results concerning volume-preserving Anosov diffeomorphisms on compact manifolds. The main theorem is that if a ${C}^{1+\alpha }$  , $\alpha >0$  , volume-preserving diffeomorphism on a compact manifold has a hyperbolic invariant set with positive volume, then the map is Anosov. This is not necessarily true for ${C}^{1}$  maps. The proof uses a special type of measure density points different from the standard Lebesgue density points.

1 Introduction and statement of main results

We consider volume-preserving or symplectic diffeomorphisms on a compact Riemannian manifold $M$  . Let $\text{Diff}r\mu \left(M\right)\text{}$  be the the set of all ${C}^{r}$  diffeomorphisms preserving a smooth volume $\mu$  on $M$  . If $r$  is not an integer, $r=k+\alpha$  for some positive integer $k$  and $0<\alpha <1$  , it is understood that the functions in $\text{Diff}r\mu \left(M\right)\text{}$  are ${C}^{k}$  functions with $\alpha$  -Hölder $k$  -th derivatives.
An invariant set $\Lambda \subset M$  is said to be hyperbolic if there is a continuous splitting of ${T}_{x}M={E}_{x}^{s}\oplus {E}_{x}^{u}$  for every $x\in \Lambda$  and constants $C>0$  , $\lambda >1$  such that
 $\begin{array}{ccc}d{f}_{x}\left({E}_{x}^{s}\right)& =& {E}_{f\left(x\right)}^{s}\text{and}d{f}_{x}\left({E}_{x}^{u}\right)={E}_{f\left(x\right)}^{u}\end{array}$
 $\begin{array}{ccc}|d{f}_{x}^{n}{v}_{x}^{s}|& \le & C{\lambda }^{-n}|{v}^{s}|,\text{for all}{v}^{s}\in {E}_{x}^{s},n\in \mathbb{N}\end{array}$
 $\begin{array}{ccc}|d{f}_{x}^{-n}{v}_{x}^{u}|& \le & C{\lambda }^{-n}|{v}^{u}|,\text{for all}{v}^{u}\in {E}_{x}^{u},n\in \mathbb{N}\end{array}$
If the whole manifold $M$  is hyperbolic for some $f\in \text{Diff}r\mu \left(M\right)\text{}$  , then $f$  is said to be Anosov. Not all manifolds can support Anosov diffeomorphisms.
Typical examples of hyperbolic invariant sets are Cantor sets. The following simple proposition explains why this is the case.
Proposition 1.1. Let $f\in \text{Diff}r\mu \left(M\right)\text{}$  , $r\ge 1$  , be a volume-preserving diffeomorphism on a compact manifold $M$  . Let $\Lambda \subset M$  be a closed hyperbolic invariant set. If the interior of $\Lambda$  is non-empty, then $f$  is Anosov on $M$  and $\Lambda =M$  .
We will give a proof of this proposition in the following section. The proof uses the fact that the recurrent points are dense on the manifold. This is a consequence of the volume-preserving property. Without the volume-preserving or the dense recurrent points condition, the proposition is not true, we refer to Fisher [3for a counter-example. Fisher also give a proof of the above proposition. On the other hand, it is an open problem whether there are any Anosov diffeomporphisms with wandering domains.
One natural question one asks is whether there is any hyperbolic invariant set with a positive measure for a volume-preserving non-Anosov diffeomorphism. The answer is yes for ${C}^{1}$  diffeomorphisms, as Bowen's example of fat horseshoe shows [2, see also Robinson & Young [7. However, if the map is assumed to be ${C}^{1+\alpha }$  for some $\alpha >0$  , then the answer is no. This is the main result of this note.
Theorem 1.2. Let $f\in \text{Diff}r\mu \left(M\right)\text{}$  , $r>1$  , be a volume preserving diffeomorphism on a compact manifold $M$  . Let $\Lambda \subset M$  be a closed hyperbolic invariant set. If $\mu \left(\Lambda \right)>0$  , then $f$  is Anosov on $M$  and $\Lambda =M$  .
It is not surprising that the map is required to be ${C}^{1+\alpha }$  . As various examples show, measure-theoretical properties are often not respected by ${C}^{1}$  maps. Additional smoothness, however little, guarantees certain regularities in measure.
Our proof uses a special type of measure density points different from the Lebesgue density points. The density basis for our density points are dynamically defined. It is similar to the Juliennes defined by Pugh & Shub [5 [6. But our case is much simpler.

2 Proof of the proposition

In this section, we give a simple proof of the Proposition  1.1 .
Let $U$  be the interior of the hyperbolic invariant set $\Lambda$  . By the assumption of the proposition, $U\ne \varnothing$  . Clearly, $U$  is invariant. We want to prove that the closure of $U$  , $\overline{U}$  is the whole manifold. We know that $\overline{U}$  is closed, it suffices to show that $\overline{U}$  is also open.
For any $x\in \overline{U}$  , there exists a sequence of points ${x}_{n}\in U$  , $i\in \mathbb{N}$  , such that ${x}_{n}\to x$  as $n\to \infty$  . As $f$  is volume preserving, by Poincaré recurrence theorem, almost every point is both forward and backward recurrent. Moreover, the set of periodic points is dense in $U$  , since $U$  is hyperbolic. We may choose $\left\{{x}_{n}{\right\}}_{n\in \mathbb{N}}$  to be periodic points. Since $U$  is invariant and each ${x}_{n}$  is an interior point in $U$  , then ${W}^{s}\left({x}_{n}\right)$  and ${W}^{u}\left({x}_{n}\right)$  are in $U$  for all $n\in \mathbb{N}$  . For each fixed $\delta >0$  small, let ${W}_{\delta }^{s}\left(x\right)$  and ${W}_{\delta }^{u}\left(x\right)$  be, respectively, the local stable manifold and unstable manifold of $x$  .
As ${x}_{n}\to x$  , as $n\to \infty$  , we have that ${W}_{\delta }^{s}\left({x}_{n}\right)\to {W}_{\delta }^{s}\left(x\right)$  and ${W}_{\delta }^{u}\left({x}_{n}\right)\to {W}_{\delta }^{u}\left(x\right)$  as $n\to \infty$  . This implies that each point on ${W}_{\delta }^{s}\left(x\right)$  or on ${W}_{\delta }^{u}\left(x\right)$  is also in the closure of $U$  . Let $y\in {W}_{\delta }^{s}\left(x\right)$  and $z\in {W}_{\delta }^{u}\left(x\right)$  , the same argument shows that ${W}_{\delta }^{u}\left(y\right)$  and ${W}_{\delta }^{s}\left(z\right)$  are both in the closure of $U$  . Consequently, ${W}_{\delta }^{u}\left(y\right)\cap {W}_{\delta }^{s}\left(z\right)\in \overline{U},$  i.e., $\overline{U}$  has the product structure. This implies that $x$  is in the interior of $\overline{U}$  .
Consequently, the set $\overline{U}$  is open. Since $\overline{U}$  is also closed and $M$  is connected, we have $\overline{U}=M$  . i.e., $f$  is hyperbolic on $M$  .
This proves the proposition.

3 Proof of the Theorem

The proof of Theorem  1.2 uses a similar idea to the proof of Proposition  1.1 , but the details are much more complicated. Here the interior points are replaced by density points. One may regard the density points as measure theoretical interior points for a set with positive measure.
We need some preliminary results from standard smooth ergodic theory. It is well-known that the stable and unstable foliations for a ${C}^{1}$  Anosov diffeomorphism may not be absolutely continuous. However, for ${C}^{1+\alpha }$  diffeomorphisms, these foliations are absolutely continuous (Anosov [1). Moreover, the stable and unstable foliations over a hyperbolic (even non-uniformly, cf Pesin [4) invariant set are also absolutely continuous for ${C}^{1+\alpha }$  diffeomorphisms. In fact, the absolute continuity of the foliations is proved by showing that the holonomy maps of these foliations are absolutely continuous.
We also need some results on density basis and density points of a measurable set. Let $A\subset {\mathbb{R}}^{n}$  be a measurable set with the standard Lebesgue measure $m$  . A point $x\in {\mathbb{R}}^{n}$  is said to be a Lebesgue density point if ${lim}_{\epsilon \to 0}\frac{m\left(B\left(x,\epsilon \right)\cap A\right)}{m\left(B\left(x,\epsilon \right)\right)}=1,$  where $B\left(x,\epsilon \right)$  is the $\epsilon$  -ball centered at $x$  . Lebesgue density theorem states that almost every point of $A$  is a density point for $A$  .
To rpove our theorem, we need a different definition of density point. The Lebesgue density point is defined by a basis of $\epsilon$  -balls. We replace it by a dynamically defined basis. Let $\Lambda \subset M$  be a hyperbolic invariant set, we first define a basis on the unstable manifold for each point $\Lambda$  .
For a fixed small real number $\delta >0$  , let ${W}_{\delta }^{u}\left(x\right)$  be the local unstable manifold of a point $x\in \Lambda$  . Let ${\mu }_{u}$  and ${\mu }_{s}$  respectively be the induced measures of the smooth volume form $\mu$  on the unstable leaves. Let ${n}_{u}$  and ${n}_{s}$  respectively be the dimensions be the unstable and stable leaves. For any positive integer $k$  , let ${B}_{k}^{u}\left(x\right)$  be a subset of ${W}^{u}\left(x\right)$  defined by ${B}_{k}^{u}\left(x\right)={f}^{-k}\left({W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)\right).$  Clearly, the cubes ${B}_{k}^{u}\left(x\right)$  , $k\in \mathbb{N}$  shrinks to the point $x$  as $k\to \infty$  . We call the collection of the sets $\left\{{B}_{k}^{u}\left(x\right)|k\in \mathbb{N},x\in \Lambda \right\}$  the unstable density basis.
Similarly, we can define the stable density basis $\left\{{B}_{k}^{s}\left(x\right)|k\in \mathbb{N},x\in \Lambda \right\}$  , by defining ${B}_{k}^{s}\left(x\right)={f}^{k}\left({W}_{\delta }^{u}\left({f}^{-k}\left(x\right)\right)\right).$  The density basis we defined has infinite eccentricity.
A point $x\in \Lambda$  is said to be a density point on the unstable foliation if ${lim}_{k\to \infty }\frac{{\mu }_{u}\left({B}_{k}^{u}\left(x\right)\cap A\right)}{{\mu }_{u}\left({B}_{k}^{u}\left(x\right)\right)}=1.$  Similarly, we can define the density points on the stable foliation.
Proposition 3.1. The set of points on $\Lambda$  that are both density points on the stable foliations and unstable foliations has the full measure in $\Lambda$  .
Our definitions of density points can be regarded as simplified versions of the Juliennes density points by Pugh & Shub [5 [6. The above proposition follows from the proof for the Jeliennes density point by Pugh & Shub. The proof itself is similar to the proof of the Lebesgue Density Theorem. The key properties for the density basis are scaling and engulfing defined as follows.
(a) Scaling: for any fixed $k\ge 0$  , $m\left({B}_{n}^{u}\left(x\right)\right)/m\left({B}_{n+k}^{u}\left(x\right)\right)$  is unformly bounded as $n\to \infty$  .
(b) Engulfing: there is a unifom $L$  such that ${B}_{n+L}^{u}\left(x\right)\cap {B}_{n+L}^{u}\left(y\right)\ne \varnothing ⇒{B}_{n+L}^{u}\left(x\right)\cup {B}_{n+L}^{u}\left(y\right)\subset {B}_{n}^{u}\left(x\right).$  We remark that our density points are defined on the stable and unstable foliations, we freely used the fact that the stable and unstable foliations are absolutely continuous.
Let $A$  be a subset of $\Lambda$  such that for any $x\in A$  , $x$  is a density point $\Lambda$  on both stable foliation and unstable foliation; and $x$  is a recurrent point, both forward and backward. By Poincaré recurrence theorem, Proposition  3.1 and the absolute continuity of the foliations, the set $A$  has the full measure in $\Lambda$  .
The following is the main lemma in proving our theorem.
Lemma 3.2. Assume that $f$  is a ${C}^{1+\alpha }$  volume-preserving diffeomorphism, for some positive number $\alpha >0$  . Fix $x\in A$  and a positive number $\delta >0$  .
Then for any $\epsilon >0$  , there exists a positive integer ${k}_{0}$  , depending on $x$  and $\epsilon$  , such that for $k\ge {k}_{0}$  , ${\mu }_{s}\left({W}_{\delta }^{s}\left({f}^{-k}\left(x\right)\right)\cap A\right)\ge \left(1-\epsilon \right){\mu }_{s}\left({W}_{\delta }^{s}\left({f}^{-k}\left(x\right)\right)\right)$  and ${\mu }_{u}\left({W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)\cap A\right)\ge \left(1-\epsilon \right){\mu }_{u}\left({W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)\right)$  i.e., for sufficiently large $k$  , the set $A$  has a very high density in ${W}_{\delta }^{s}\left({f}^{-k}\left(x\right)\right)$  and ${W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)$  .
Proof of the lemma: We first prove the lemma for the unstable foliation. Let ${n}_{u}$  be the dimension of the leaves of the foliation, local unstable manifold ${W}_{\delta }^{u}\left(x\right)$  can be identified with a cube in ${E}_{x}^{u}={\mathbb{R}}^{{n}_{u}}$  by the exponential map from ${E}_{x}^{u}$  to ${W}^{u}\left(x\right)$  .
Since the leaves of the unstable foliation is smooth, the conditional measure ${\mu }_{u}$  are smoothly equivalent to the standard Lebesgue measure $m$  on ${\mathbb{R}}^{{n}_{u}}$  . i.e., for any point $x\in \Lambda$  , there is a smooth function ${g}_{u}\left(y\right)$  defined for $y\in {W}_{\delta }^{u}\left(x\right)$  on the local unstable manifold, uniformly bounded away from zero and infinity, such that ${\mu }_{u}\left(E\right)={\int }_{E}{g}_{u}dm,$  where $E$  is a measurable set in ${W}_{\delta }^{u}\left(x\right)$  and $m$  is the standard Lebesgue measure in ${\mathbb{R}}^{{n}_{u}}$  .
For any positive integer $k$  , we want to estimate the measure of the set ${W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)\cap A$  . Let ${B}_{0}^{k}={f}^{-k}\left({W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)\right)$  , obviously ${B}_{0}^{k}\subset {W}_{\delta }^{u}\left(x\right)$  . In fact, ${B}_{0}^{k}$  is the set ${B}_{k}^{u}\left(x\right)$  in our definition of density basis. We iterate ${B}_{0}^{k}$  under $f$  and obtain a sequence of sets ${B}_{i}^{k}={f}^{i}\left({B}_{0}^{k}\right)$  , for $i=1,2,\dots ,k$  . The last set in the sequence is ${B}_{k}^{k}={W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)$  .
Let ${\eta }_{k}=1-{\mu }_{u}\left({B}_{0}^{k}\cap A\right)/{\mu }_{u}\left({B}_{0}^{k}\right)$  , then $0\le {\eta }_{k}\le 1$  . As $x$  is a density point on the unstable foliation, ${lim}_{k\to \infty }\frac{{\mu }_{u}\left({B}_{0}^{k}\cap A\right)}{{\mu }_{u}\left({B}_{0}^{k}\right)}=1,$  The number ${\eta }_{k}$  is small for large $k$  , and ${lim}_{k\to \infty }{\eta }_{k}=0$  . Since $f$  is ${C}^{1+\alpha }$  , there exists a constant ${C}_{1}>0$  such that $||d{f}_{y}-d{f}_{z}||\le {C}_{1}|y-z{|}^{\alpha }$  . Here we abuse the notation a little by writing $|y-z|$  as the distance between $y$  and $z$  .
Let ${\rho }_{i}^{k}$  be the maximum distance from ${f}^{i}\left(x\right)$  to the boundary of ${B}_{i}^{k}$  , i.e., ${\rho }_{i}^{k}={max}_{y\in {B}_{i}^{k}}\left\{d\left({f}^{i}\left(x\right),y\right)\right\}.$  For any $y\in {B}_{0}^{k}$  , $||d{f}_{y}-d{f}_{x}||\le {C}_{1}\left({\rho }_{0}^{k}{\right)}^{\alpha }$  . In general, for any $y\in {B}_{i}^{k}$  , $||d{f}_{y}-d{f}_{{f}^{i}\left(x\right)}||\le {C}_{1}\left({\rho }_{i}^{k}{\right)}^{\alpha }$  .
Let ${J}_{u}\left(y\right)=|det\left(d{f}_{y}{|}_{{E}_{y}^{u}}\right)|$  be the Jacobian of the map $f$  at $y$  restricted on the unstable manifold of $y$  . Then $|{J}_{u}\left(x\right)-{J}_{u}\left(y\right)|\le {C}_{2}|x-y{|}^{\alpha }$  , for some positive constant ${C}_{2}>0$  .
Let ${D}_{0}={B}_{0}^{k}\A$  and ${D}_{i}={B}_{i}^{k}\A$  , for $i=1,2,\dots ,k$  . These are the complements of $A$  in ${B}_{i}^{k}$  . By the definition of ${\eta }_{k}$  , ${\mu }_{u}\left({D}_{0}\right)={\eta }_{k}{\mu }_{u}\left({B}_{0}^{k}\right)$  . We need to estimate the measure of ${D}_{1}$  . For any set $E\subset {B}_{0}^{k}$  , ${\mu }_{u}\left(f\left(E\right)\right)={\int }_{f\left(E\right)}{g}_{u}dm={\int }_{E}{J}_{u}\left({g}_{u}\cdot f\right)dm={\int }_{E}{J}_{u}\left({g}_{u}\cdot f\right){g}_{u}^{-1}d{\mu }_{u}$  Since the functions ${g}_{u}$  and ${g}_{u}^{-1}$  are smooth on any unstable manifold, the integrand in the above integral is ${C}^{1+\alpha }$  , there is a constant ${C}_{3}>0$  such that $|{J}_{u}\left(y\right)\left({g}_{u}\cdot f\right)\left(y\right){g}_{u}^{-1}\left(y\right)-{J}_{u}\left(x\right)\left({g}_{u}\cdot f\right)\left(x\right){g}_{u}^{-1}\left(x\right)|\le {C}_{3}|x-y{|}^{\alpha },$  for all $x\in \Lambda$  , $y\in {W}_{\delta }^{u}\left(x\right)$  . Therefore, $|{\mu }_{u}\left(f\left(E\right)\right)-\left({J}_{u}\left(x\right)\left({g}_{u}\cdot f\right)\left(x\right){g}_{u}^{-1}\left(x\right)\right){\mu }_{u}\left(E\right)|\le {C}_{3}\left({\rho }_{0}^{k}{\right)}^{\alpha }{\mu }_{u}\left(E\right).$  Consequently, ${\mu }_{u}\left({D}_{1}\right)\le \left({J}_{u}\left(x\right)\left({g}_{u}\cdot f\right)\left(x\right){g}_{u}^{-1}\left(x\right)\right){\mu }_{u}\left({D}_{0}\right)+{C}_{3}\left({\rho }_{0}^{k}{\right)}^{\alpha }{\mu }_{u}\left({D}_{0}\right)$  and ${\mu }_{u}\left({B}_{1}^{k}\{D}_{1}\right)\ge \left({J}_{u}\left(x\right)\left({g}_{u}\cdot f\right)\left(x\right){g}_{u}^{-1}\left(x\right)\right){\mu }_{u}\left({B}_{0}^{k}\{D}_{0}\right)-{C}_{3}\left({\rho }_{0}^{k}{\right)}^{\alpha }{\mu }_{u}\left({B}_{0}^{k}\{D}_{0}\right)$  and therefore ${\mu }_{u}\left({D}_{1}\right)\le {\eta }_{k}\frac{\left(1+{C}_{3}\left({\rho }_{0}^{k}{\right)}^{\alpha }\right)}{\left(1-{C}_{3}\left({\rho }_{0}^{k}{\right)}^{\alpha }\right)}{\mu }_{u}\left({B}_{1}^{k}\right).$  By induction on $i$  , we have ${\mu }_{u}\left({D}_{k}\right)\le {\eta }_{k}\frac{\left(1+{C}_{3}\left({\rho }_{0}^{k}{\right)}^{\alpha }\right)}{\left(1-{C}_{3}\left({\rho }_{0}^{k}{\right)}^{\alpha }\right)}\frac{\left(1+{C}_{3}\left({\rho }_{1}^{k}{\right)}^{\alpha }\right)}{\left(1-{C}_{3}\left({\rho }_{1}^{k}{\right)}^{\alpha }\right)}\cdots \frac{\left(1+{C}_{3}\left({\rho }_{k-1}^{k}{\right)}^{\alpha }\right)}{\left(1-{C}_{3}\left({\rho }_{k-1}^{k}{\right)}^{\alpha }\right)}{\mu }_{u}\left({B}_{k}^{k}\right).$  The map $df:{T}_{\Lambda }M\to {T}_{\Lambda }M$  uniformly expands vectors on the unstable splitting.
That uniform expansion extends to local unstable manifolds ${W}_{\delta }^{u}\left(x\right)$  , $x\in \Lambda$  if $\delta$  is chosen small enough. There exist positive real numbers ${C}_{4}>C>0$  and $\lambda >{\lambda }_{1}>1$  (here $C$  and $\lambda$  are the same as those in the definition of the hyperbolic invariant set) such that ${\rho }_{k-1}^{k}\le {C}_{4}{\lambda }_{1}^{-1}\delta$  and ${\rho }_{i}^{k}\le {C}_{4}{\lambda }_{1}^{-\left(k-i\right)}\delta$  , for $i=0,1,\dots ,k-1$  . This implies that
 $\begin{array}{ccc}{\mu }_{u}\left({D}_{k}\right)& \le & {\eta }_{k}\frac{\left(1+{C}_{3}\left({C}_{4}{\lambda }_{1}^{-k}\delta {\right)}^{\alpha }\right)}{\left(1-{C}_{3}\left({C}_{4}{\lambda }_{1}^{-k}\delta {\right)}^{\alpha }\right)}\frac{\left(1+{C}_{3}\left({C}_{4}{\lambda }_{1}^{-k+1}\delta {\right)}^{\alpha }\right)}{\left(1-{C}_{3}\left({C}_{4}{\lambda }_{1}^{-k+1}\delta {\right)}^{\alpha }\right)}\end{array}$
 $\begin{array}{ccc}& & \cdot \cdot \cdot \frac{\left(1+{C}_{3}\left({C}_{4}{\lambda }_{1}^{-1}\delta {\right)}^{\alpha }\right)}{\left(1-{C}_{3}\left({C}_{4}{\lambda }_{1}^{-1}\delta {\right)}^{\alpha }\right)}{\mu }_{u}\left({B}_{k}^{k}\right)\end{array}$
 $\begin{array}{ccc}& <& {\eta }_{k}{\mu }_{u}\left({B}_{k}^{k}{\right)}^{\infty }{\prod }_{i=1}\left(1+{C}_{3}\left({C}_{4}{\lambda }_{1}^{-i}\delta {\right)}^{\alpha }\right){/}^{\infty }{\prod }_{i=1}\left(1-{C}_{3}\left({C}_{4}{\lambda }_{1}^{-i}\delta {\right)}^{\alpha }\right)\end{array}$
Since $\lambda >1$  , then ${\lambda }^{\alpha }>1$  , the infinite products converge. We have ${\mu }_{u}\left({D}_{k}\right)<{C}_{5}{\eta }_{k}{\mu }_{u}\left({B}_{k}^{k}\right).$  for some constant ${C}_{5}>0$  .
Choose a positive integer ${k}_{0}$  such that for $k\ge {k}_{0}$  , ${\eta }_{k}<\epsilon /{C}_{5}$  , then we have ${\mu }_{u}\left({W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)\cap A\right)\ge \left(1-\epsilon \right){\mu }_{u}\left({W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)\right),$  for all $k\ge {k}_{0}$  .
This proves the statement of the lemma on the unstable foliation. The part on the stable foliation can be proved in the same way by considering ${f}^{-1}$  .
This proves the lemma.
We return to the proof of the theorem. Let $E$  be the closure of $A$  in $M$  . We claim that if $y\in E$  , then ${W}_{\delta }^{s}\left(y\right)\subset E$  and ${W}_{\delta }^{u}\left(y\right)\subset E$  .
Suppose that this is not true. i.e., there is a point $z\in {W}^{u}\left(y\right)$  and a small ${\epsilon }_{1}$  -ball around $z$  , $B\left(z,{\epsilon }_{1}\right)\subset M$  such that $B\left(z,\epsilon \right)\cap A=\varnothing$  . Consequently, there are constants ${\epsilon }_{2}>0$  , depending on ${\epsilon }_{1}$  , and ${C}_{6}>0$  , independent of ${\epsilon }_{1}$  , such that if $x\in \Lambda$  , $|x-y|\le {\epsilon }_{2}$  , then ${\mu }_{u}\left({W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)\cap A\right)<\left(1-{C}_{6}{\epsilon }_{1}^{{n}_{u}}\right){\mu }_{u}\left({W}_{\delta }^{u}\left({f}^{k}\left(x\right)\right)\right).$  On the other hand, since $y\in E$  , there is a sequence of points ${x}_{i}\in A$  , $i\in \mathbb{N}$  such that ${x}_{i}\to y$  and $i\to \infty$  . For the above ${\epsilon }_{1}$  , there is positive integer ${i}_{0}$  such that if $i\ge {i}_{0}$  , $d\left({x}_{i},y\right)\le {\epsilon }_{1}/3$  . For any fixed $\epsilon >0$  , by the lemma above and the recurrence of ${x}_{i}$  , there is a positive integer $k>0$  such that $d\left({x}_{i},{f}^{k}\left({x}_{i}\right)\right)\le {\epsilon }_{1}/3$  and ${\mu }_{u}\left({W}_{\delta }^{u}\left({f}^{k}\left({x}_{i}\right)\right)\cap A\right)\ge \left(1-\epsilon \right){\mu }_{u}\left({W}_{\delta }^{u}\left({f}^{k}\left({x}_{i}\right)\right)\right).$  Since $d\left(y,{f}^{k}\left({x}_{i}\right)\right)\le \frac{2{\epsilon }_{2}}{3}<{\epsilon }_{2}$  , choosing $\epsilon ={C}_{6}{\epsilon }_{1}^{{n}_{u}}$  leads to a contradiction. This contradiction show that if $y\in E$  , then ${W}_{\delta }^{u}\left(y\right)\subset E$  . Similarly by considering ${f}^{-1}$  , we have ${W}_{\delta }^{s}\left(y\right)\subset E$  . To conclude our proof, for any $y\in E$  , ${W}_{\delta }^{u}\left(y\right)\subset E$  and $V={\bigcup }_{z\in {W}_{\delta }^{u}\left(y\right)}{W}_{\delta }^{s}\left(z\right)\subset E.$  Since $V$  is hyperbolic, $y$  is in the interior of $V$  . This implies that $E$  is an open set.
But $V$  is also closed and non-empty and $M$  is connected, $E=M$  . This implies that $f$  is Anosov. Finally, the reason that $\Lambda =M$  in the first place is that $f$  is ergodic, any invariant set with positive measure must have full measure and its closure must be the whole manifold.
This proves the theorem.
References

1. D.V. Anosov. Geodesic flows on closed riemannian manifolds of negative curvature. Proc. Steklov. Inst. Math., 90, 1967.
2. R. Bowen. A horseshoe with positive measure. Invent. Math., 29:203–204, 1975.
3. T. Fisher. On the structure of hyperbolic sets. Thesis, Northwestern University, 2004.
4. Ya. Pesin. Characteristic lyapunov exponents and smooth ergodic theory. Russian Math. Surveys, 32(4):55–112, 1977.
5. C. Pugh and M. Shub. Stable ergodicity and julienne quasiconformality. J. Eur. Math. Soc., 2:1–52, 2000.
6. C. Pugh and M. Shub. Stable ergodicity. Bulletin Amer. Math. Soc., 41(1):1–41, 2003.
7. C. Robinson and Lai-Sang Young. Nonabsolutely continuous foliations for an anosov diffeomorphisms. Inventiones Math., 61:159–176, 1980.

Department of Mathematics, Northwestern University, Evanston, Illinois 60208 E-mail address : xia@math.northwestern.edu