

===== SOURCE: bw-journal.txt =====


=== PDF PAGE 1 OF 42 ===
 ANNALS OF
MATHEMATICS





On the ergodicity of partially hyperbolic
               systems

            By Keith Burns and Amie Wilkinson





        SECOND SERIES, VOL. 171, NO. 1

                    January, 2010

               anmaah


=== PDF PAGE 2 OF 42 ===



=== PDF PAGE 3 OF 42 ===
                                   Annals of Mathematics, 171 (2010), 451–489




On the ergodicity of partially hyperbolic systems

             By KEITH BURNS and AMIE WILKINSON


                               Abstract

   Pugh and Shub have conjectured that essential accessibility implies ergodicity
for a C 2, partially hyperbolic, volume-preserving diffeomorphism. We prove this
conjecture under a mild center bunching assumption, which is satisﬁed in particular
by all partially hyperbolic systems with 1-dimensional center bundle. We also
                         1Cıobtain ergodicity results for C     partially hyperbolic systems.


                              Introduction

     In [Hop39], Eberhard Hopf introduced a simple argument that proved the
ergodicity (with respect to Liouville measure) of the geodesic ﬂow of a compact,
negatively curved surface. The argument has since been applied to increasingly
general classes of dynamical systems. The key feature that these systems possess
is hyperbolicity. The strongest form of hyperbolicity is uniform hyperbolicity. A
diffeomorphism f  W M ! M of a compact manifold M is uniformly hyperbolic or
Anosov if there exists a splitting of the tangent bundle into Tf -invariant subbundles,
given by
                TM D Es ˚ Eu;
and a continuous Riemannian metric such that, for every unit vector v 2 TM,
                      kTf vk < 1   if v 2 Es;
                      kTf vk > 1   if v 2 Eu:
Anosov ﬂows are deﬁned similarly, with Es ˚ Eu complementary to the bundle
E0 that is tangent to the ﬂow direction. The bundles Es and Eu of an Anosov
system are tangent to the stable and unstable foliations Ws and Wu, respectively.
The properties of these foliations play a crucial role in the Hopf argument.


Keith Burns was supported by NSF grants DMS-0100416 and DMS-0408704, and Amie Wilkinson
by NSF grants DMS-0100314 and DMS-0401326.

                                   451


=== PDF PAGE 4 OF 42 ===
452                      KEITH BURNS and AMIE WILKINSON

    Hopf’s original argument established ergodicity for volume-preserving, uni-
formly hyperbolic systems under the assumption that the foliations Ws and Wu
are C 1. While the leaves of these foliations are always as smooth as the diffeomor-
phism, the foliations are usually only H¨older continuous in the direction transverse
to the leaves. In particular, for geodesic ﬂows on arbitrary compact manifolds of
negative sectional curvature, these foliations are not always C 1.
    Anosov and Sinai [AS67], [Ano67] observed that the C 1 condition on the
stable and unstable foliations in the Hopf argument could be replaced by the weaker
condition of absolute continuity, which we discuss in Section 2.2. They showed
that Ws and Wu are absolutely continuous if the system is C 2, thereby establishing
ergodicity of all C 2 volume-preserving, uniformly hyperbolic systems, including
geodesic ﬂows for compact manifolds of negative sectional curvature.
    At this point, it became clear that the Hopf argument should extend to even
more general settings. Two natural generalizations of uniform hyperbolicity are

    nonuniform hyperbolicity, which requires hyperbolicity along almost every
     orbit, but allows the expansion of Eu and the contraction of Es to weaken
     near the exceptional set where there is no hyperbolicity; and

     partial hyperbolicity, which requires uniform expansion of Eu and uniform
     contraction of Es, but allows central directions at each point, in which the
    expansion and contraction is dominated by the behavior in the hyperbolic
     directions.

The ﬁrst direction is Pesin theory; the second is the subject of this paper.
     Brin and Pesin [BP74] and independently Pugh and Shub [PS72] ﬁrst exam-
ined the ergodic properties of partially hyperbolic systems soon after the work
of Anosov and Sinai. The current deﬁnition of partial hyperbolicity is more gen-
eral than theirs, but has the same basic features.1 We say that a diffeomorphism
f  W M ! M of a compact manifold M is partially hyperbolic if the following
conditions hold. There is a nontrivial splitting of the tangent bundle, TM D
Es ˚ Ec ˚ Eu, that is invariant under the derivative map Tf . Further, there is
a Riemannian metric for which we can choose continuous positive functions   , O ,
  and O with

(1)                             ; O < 1   and   < < O  1 < O  1

such that, for any unit vector v 2 TpM,


   1The difference is that in [BP74] the functions   ; O ;  and O in the deﬁnition of partial hyperbol-
icity are assumed to be constant.


=== PDF PAGE 5 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           453

(2)                   kTf vk <  .p/          if v 2 Es.p/;
(3)                .p/ < kTf vk < O .p/ 1       if v 2 Ec.p/;
(4)                  O .p/ 1 < kTf vk                     if v 2 Eu.p/:

Partial hyperbolicity is a C 1-open condition: any diffeomorphism sufﬁciently C 1-
close to a partially hyperbolic diffeomorphism is itself partially hyperbolic. Par-
tially hyperbolic ﬂows are deﬁned similarly. For an extensive discussion of exam-
ples of partially hyperbolic dynamical systems, see the survey article [BPSW01]
and the book [Pes04]. Among these examples are the time-1 map of an Anosov
ﬂow, the frame ﬂow for a compact manifold of negative sectional curvature, and
many afﬁne transformations of compact homogeneous spaces.
     All of these examples preserve the volume induced by a Riemannian met-
ric on M. The methods based on Hopf’s argument actually apply to a slightly
larger class of invariant measures, namely those that lie in the measure class of
a volume, meaning that they have the same sets of measure 0 as a volume (note
that the volumes of any two Riemannian metrics lie in the same measure class).
By a slight abuse of notation, we say that f  is volume-preserving if it preserves
a probability measure in the measure class of a volume. In this paper, m will
denote a Riemannian volume on M (not necessarily invariant), and   will denote
an invariant probability measure that lies in the measure class of m. Measurability
is with respect to the Borel  -algebra on M.
    As in the Anosov case, the stable and unstable bundles Es and Eu of a par-
tially hyperbolic diffeomorphism are tangent to foliations, which we again denote
by Ws and Wu, respectively [BP74]. Brin-Pesin and Pugh-Shub proved that these
foliations are absolutely continuous.
   By its very nature, the Hopf argument shows that for almost every p 2 M,
almost every point of Ws.p/ and almost every point of Wu.p/ lies in the ergodic
component of p. Thus we can only hope to prove ergodicity using a Hopf argument
if something close to the following condition holds.
     Deﬁnition. A partially hyperbolic diffeomorphism f  W M ! M is accessible
if any point in M can be reached from any other along an su-path, which is a
concatenation of ﬁnitely many subpaths, each of which lies entirely in a single leaf
of Ws or a single leaf of Wu.
    The accessibility class of p 2M is the set of all q 2M that can be reached from
p along an su-path. Accessibility means that there is one accessibility class, which
contains all points. The following notion is a natural weakening of accessibility.
     Deﬁnition. A partially hyperbolic diffeomorphism f  W M ! M is essentially
accessible if every measurable set that is a union of entire accessibility classes has
either full or zero volume.


=== PDF PAGE 6 OF 42 ===
454                      KEITH BURNS and AMIE WILKINSON

    Pugh and Shub have conjectured that essential accessibility implies ergodicity
for a C 2, partially hyperbolic, volume-preserving diffeomorphism [PS97]. We
prove this conjecture under one, rather mild additional assumption.

     Deﬁnition. A partially hyperbolic diffeomorphism is center bunched if the
functions   ; O ;   , and O can be chosen so that
(5)              <   O    and      O <   O :

   We can then state our main result:
   THEOREM 0.1. Let f be C 2, volume-preserving, partially hyperbolic, and
center bunched. If f is essentially accessible, then f is ergodic and in fact has the
Kolmogorov property.

     This result extends earlier results about ergodicity of partially hyperbolic sys-
tems. Brin and Pesin [BP74] proved in the early 1970s that a C 2 volume-preserving
partially hyperbolic diffeomorphism that is essentially accessible is ergodic if it
satisﬁes the following additional conditions:

    Center bunching: Inequalities (5) hold.
    Dynamical coherence: There are foliations Wc, Wcs, and Wcu tangent to Ec,
   Ec ˚ Es, and Ec ˚ Eu, respectively.
     Lipschitzness of Wc: There are Lipschitz foliation charts for Wc.

While Brin and Pesin’s argument applies to many examples of partially hyperbolic
diffeomorphisms, the third condition is in some ways very restrictive, as Lipschitz-
ness of Wc can be destroyed by arbitrarily small perturbations [SW00a]. Brin
and Pesin’s theorem applies in particular to the time-1 map '1 of the geodesic
ﬂow for a compact surface of constant negative curvature. If we make a C 1 small
perturbation to '1, all of their hypotheses continue to hold, except Lipschitzness
of Wc (this fact follows from combining results in [Dol04] with an argument of
Ma˜n´e — see [BPSW01, p. 352]).
       It was not until the 1990s that Grayson, Pugh, and Shub [GPS94] were able
to show that any small perturbation of '1 is ergodic; in other words, '1 is stably
ergodic: any C 2 volume-preserving diffeomorphism sufﬁciently C 1-close to '1
is ergodic. The ideas in this groundbreaking paper have been generalized in sev-
eral stages [Wil98; PS97], culminating in [PS00]. The main result there assumes
dynamical coherence and uses a signiﬁcantly stronger version of center bunching
than inequalities (5), and its center bunching hypothesis requires that the action of
Tf on Ec be close to isometric — that is, both  and O (and not just their product)
must be close to 1.
   By contrast, our center bunching hypothesis requires only that the action of
Tf on Ec be close enough to conformal that the hyperbolicity of f dominates the


=== PDF PAGE 7 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           455

nonconformality of Tf on Ec. Center bunching always holds when Tf jEc is con-
formal. For then we have kTpf vk D kTpf jEc.p/k for any unit vector v 2 Ec.p/,
and hence we can choose  .p/ slightly smaller and O .p/ 1 slightly bigger than
kTpf jEc.p/k. By doing this we may make the ratio  .p/= O .p/ 1 D  .p/ O .p/
arbitrarily close to 1, and hence larger than both  .p/ and O .p/.
     In particular, center bunching holds whenever Ec is one-dimensional, and we
obtain a corollary:
   COROLLARY 0.2. Let f be C 2, volume-preserving, and partially hyperbolic
with dim.Ec/ D 1. If f is essentially accessible, then f is ergodic and in fact has
the Kolmogorov property.

     This establishes the Pugh-Shub conjecture mentioned above in the case where
Ec is 1-dimensional.
     Corollary 0.2 has also been recently proved by F. Rodr´ıguez Hertz, J. Rodr´ı-
guez Hertz, and R. Ures [RHRHU08]. Their argument is mainly based on tech-
niques in an earlier version2 of the present paper [BW05]. They prove in addition
that stable accessibility is C r dense among the C r partially hyperbolic diffeomor-
phisms with 1-dimensional center, which implies, for r   2, that stable ergodicity is
C r dense among the volume-preserving C r partially hyperbolic diffeomorphisms
with one-dimensional center. Their work establishes the main stable ergodicity
conjectures of Pugh and Shub [PS00, Conjectures 1–3] in the case where Ec is
one-dimensional.
    There is only one place in the proof of Theorem 0.1 where we need the diffeo-
                 2              1Cımorphism to be C  as opposed to C     for some ı > 0. This is when we use the
fact that center bunching implies that the stable and unstable holonomies between
center leaves are Lipschitz. This fact is proved using a graph transform argument
in [BP74] and also (in a slightly more general setting) in [PSW97], [PSW00]. In
[BW06b], we show that the same result about holonomies holds, at the expense of
a more stringent bunching hypothesis, when C 2 is replaced by C 1Cı. Plugging
this result into the proof of Theorem 0.1 gives another result:
   THEOREM 0.3. Let f be C 1Cı, volume-preserving, and partially hyperbolic.
Let   ; O be continuous functions satisfying
                   .p/ <kTf vk                   if v 2 Es.p/;
(6)
                     kTf vk < O .p/ 1    if v 2 Eu.p/:

Suppose that f satisﬁes the strong center bunching condition
(7)               <   O  and   O <   O ;

   2This earlier version proved the same result as the present paper but under the additional hypoth-
esis of dynamical coherence, i.e. the existence of foliations tangent to the bundles Ec ˚ Eu and
Ec ˚ Es.


=== PDF PAGE 8 OF 42 ===
456                      KEITH BURNS and AMIE WILKINSON

where  2 .0; ı/ satisﬁes the inequalities

(8)                         1 <     and   O O  1 < O  :

       If f is essentially accessible, then f is ergodic and in fact has the Kolmogorov
property.

   We remark that any   satisfying the conditions in (8) is a H¨older exponent
for the central distribution Ec (see e.g. [PSW97, Th. A]). It would be interesting
to know whether (7) could be replaced by (5) in Theorem 0.3. The strong center
bunching condition in (7) is automatically satisﬁed when Ec is one-dimensional,
since, as above, we may then choose  and O  1 so that   O is arbitrarily close to 1.
Theorem 0.3 has this corollary:

   COROLLARY 0.4. Let f be C 1Cı, volume-preserving, and partially hyper-
bolic with dim.Ec/ D 1. If f  is essentially accessible, then f  is ergodic and in
fact has the Kolmogorov property.

    Our arguments rely upon the same basic strategy as those of Pugh and Shub in
[PS00]. We use a similar notion of center-unstable juliennes, and show that they are
quasi-preserved by stable holonomies. But our analysis of the relationship between
juliennes and density points is fundamentally different from theirs. One essential
novelty is the use of a version of Cavalieri’s principle, introduced in Section 2.3.
Another is the use of the fake foliations constructed in Section 3. The use of Cava-
lieri’s principle greatly simpliﬁes the analysis of the relationship between juliennes
and Lebesgue density points of invariant sets, which is the heart of the proof. It
allows us to weaken the center bunching hypothesis to the minimum condition
needed to obtain smoothness of holonomy along stable leaves inside a leaf of a
center stable foliation. The fake foliations and the simpliﬁcations using Cavalieri’s
principle make it possible to eliminate the hypothesis of dynamical coherence. It
is interesting to note that both advances, fake foliations and Cavalieri’s principle,
are used in an essential way to remove the dynamical coherence assumption; the
technique of fake foliations cannot be used directly to modify or improve the Pugh
Shub approach to remove dynamical coherence.
    Our ergodicity result, Theorem 0.1, is proved in Section 5 as a consequence
of Theorem 5.1, which is really the central result of the paper. Theorem 5.1 is
proved in Sections 6, 7 and 8.


                                   1. Preliminaries

     1.1. Notations. We use the convention that if q is a point in M and j is an
integer, then qj denotes the point f j .q/, with q0 D q. If ˛ W M ! R is a positive


=== PDF PAGE 9 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           457

function and j   1 is an integer, let
                 ˛j .p/ D ˛.p/˛.p1/   ˛.pj  1/;
            ˛  j .p/ D ˛.p  j / 1˛.p jC1/ 1   ˛.p 1/  1:
We set ˛0.p/ D 1. Observe that ˛j is a multiplicative cocycle; in particular, we
have ˛  j .p/ 1 D ˛j .p  j /. Note also that .˛ˇ/j D ˛j ˇj , and if ˛ is a constant
function, then ˛n D ˛n.
    The notation ˛ < ˇ, where ˛ and ˇ are continuous functions, means that the
inequality holds pointwise. The function minf˛; ˇg has the value min.˛.p/; ˇ.p//
at the point p.
    As usual P D O.Q/ means that there is a constant C > 0 such that jP j  CQ.
Usually P and Q will depend on an integer n and one or more points in M. The
constant C must be independent of n and the choice of the points.

     1.2. Foliation boxes and local leaves. Let F be a foliation of an n-manifold
M with d-dimensional smooth leaves. For r > 0, we denote by F.x; r/ the con-
nected component of x in the intersection of F.x/ with the ball B.x; r/.
   A foliation box for F is the image U of Rn d  Rd under a homeomorphism
that sends each vertical Rd-slice into a leaf of F. The images of the vertical Rd-
slices will be called local leaves of F in U .
   A smooth transversal to F in U is a smooth codimension-d disk in U that
intersects each local leaf in U exactly once and whose tangent bundle is uniformly
transverse to T F. If  1 and  2 are two smooth transversals to F in U , we have the
holonomy map hF W  1 !  2, which takes a point in  1 to the intersection of its
local leaf in U with  2.

      1.3. Adapted metrics. We assume that the Riemannian metric on M is chosen
so that the inequalities in (1)–(4) involving   ;  ; O ; O hold. Such a metric will be
called adapted. Note that a rescaling of an adapted metric is still adapted. It will be
convenient to assume that the metric is scaled so that the geodesic balls of radius 1
are very small neighborhoods of their centers. Distance with respect to the metric
will be denoted by d.
    There is no harm in increasing  and O and decreasing  and O slightly, pro-
vided that the inequalities (1)–(4) still hold. If f  is center bunched, the change
must also be small enough so that inequalities (5) still hold. Similarly, if f  is
strongly center bunched, the change must also be small enough so that inequalities
(7) still hold.
   By rescaling the metric on M, we may assume that for some R > 1 and any
x 2 M, the Riemannian ball B.x; R/ is contained in foliation boxes for both Ws
and Wu. We assume that R is large enough so that all the objects considered in the
sequel are small compared with R. Having ﬁxed such an R, we deﬁne, for a D s


=== PDF PAGE 10 OF 42 ===
458                      KEITH BURNS and AMIE WILKINSON

or u, the local leaf of Wa through x by Waloc.x/ D Wa.x; R/. Any foliation box U
for either Ws or Wu that we consider in the rest of the paper will be small enough
so that Waloc.x/ \ U is a local leaf of Wa in U for each x 2 U . By (if necessary)
further rescaling the metric to make the local leaves smaller, we may assume that
our metric is still adapted, and that for all p 2 M and q; q0 2 B.p; R/, we have
that
(9)         q 2 Wsloc.q0/ H)  d.f .q/; f .q0//    .p/d.q; q0/;

and similarly
      q 2 Wuloc.q0/ H)  d.f  1.q/; f  1.q0//    O .f  1.p//d.q; q0/:
In particular, f .Wsloc.p//  Wsloc.f .p// and f 1.Wuloc.p//  Wuloc.f  1.p// for
all p 2 M. This is possible because  and O are continuous, and the inequalities
(2) and (4) that they satisfy are strict.
   An inductive argument then gives this:
   LEMMA 1.1.  If qj ; q0j 2 B.pj ; R/ for j D 0; : : : ; n   1, and q 2 Wsloc.q0/, then
                        d.qn; q0n/    n.p/d.q; q0/:
       If q  j ; q0 j 2 B.p  j ; R/ for j D 0; : : : ; n   1, and q 2 Wuloc.q0/, then
                   d.q n; q0 n/    O n.p/ 1d.q; q0/:

     Proof. We prove the ﬁrst claim; the second follows from the ﬁrst, with f
replaced by f  1. The proof is by induction on n. The claim is vacuously true for
n D 0. Suppose the claim holds for n D k. The inductive assumption gives that
      q0                        q0/ and q0                       2 Wsloc.qk/. Then (9), applied at pk, impliesd.qk;       k/    k.p/d.q;                              k
that q0       2 Wsloc.qkC1/, and    kC1
          d.qkC1; q0kC1/    .pk/d.qk; q0k/
                          .pk/ k.p/d.q; q0/ D  .pkC1/d.q; q0/:

                     2. Density points and absolute continuity

     2.1. Volume and density. Recall that when we say that the diffeomorphism
f  is volume-preserving, we mean that f preserves a probability measure   that
lies in the measure class of a Riemannian volume m on M.
      If S  M is a smooth submanifold, we denote by mS the volume of the
induced Riemannian metric on S. If F is a foliation with smooth leaves, and A
is contained in a single leaf of F and is measurable in that leaf, then we denote
by mF.A/ the induced Riemannian volume of A in that leaf. A set is said to be
saturated by a foliation F or F-saturated if it is a union of entire leaves of F. A set


=== PDF PAGE 11 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           459

A is essentially F-saturated if there exists a measurable F-saturated set A0, which
we call an essential F-saturate of A, with m.A  A0/ D 0.
      If   is a measure and A and B are  -measurable sets with  .B/ > 0, we deﬁne
the density of A in B by  .A W B/ D .A \ B/= .B/. We say a point x 2 M is a
Lebesgue density point of a measurable set X M if limr!0 m.X W Br.x// D 1.
Notice that the notion of Lebesgue density point depends only on the smooth struc-
ture of M, because any two Riemannian metrics have the same Lebesgue density
points.
    The Lebesgue density theorem implies that if A is a measurable set and bA is
the set of Lebesgue density points of A, then m.A  bA/ D 0.
    Lebesgue density points can be characterized using nested sequences of mea-
surable sets. We say that a sequence of measurable sets Yn nests at a point x if
Y0   Y1   Y2        fxg and Tn Yn D fxg. A sequence Yn with m.Yn/ > 0 that
nests at x is a Lebesgue density sequence at x if for every measurable set X, x is
a Lebesgue density point of X if and only if limn!1 m.X W Yn/ D 1. It is easily
shown that a Lebesgue density sequence Yn must be regular, a term we now deﬁne.

     Deﬁnition. A nested sequence of measurable sets Yn is regular if there exists
ı > 0 such that, for all n   0, we have

                m.Yn/ > 0  and  m.YnC1/  ım.Yn/:

The simplest example of a Lebesgue density sequence at x is the sequence of balls
B.x;  n/, where  2 .0; 1/.
     In proving Theorem 0.1, we characterize the Lebesgue density points of a
special class of measurable sets, those that are both essentially Ws-saturated and
essentially Wu-saturated. Such sets will be called bi-essentially saturated. These
sets arise when we consider Birkhoff averages of continuous test functions; the
sublevel sets of such averages are bi-essentially saturated.
   We say Yn is a Lebesgue density sequence at x for bi-essentially saturated
sets if Yn nests at x, Yn is regular, and, for every bi-essentially saturated set X, x
is a Lebesgue density point of X if and only if limn!1 m.X W Yn/ D 1. Note that
a Lebesgue density sequence for bi-essentially saturated sets is not necessarily a
Lebesgue density sequence, because only certain measurable sets are considered.
In fact, many of the Lebesgue density sequences for bi-essentially saturated sets
constructed in this paper are not Lebesgue density sequences.
     In our proof, we will frequently have to pass the property of being a Lebesgue
density sequence for bi-essentially saturated sets from one sequence Yn that nests
at x to another sequence Zn that nests at x. To do so, we have to show that Zn is
also regular and that, for every measurable set X that is bi-essentially saturated,

(10)          lim m.X W Yn/ D 1 ”    lim m.X W Zn/ D 1:
        n!1               n!1


=== PDF PAGE 12 OF 42 ===
460                      KEITH BURNS and AMIE WILKINSON

We have two techniques for doing this. Which technique we use depends on the
construction of Yn and Zn.
    The ﬁrst technique is very simple. Two nested sequences of sets Yn and Zn
are internested if there exists a k  1 such that, for all n   0, we have

                  YnCk  Zn  and  ZnCk   Yn:

Internestedness is an equivalence relation. The following lemma is a straightfor-
ward consequence of the deﬁnitions.

   LEMMA 2.1. Let Yn and Zn be internested sequences of measurable sets,
with Yn regular. Then Zn is also regular. If the sets Yn have positive measure, then
so do the Zn, and, for any measurable set X,
              lim m.X W Yn/ D 1 ”    lim m.X W Zn/ D 1:
        n!1               n!1

   COROLLARY 2.2. Suppose Yn and Zn both nest at x and are internested.
Then Yn is a Lebesgue density sequence for bi-essentially saturated sets if and only
if Zn is a Lebesgue density sequence for bi-essentially saturated sets.

    The second technique uses the absolute continuity of the foliations Ws and Wu,
plus the saturation properties of X; it is developed in the next two subsections.

      2.2. Absolute continuity. Our arguments in this paper use two versions of the
property of absolute continuity of a foliation.
    The ﬁrst version of absolute continuity involves holonomy maps between
transversals. A foliation F with smooth leaves is transversely absolutely contin-
uous with bounded Jacobians if for every angle ˛ 2 .0;  =2, there exists C   1
and R0 > 0 such that, for every foliation box U of diameter less than R0, any
two smooth transversals  1 and  2 to F in U of angle at least ˛ with F, and any
m 1-measurable set A contained in  1, we have
(11)          C 1m 1.A/ m 2.hF.A// Cm 1.A/:

    The second version involves a Fubini-like property. A foliation F with smooth
leaves is absolutely continuous with bounded Jacobians if, for every ˛ 2 .0;  =2,
there exist C   1 and R0 > 0 such that, for every foliation box U of diameter less
than R0, any smooth transversal   to F in U of angle at least ˛ with F, and any
measurable set A contained in U , we have the inequality

                     Z
(12)      C 1m.A/    mF.A \ Floc.x// dm .x/  Cm.A/:

       It is not difﬁcult to verify that these two properties are equivalent. Note that
the minimal C for which (11) holds is not necessarily the same as the minimal C
for which (12) holds.


=== PDF PAGE 13 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           461

    The foliations Ws and Wu for a partially hyperbolic diffeomorphism are trans-
versely absolutely continuous with bounded Jacobians. This was shown in the
Anosov case by Anosov himself [Ano67], and in the case of partial hyperbolicity
by Brin and Pesin [BP74] and by Pugh and Shub [PS72]. Their proofs were written
under the assumption that the functions   , O ,    , and O are constant. In the general
case of partial hyperbolicity, where these functions are not constant, absolute con-
tinuity of Ws and Wu follows from Pesin theory. A direct proof in this context
has been given by Abdenur and Viana [AV08]. All of these results show that the
Jacobians are continuous functions and so are bounded, since M is compact. In
general, Wc does not have either absolute continuity property, even when f  is
dynamically coherent (examples were ﬁrst constructed by Katok [Mil97]; open
sets of examples were made by Shub and Wilkinson [SW00a]).
      2.3. Saturated sets and Cavalieri’s principle. The second technique mentioned
in Section 2.1 involves decomposing the sets in a sequence nesting at x along
the leaves of an absolutely continuous foliation. The method is reminiscent of
Cavalieri’s principle, which states that the volume of a 3-dimensional solid may
computed by taking 2-dimensional slices of the solid. If the areas of these slices
are approximately equal, then the volume of the solid is approximately equal to
the product of the slice area and the total height of the slices.
     In our setting, we use the leaves of an absolutely continuous foliation to de-
compose a subset of a foliation box into slices, which we call ﬁbers. The idea
of approximately equal area of the slices translates into c-uniformity of the ﬁbers,
which we deﬁne below. When the foliation is absolutely continuous, the volume
of a subset of the foliation box with c-uniform ﬁbers will be approximately the
volume of one of the ﬁbers times the volume of the projection of the set onto a
transversal.
     Let F be an absolutely continuous foliation, and let U be a foliation box for F.
Let  be a smooth transversal to F in U . Let Y  U be a measurable set. For a
point q 2   , we deﬁne the ﬁber Y.q/of Y over q to be the intersection of Y with
the local leaf of F in U containing q. The base Y of Y is the set of all q 2  such
that the ﬁber Y.q/ is mF-measurable and mF.Y.q// > 0. The absolute continuity
of F implies that Y is m -measurable. We say “Y ﬁbers over Z” to indicate that
Z D Y .
       If, for some c   1, the inequalities c 1  mF.Y.q//=mF.Y.q0//  c hold for all
q; q0 2 Y , then we say that Y has c-uniform ﬁbers. A sequence of measurable sets
Yn contained in U has c-uniform ﬁbers if each set in the sequence has c-uniform
ﬁbers, with c independent of n.
    PROPOSITION 2.3. Suppose that the foliation F is absolutely continuous with
bounded Jacobians. Let U be a foliation box for F, and let  be a smooth transver-
sal to F in U . Then there is a constant C   1 such that for any c    1, any


=== PDF PAGE 14 OF 42 ===
462                      KEITH BURNS and AMIE WILKINSON

measurable set Y  U with c-uniform ﬁbers, and any point q0 2 Y , we have
             .Cc/ 1m.Y /  mF.Y.q0//m . Y /  Ccm.Y /:

     Proof. Absolute continuity of F with bounded Jacobians implies that there
exists a C   1 that depends only on F, U , and   , and does not depend on Y , such
that
                        Z
          C 1m.Y /     mF.Y.q// dm .q/  Cm.Y /:

Since the ﬁbers of Y are c-uniform, we have, for any q 2 Y , that
                 c 1mF.Y.q0//  mF.Y.q//  cmF.Y.q0//:

Combining these two sets of inequalities gives the desired result.

     Regularity of a sequence with c-uniform ﬁbers can be obtained from regularity
of its ﬁbers and bases.

    PROPOSITION 2.4. Suppose that the foliation F is absolutely continuous with
bounded Jacobians. Let U be a foliation box for F, and let  be a smooth transver-
sal to F in U . Let Yn be a sequence of subsets of U with c-uniform ﬁbers. Suppose
that there exists a ı > 0 such that, for all n   0,

  (i) m . YnC1/  ım . Yn/ and
                         and z0 2 (ii) there are points z 2                                          with mF.YnC1.z//   ımF.Yn.z0//:                                      Yn                       YnC1

Then Yn is regular.

     Proof. It follows from Proposition 2.3 that
              m.YnC1/  .cC/ 1mF.YnC1.z//m . YnC1/;
                m.Yn/  cCmF.Yn.z0//m . Yn/:
Using the two properties of ı, we then obtain m.YnC1/  .cC/ 2ı2m.Yn/, which
says that Yn is regular.

   We now turn to the second technique mentioned above for proving an equiv-
alence of the form (10). It leads to the main result we prove in this section:

    PROPOSITION 2.5. Let F be absolutely continuous with bounded Jacobians,
and let U be a foliation box for F with smooth transversal   . Let Yn and Zn
be sequences of measurable subsets of U with c-uniform ﬁbers for some c    1.
Suppose that  D     for all n. Then, for any essentially F-saturated set X  U ,             Yn    Zn
we have that

              lim m.X W Yn/ D 1 ”    lim m.X W Zn/ D 1:
        n!1               n!1


=== PDF PAGE 15 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           463

    COROLLARY 2.6. Fix F D Ws or Wu. Suppose that Yn and Zn satisfy the hy-
potheses of Proposition 2.5 and that Yn and Zn both nest at x. If Yn is a Lebesgue
density sequence at x for bi-essentially saturated sets, then so is Zn, provided it is
regular.

    Before proving Proposition 2.5, we establish a related result, which will also
be used in the proof of Theorem 0.1.

    PROPOSITION 2.7. Let F be absolutely continuous with bounded Jacobians,
and let U be a foliation box for F with smooth transversal   . Suppose that fYngn 0
is a sequence of measurable sets in U with c-uniform ﬁbers for some c   1. Then,
for every F-saturated measurable set X, we have that
             lim m.X W Yn/ D 1 ”                                         lim m . X W Yn/ D 1:        n!1                          n!1
    Remark. The hypothesis that X is F-saturated can be weakened. In fact, it
sufﬁces for X \ U to be a union of local leaves of F in U .

     Proof of Proposition 2.7. Let X  be the complement of X in M. Then X   is
also F-saturated. The proposition can be reformulated in terms of X  . We have to
prove that
            lim m.X   W Yn/ D 0 ”                                          lim m . X   W Yn/ D 0:       n!1                          n!1

    For each n, let
                  mn D  inf  mF.Yn.q//:                              q2                                     Yn
Since the ﬁbers of Yn are c-uniform, it follows that mn > 0 for all n, and mn
mF.Yn.q//  cmn for all q 2  Yn. Absolute continuity implies that there exists a
C   1 such that
                       Z
          C 1m.Yn/     mF.Yn.q// dm .q/  Cm.Yn/:

Together, these inequalities imply that
(13)         C 1mnm . Yn/  m.Yn/  Ccmnm . Yn/:
     Since X   is F-saturated, the F-ﬁber of X \Yn over a point q 2 Yn is either
empty or equal to Yn.q/. Thus X \ Yn also has c-uniform ﬁbers, and, as above,
we obtain
(14)    C 1mnm . X \Yn/  m.X \ Yn/  Ccmnm . X \Yn/:
    Noting that X \Yn D X \ Yn and dividing the inequalities in (14) by those
in (13), we obtain
        .C 2c/ 1m . X   W Yn/  m.X   W Yn/  C 2cm . X   W  Yn/:

The result follows easily from this.


=== PDF PAGE 16 OF 42 ===
464                      KEITH BURNS and AMIE WILKINSON

     Proof of Proposition 2.5. Let X0 be an essential F-saturate of X. Using
Proposition 2.7, we have the equivalences
   lim m.X W Yn/ D 1 ” lim m.X0 W Yn/ D 1
 n!1            n!1
     ”                           lim m . X0 W Yn/ D 1                n!1
     ”                           lim m . X0 W Zn/ D 1                n!1
     ” lim m.X0 W Zn/ D 1 ” lim m.X W Zn/ D 1:
                n!1             n!1

                             3. Fake invariant foliations

    The Lebesgue density sequences for bi-essentially saturated sets that we will
use in this proof will be constructed using dynamical foliations with uniform conti-
nuity properties. If f happens to be dynamically coherent, then we are free to use
the foliations Ws; Wu; Wc; Wcs, and Wcu for these constructions. Since we are not
assuming dynamical coherence, we must ﬁnd substitutes for Wc; Wcs, and Wcu
to make our proof work in general. It turns out to be simplest to ﬁnd substitutes
for all invariant foliations Ws; Wu; Wc; Wcs, and Wcu. We call these substitutes
“fake invariant foliations.” There are a few key places in the argument where we
will have to use the real invariant foliations Wu and Ws, rather than their fake
counterparts. We will indicate where this is the case. The reader should recall the
choice of the constant R from Section 1.3.
    PROPOSITION 3.1. Let f  W M ! M be a C 1 partially hyperbolic diffeomor-
phism. For any " > 0, there exist constants r and r1 with R > r > r1 > 0 such that,
for every p 2 M, the neighborhood B.p; r/ is foliated by foliations Wup, Wsp, Wcp,                                     b  b  b
Wcup  and Wcsp with the following properties, for each ˇ 2 fu; s; c; cu; csg.b    b
  (i) Almost tangency to invariant distributions. For each q 2 B.p; r/, the leaf
   Wˇp .q/ is C 1, and the tangent space Tq Wˇp .q/ lies in a cone of radius " about   b                    b
    Eˇ.q/.
 (ii) Local invariance. For each q 2 B.p; r1/,
     p .q; r1//  Wˇ                                      p .q; r1//  Wˇ  f .bWˇ           bf .p/.f .q//  and  f 1.bWˇ                                 bf  1.p/.f  1.q//:
(iii) Exponential growth bounds at local scales. The following hold for all n   0:
      (a) Suppose that qj 2 B.pj ; r1/ for 0   j  n   1.
            If q0 2 Wsp.q; r1/, then         b
                    q0n 2 Wsp.qn; r1/  and  d.qn; q0n/    n.p/d.q; q0/:             b
            If q0j 2 Wcsp .qj ; r1/ for 0   j  n   1, then         b
                    q0n 2 Wcsp .qn/    and  d.qn; q0n/    On.p/ 1d.q; q0/:             b


=== PDF PAGE 17 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           465

      (b) Suppose that q  j 2 B.p  j ; r1/ for 0   j  n   1.
             If q0 2 Wup.q; r1/, then         b
                   q0 n 2 Wup.q n; r1/  and  d.q n; q0 n/    O n.p/ 1d.q; q0/:              b
             If q0 j 2 Wcup .q  j ; r1/ for 0   j  n   1, then          b
                   q0 n 2 Wcup .q n/    and  d.q n; q0 n/     n.p/d.q; q0/:              b
 (iv) Coherence. Wsp and Wcp subfoliate Wcsp  , and Wup and Wcp subfoliate Wcup  .          b    b       b    b    b       b
 (v) Uniqueness. Wsp.p/ D Ws.p; r/, and Wup.p/ D Wu.p; r/.          b             b
 (vi) Regularity.  If f  is C 1Cı, then the foliations Wup, Wsp, Wcp, Wcup  , and Wcsp                            b  b  b  b     b
    and their tangent distributions are uniformly Hölder continuous.
(vii) Regularity of the strong foliation inside weak leaves. If f  is C 2 and center
     bunched, then each leaf of Wcsp  is C 1 foliated by leaves of the foliation Wsp,                  b                        b
    and each leaf of Wcup  is C 1 foliated by leaves of the foliation Wup.  If f  is             b                        b      1Cı   C    and strongly center bunched, then the same conclusion holds.

The regularity statements in (vi) and (vii) hold uniformly in p 2 M.

     Proof. Suppose that f  is C k for some k   1. After possibly reducing ", we
can assume that inequalities (1)–(4) hold for unit vectors in the "-cones around the
 spaces in the partially hyperbolic splitting.
    The construction of the leaves of Wcup  and Wcsp through p is essentially the                      b    b
same as the proof of the existence of pseudo-hyperbolic plaque families in [HPS77].
They are obtained as ﬁxed points of graph transforms of a map that coincides with
f in a neighborhood of the orbit of p. We take the argument one step further and
 consider all ﬁxed points of these graph transforms in the entire neighborhood of p.
    Our construction will be performed in two steps. In the ﬁrst, we construct
 foliations of each tangent space TpM. In the second step, we use the exponential
map expp to project these foliations from a neighborhood of the origin in TpM to
 a neighborhood of p.

     Step 1. We choose an r0 > 0 such that expp 1 is deﬁned on B.p; 2r0/. For
 r 2 .0; r0, we deﬁne, in the standard way, a map

                                      Fr
                TM     TM


                                   f
              M     M;

which is uniformly C k on ﬁbers, satisfying

  (a) Fr.p; v/ D expf .p/1  ıf ı expp.v/ for kvk   r,


=== PDF PAGE 18 OF 42 ===
466                      KEITH BURNS and AMIE WILKINSON

 (b) Fr.p; v/ D Tpf .v/ for kvk   2r, and
 (c) kFr.p; /  Tpf . /kC 1 ! 0 as r ! 0, uniformly in p.

    Endowing M with the discrete topology, we regard TM as the disjoint union
of its ﬁbers. Property (c) implies that, if r is small enough, then Fr is partially
hyperbolic, and each bundle in the partially hyperbolic splitting for Fr at v 2 TpM
lies within the "=2-cone about the corresponding subspace of TpM in the partially
hyperbolic splitting for f at p (we are making the usual identiﬁcation of TvTpM
with TpM). If r is small enough, the equivalents of inequalities (1)–(4) will hold
with Tf replaced by TFr.
      If r is sufﬁciently small, standard graph transform arguments give stable, un-
stable, center-stable, and center-unstable foliations for Fr inside each TpM. These
foliations are uniquely determined by the extension Fr and the requirement that
their leaves be graphs of functions with bounded derivatives. We obtain a center fo-
liation by intersecting the leaves of the center-unstable and center-stable foliations.
While TM is not compact, all of the relevant estimates for Fr are uniform, and it
is this, not compactness, that counts. The proof of the Hadamard-Perron theorem
in [KH95] contains many of the details of this argument.
    The uniqueness of the stable and unstable foliations implies, via a standard
argument (see e.g. [HPS77, Th. 6.1(e)]), that the stable foliation subfoliates the
center-stable, and the unstable subfoliates the center-unstable.
   We now discuss the regularity properties of these foliations of TM. Recall the
standard method for determining the regularity of invariant bundles and foliations.
Suppose that TX D E1 ˚ E2 is a Tg-invariant splitting of the tangent bundle for a
C k diffeomorphism g W X ! X satisfying, for every p 2 X and every unit vector
v 2 TpX, the inequalities
                ˛1.p/ < kTgvk < ˇ1.p/;   if v 2 E1.p/;
                ˛2.p/ < kTgvk                if v 2 E2.p/;
where 0 < ˛1.p/ < ˇ1.p/ < ˛2.p/. Then the bundle E2 is C a for any a  k  1
that satisﬁes
                               ˇ1.p/
                          sup        < 1:
                  p2M ˛1.p/a˛2.p/

   When the functions ˛1, ˛2, and ˇ2 are constant, this fact is classical — see
e.g. the C r Section Theorem in [HPS77]. The general case of this result appears
more recently in the literature [SSS92], [Has94], [Wil98], [PSW97]. This result
extends, at least in part, to give regularity of invariant foliations. In particular,
when there is a foliation F2 tangent to E2 that arises at the unique ﬁxed point of
a nonlinear graph transform, then F2 is a C a foliation [PSW97], [PSW00]. These
results are proved in the compact case, but compactness is used only to obtain


=== PDF PAGE 19 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           467

uniform estimates on the functions ˛1; ˇ1; ˛2 and the derivative of g; the results
carry over as long as such uniform estimates hold.
    Our foliations of TM have been constructed as the unique ﬁxed points of
graph transform maps. We can apply the above results to the Fr-invariant splittings
of T TM as the sum of the stable and center-unstable bundles for Fr and as the
sum of the center-stable and unstable bundles for Fr. It follows immediately that
both the center-unstable and unstable bundles and the corresponding foliations are
                                1CıH¨older continuous as long as Fr is C     for some ı > 0. We obtain the H¨older
continuity of the center-stable and stable bundles for Fr and the corresponding
foliations by thinking of the same splittings as Fr 1 -invariant. H¨older regularity of
the center bundle and foliation is obtained by noticing the center is the intersection
of the center-stable and center-unstable.
   When k   2, a similar estimate gives the C 1 regularity of the unstable bundle
along the leaves of the center-unstable foliation. The manifold X is the disjoint
union of the leaves of the center-unstable foliation for Fr, E2 is the unstable
bundle, and E1 is the center bundle. We have ˛1 D   , ˇ1 D O  1, and ˛2 D O  1.
The center bunching hypothesis O <   O implies that
                                               O .p/
                           sup       < 1
                    p2M  .p/a O .p/
for some a > 1. It follows that E2 D Eu is a C 1 bundle over X, the leaves of the
center-unstable foliation. Similarly, this argument shows that the center bunching
hypothesis <   O implies that Es is a C 1 bundle over the leaves of the center-
stable foliation. There is an additional difﬁculty in the proof of this estimate, which
is that X is not a C k manifold in general, even when Fr is C k. This difﬁculty is
dealt with in [PSW97], [PSW00].
   When 1 < k < 2, this type of estimate does not work at all. A different argu-
ment, working with holonomy maps instead of bundles, is presented in [BW06b].
The main result there implies that under the strong center bunching hypothesis
on f , which carries over to Fr, the unstable foliation C 1-subfoliates the center-
unstable, and the stable foliation C 1-subfoliates the center-stable.
     Step 2. We now have foliations of TpM for each p 2 M. We obtain the
foliations Wup; Wcp; Wsp; Wcup  , and Wcsp by applying the exponential map expp to      b  b  b  b     b
the corresponding foliations of TpM inside the ball around the origin of radius r.
      If r is sufﬁciently small, then the distribution Eˇ.q/ lies within the angular
"=2-cone about the parallel translate of Eˇ.p/ for every ˇ 2 fu; s; c; cu; csg and
all p; q with d.p; q/   r. Combining this fact with the preceding discussion, we
obtain that property (i) holds if r is sufﬁciently small.
     Property (ii), local invariance, follows from invariance under Fr of the folia-
tions of TM and the fact that expf .p/.Fr.p; v// D f .expp.p; v// if kvk   r.


=== PDF PAGE 20 OF 42 ===
468                      KEITH BURNS and AMIE WILKINSON

    Having chosen r, we now choose r1 small enough so that

    f .B.p; 2r1//  B.f .p/; r/  and  f  1.B.p; 2r1//  B.f  1.p/; r/;

and so that, for all q 2 B.p; r1/,
        q0 2 Wsp.q; r1/ H) d.f .q/; f .q0//    .p/ d.q; q0/;       b
        q0 2 Wup.q; r1/ H) d.f  1.q/; f  1.q0//    O .f  1.p// d.q; q0/;       b
        q0 2 Wcsp .q; r1/ H) d.f .q/; f .q0//    O .p/ 1 d.q; q0/;       b
       q0 2 Wcup .q; r1/ H) d.f  1.q/; f  1.q0//    .f  1.p// 1 d.q; q0/:       b

     Property (iii), exponential growth bounds at local scales, is now proved by an
inductive argument similar to the proof of Lemma 1.1.
     Properties (iv)–(vii) — coherence, uniqueness, regularity, and regularity of the
strong foliation inside weak leaves — follow immediately from the corresponding
properties of the foliations of TM discussed above.

    Remark. Note that the system of local foliations constructed in Proposition 3.1
is not unique; it depends on the extension of Fr outside of a neighborhood of the
zero-section of TM. Also, even when f  is dynamically coherent, in general there
is no reason to expect the fake invariant foliations Wcsp ; Wcup  , and Wcp to coincide                            b  b     b
with the local leaves of the real invariant foliations Wcs; Wcup  , and Wcp, even at p.
For the leaf Wcsp .p/ to coincide with the local leaf of Wcs through p, it is necessary       b
that every iterate f  n.Wcs.pn; r// overﬂow the neighborhood B.p; r/.

     For the rest of the paper, Wsp, Wcp, Wsp, Wcsp and Wcup  will denote fake invari-                b  b  b  b    b
ant foliations given by Proposition 3.1, with " > 0 much less than the angle between
any two of the subspaces in the partially hyperbolic splitting. We may rescale the
metric so that the radius r1 in conclusion (ii) of Proposition 3.1 is much bigger
than 1. This will ensure that all of the objects used in the rest of the paper are well
deﬁned. We may assume that if maxfd.x; p/; d.y; p/g   3, then Wcsp .x/\ Wup.y/,                                   b     b
Wcsp .x/ \ Wuloc.y/, Wcup .x/ \ Wsp.y/, and Wcup .x/ \ Wsloc.y/ are single points.b          b     b      b
We denote by Oma the measure mbWa.

                4. Distortion estimates inside thin neighborhoods

     4.1. A simple distortion lemma. The next lemma will be used to compare
values of H¨older cocycles at nearby points.

   LEMMA 4.1. Let ˛ W M ! R be a positive Hölder continuous function, with
exponent > 0. Then there exists a constant H > 0 such that the following hold


=== PDF PAGE 21 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           469

for all p; q 2 M, B > 0 and n   1:

             n 1
                                           ˛n.p/
                                               eHB:    X d.pi; qi/  B  H)   e HB
                                              ˛n.q/             iD0
           n
                                      ˛ n.p/
                                                eHB:   X d.p  i; q  i/  B  H)   e HB
                                      ˛ n.q/          iD1

     Proof. We prove the ﬁrst part of the lemma. The second part is proved sim-
ilarly. The function log ˛ is also H¨older continuous with exponent   . Let H > 0
be the H¨older constant of log ˛ such that, for all x; y 2 M,

                        jlog ˛.x/  log ˛.y/j  Hd.x; y/  :

The desired inequalities are equivalent to jlog ˛n.p/   log ˛n.q/j  HB: By ex-
panding log ˛n as a series, we obtain

                                    n 1
               jlog ˛n.p/  log ˛n.q/j X jlog ˛.pi/  log ˛.qi/j
                                   iD0

                                      n 1
                   H X d.pi; qi/   HB;
                                     iD0
since PniD01 d.pi; qi/  B by the hypothesis of the lemma.

     4.2. Thin neighborhoods of Ws.p; 1/. We next identify, for each n  0 and
p 2 M, a neighborhood of p whose ﬁrst n iterates remain in a uniform neighbor-
hood of the corresponding iterates of p. We give an exponential estimate of the size
of the ﬁrst n iterates of the n-th such neighborhood. In our proof of Theorem 0.1,
we construct sequences of geometric objects; the n-th term in the sequence of
objects for any x 2 Ws.p; 1/ will lie in the n-th neighborhood of p.
    Let  < 1 be a continuous function. For n   0 and p 2 M, deﬁne the set
Sn; .p/ by
                     Sn; .p/ D [  Wcp.x;  n.p//:                        b
                                x2Ws.p;1/

   LEMMA 4.2. Suppose that   satisﬁes < minf O ; 1g. Then

              f j .Sn; .p//  B.pj ; 2/   for j D 0; : : : ; n.

     Further, there exist positive constants < 1 and C > 0 such that, for every
n   0,
            f j .Sn; .p//  B.pj ; C  j /   for j D 0; : : : ; n:


=== PDF PAGE 22 OF 42 ===
470                      KEITH BURNS and AMIE WILKINSON

     Proof. Suppose that x 2 Ws.p; 1/ and y 2 Wcp.x;  n.p//. By part (iii.a) of                           b
Proposition 3.1, we then have
     yj 2 Wcpj .xj ; Oj 1 .p/ n.p//  Wcpj .xj ; 1/  B.pj ; 2/   for 0   j   n.      b              b
In fact, since  < minf O ; 1g, the quantity Oj 1 .p/ n.p/ < Oj 1 .p/ j .p/ is expo-
nentially small in j, as is the diameter of f j .Ws.p; 1//. This implies the second
conclusion.
   Now let    1 be another continuous function. For every x 2 Sn; .p/, we
have that B.xn; n.p//  B.pn; r/, and so the set
        Tn;  ; .p/ D f  n  [   Wupn.z; n.p// [ Wu.z; n.p//                      b
                       z2f n.Sn; .p//
is well deﬁned. Proposition 3.1 and Lemma 1.1 imply that the leaves of Wupj and                                        b
Wuloc are uniformly contracted by f  1 as long as they stay near the orbit of p;
combining these facts with Lemma 4.2, we get:
   LEMMA 4.3. For every continuous function   satisfying  < minf O ; 1g, the
set Tn; ;1.p/ satisﬁes
             f j .Tn; ;1.p//  B.pj ; 3/   for j D 0; : : : ; n.

     Further, for every such  and every continuous function  < 1, there exist
positive constants < 1 and C > 0 such that, for every n   0,
            f j .Tn;  ; .p//  B.pj ; C  j /   for j D 0; : : : ; n.

    The dimensions of the neighborhoods Tn;  ; .p/ and their iterates are illus-
trated in Figure 1, in the case where   ; < 1.
   We then obtain a simple corollary of this lemma and Lemma 4.1:
   LEMMA 4.4. Let ˛ W M ! R be a positive, uniformly Hölder continuous func-
tion, and let  and  be continuous functions satisfying < minf O ; 1g and < 1.
Then there is a constant C   1 such that, for all n  0 and all x; y 2 Tn;  ; .p/,
                                 1   ˛n.y/                  C            C:
                                ˛n.x/

                                5. The main theorem

    The properties of accessibility and essential accessibility can be reformulated
using the notion of saturation. We say that a set is bisaturated if it is both Wu-
saturated and Ws-saturated. Accessibility means that a set which is bisaturated
must be either empty or all of M. Essential accessibility means that a measurable
set which is bisaturated must have either 0 or full volume.
   We can then state the central result of this paper:


=== PDF PAGE 23 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           471



                  p
             p bWs

       p         bWu     bWc              T1,σ,τ(p)                                             Tn,σ,τ(p)                Tn+1,σ,τ(p)


          f (T1,σ,τ(p))                                      f (Tn,σ,τ(p))          f (Tn+1,σ,τ(p))



                                     f n (Tn,σ,τ(p))         f                                                             n (Tn+1,σ,τ(p))



                                                         f n+1 (Tn+1,σ,τ(p))



    Figure 1. Dimensions of the neighborhoods Tn;  ; .p/ and their iterates.


   THEOREM 5.1. Let f be C 2, partially hyperbolic and center bunched (or
  1CıC    and strongly center bunched). Let A be a measurable set that is both es-
sentially Wu-saturated and essentially Ws-saturated. Then the set of Lebesgue
density points of A is Wu-saturated and Ws-saturated.

    Theorem 5.1 does not assume that f  is volume-preserving.
    The central result of Pugh and Shub in [PS00] is a version of Theorem 5.1,
which involves a different notion of density point (which they deﬁne) and a slightly
different hypothesis:

    For a D u or s, if A is essentially Wa-saturated, then the set of julienne density
points of A is Wa-saturated.

In contrast, Theorem 5.1 requires that A be both essentially Wu-saturated and
essentially Ws-saturated in order to conclude anything.
    With the obvious deﬁnitions, Theorem 5.1 has a corollary:

   COROLLARY 5.2. If f  is as in Theorem 5.1, then every bi-essentially satu-
rated set is essentially bisaturated.


=== PDF PAGE 24 OF 42 ===
472                      KEITH BURNS and AMIE WILKINSON

     Proof. Let A be a bi-essentially saturated set. Theorem 5.1 implies that the set
OA of Lebesgue density points of A is bisaturated. The Lebesgue density theorem
implies that m.A  OA/ D 0. Thus A is essentially bisaturated: it differs by a zero
set from a bisaturated set.

     Essential accessibility tells us that essentially bi saturated sets have 0 or full
measure. A priori, the notion of bi-essential saturation is weaker than the notion of
essential bisaturation. Corollary 5.2 says that the two concepts coincide when f  is
center bunched, so that if f  is essentially accessible, then bi-essentially saturated
sets must have 0 or full measure.
    Theorem 0.1 follows easily from Corollary 5.2 by a version of the Hopf argu-
ment and a result of Brin and Pesin (see [BPSW01, 2] for more details).

    Proof of Theorem 0.1. Let   be the f -invariant probability measure in the
measure class of m. The concepts of  -almost everywhere and m-almost every-
where are the same. To prove that f is ergodic with respect to    , it sufﬁces to show
that the Birkhoff averages of continuous functions are almost everywhere constant.
Let ' be a continuous function, and let

                    n                                   n
                1                                1   O's.p/ D lim sup X '.f i.p//  and  O'u.p/ D lim sup X '.f  i.p//
       n!1 n                   n!1 n                   iD0                                 iD0
be the forward and backward Birkhoff averages of ' under f . The function O's is
constant along Ws-leaves, and O'u is constant along Wu-leaves. It follows that for
any a 2 R, the sets
            As.a/ D O' s 1 . 1; a  and  Au.a/ D O' u 1 . 1; a:
are Ws-saturated and Wu-saturated, respectively.
    Since f preserves    , the Birkhoff ergodic theorem implies that O's D O'u
almost everywhere. Consequently m.As.a/  Au.a// D 0, so that the set A.a/ D
Au.a/ \ As.a/ has Au.a/ as an essential Wu-saturate and As.a/ as an essential
Ws-saturate. Thus A.a/ is bi-essentially saturated.
       It follows from Corollary 5.2 that A.a/ is essentially bisaturated. Essential
accessibility implies that A.a/ has 0 or full measure. Since a was arbitrary, it
follows that O's and O'u are almost everywhere constant, and so f  is ergodic.
    To prove that f has the Kolmogorov property, it sufﬁces to show that all sets in
the Pinsker subalgebra P have 0 or full measure. According to [BP74, Proposition
5.1], if f  is partially hyperbolic, then any set in P 2 P is bi-essentially saturated.
It again follows from Corollary 5.2 and essential accessibility that P has 0 or full
measure.

     In order to prove Theorem 5.1, it sufﬁces to show that the set of Lebesgue
density points of A is Ws-saturated; applying this result with f replaced by f  1


=== PDF PAGE 25 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           473

then shows that the set of Lebesgue density points of A is also Wu-saturated. More
precisely, it sufﬁces to show that, for any p 2 M, if x; x0 2 Ws.p; 1/ and x is a
Lebesgue density point of A, then so is x0.
     Let N
                 N D G B.pj ; r/
                                        j  0

be the disjoint union over j   0 of the balls B.pj ; r/, where r   1 is given by
Proposition 3.1. Everything we do in the rest of this paper takes place inside N,
and we drop the dependence on p when doing so is not confusing.
   We let Wu be the locally invariant foliation of N whose restriction to B.pj ; r/       b
is Wupj . Similarly we deﬁne foliations Wc; Ws; Wcu and Wcs. By Proposition 3.1, b                  b b b    b
all of these foliations are uniformly H¨older continuous, Wu uniformly C 1 subfo-
                                1            bliates Wcu, and Ws uniformly C  subfoliates Wcs. The foliations Wu and Ws
    b     b                b
induce foliations of N , which will again be denoted by Wu and Ws. Hence Ws.p/
is used to denote the local leaf Ws.p; r/. Note that

       Ws.pj / D Ws.pj /  and  Wu.pj / D Wu.pj /   for all j    0.     b              b

Note also that, since we are not assuming that f  is dynamically coherent, there are
no foliations Wc; Wcu, or Wcs.
    Most leaves of these foliations (with the notable exception of the leaves of
Ws and Ws passing through the orbit of p) are invariant under only ﬁnitely many
     b
iterates of f , until their orbits leave N. In our proof, the geometric objects in
N that we need to iterate n times always start out in a neighborhood Tn;  ;  of
the type deﬁned in Section 4.2 with  < 1 and  < minf O ; 1g. By Lemma 4.3,
these objects remain inside of N for n iterates. As long as their orbits remain
inside of N, the locally invariant foliations Wu, Wc, Ws, Wcu, and Wcs used to
                        b  b  b  b     b
construct these geometric objects are nearly indistinguishable from their invariant
counterparts Wu, Wc, Ws, Wcu, and Wcs in the dynamically coherent setting.
    These locally invariant foliations differ from true invariant foliations in one
key respect. The measurable set A in Theorem 5.1 is essentially Ws-saturated
and essentially Wu-saturated. However, the property of (essential) Wu-saturation
neither implies nor is implied by (essential) Wu-saturation. Similarly, (essential)
                         b
Ws-saturation neither implies nor is implied by (essential) Ws-saturation. While
                                 b
it is possible to prove a version of Theorem 5.1 in which Wu and Ws are replaced
by Wu and Ws, such a theorem is not enough to prove ergodicity using a Hopf
  b    b
argument. This is because the Hopf argument uses inﬁnitely many iterates of f
and is thus global in nature. Therefore, wherever we use explicitly the fact that our
set A is Wu-saturated, in particular, in our arguments that use Proposition 2.5, we
must switch from using Wu to using Wu, and similarly for stable foliations.
             b


=== PDF PAGE 26 OF 42 ===
474                      KEITH BURNS and AMIE WILKINSON

     Following Pugh and Shub, we consider for each x 2 Ws.p; 1/ a sequence of
sets, called center-unstable juliennes, that lie in the fake center-unstable manifold
Wcu.x/ and shrink exponentially as n ! 1 while becoming increasingly thin in
b
the Wu-direction. Pugh and Shub’s construction assumes dynamical coherence
  b
and uses the true invariant foliations Wcu, Wc, and Wu; here we use their fake
counterparts. While the objects we work with, such as center-unstable juliennes,
will depend on the fake invariant foliations given by Proposition 3.1, the ﬁnal con-
clusion of Theorem 5.1 does not, since we always have Ws.p/ D Ws.p/.
                               b
     Recall the center bunching assumptions from (5) that

                 <   O  and   O <   O :

In what follows, we will use only the ﬁrst of these inequalities. The second in-
equality is used to prove Wu saturation of density points.
   We choose continuous functions  and   such that

             < <     and  < minf O ; 1g:

Note that these inequalities also imply that   O <     O <     O        . The choice
of  and  with the desired properties is possible because of the center bunching
assumption. The reader should think of   as being just a little bigger than  and
  as just a little bit less than minf O ; 1g. The reader might also choose to keep in
mind the case where the functions   , O ,   , and O are constants, and where  and
  can be chosen to be constant. In this case, the cocycles  n and  n are just the
constants  n and  n.
    Using the notation deﬁned in Section 4.2, let Sn D Sn; .p/, and let Tn D
Tn;  ; .p/. For the rest of the paper, except where we indicate otherwise, we will
evaluate cocycles at the point p. We will also drop the dependence on p from the
notation; thus, if ˛ is a cocycle, then ˛n.p/ will be abbreviated to ˛n.
    The center-unstable juliennes OJncu .x/ that we construct will be contained in
Tn. We now describe the construction of center-unstable juliennes.
    For all x 2 Ws.p; 1/, deﬁne

                                bBcn.x/ D Wc.x;b      n/:
Note that Sn D S x2Ws.p;1/ bBcn.x/. For y 2 Sn, we may then deﬁne two types of
unstable juliennes by

                                n//  and  Jnu .y/ D f  n.Wu.yn;  n//:          OJnu .y/ D f n.bWu.yn;
Observe that for all y 2 Sn, the sets OJnu .y/ and Jnu .y/ are contained in Tn.


=== PDF PAGE 27 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           475


             Wu



                                           n
                               Bcn bBc

                                Jun  ˆJun               x     Wc           x





    Center-unstable julienne Jncu .x/ in [PS00].               OJncu .x/ in this paper.

       Figure 2. Two types of center-unstable juliennes, when  and
        are constant.


    For each x 2 Ws.p; 1/ and n   0, we then deﬁne the center-unstable julienne
centered at x of order n by
                                OJncu .x/ D [    OJnu .y/:
                              y2 bBcn.x/
By their construction, the sets OJncu .x/ are contained in Tn for all n   0 and x 2
Ws.p; 1/. Also OJncu .x/ is contained in the smooth submanifold Wcu.x/, which                                    b
                                                           .x/ has positivecarries the restricted Riemannian volume Omcu D  Wcu, and OJncu
                                                 u is that the distributionOmcu-measure. The reason for using OJ u here insteadmbof J
T Wc ˚ Eu is not necessarily integrable. The set deﬁned by replacing OJnu by J nu b
in the deﬁnition of OJncu might not be a C 1 submanifold of M.
    Our cu-juliennes are closely related to, but not exactly the same as, those of
Pugh and Shub. In the case where   and   are constant functions and f  is dynami-
cally coherent, their center-unstable julienne is the foliation product of Wc.x;  n/
and f  n.Wu.xn;  n//; see Figure 2. In this case, the image under f n of our
Jncu .p/ appears in [PS00] as a tubelike approximation to Pugh and Shub’s center-
unstable postjulienne of rank n. The results of [PS00] show that the cu-juliennes
deﬁned here and in [PS00] are internested. Thus our cu-juliennes could be replaced
by the Pugh-Shub cu-juliennes in Propositions 5.3 and 5.5.


=== PDF PAGE 28 OF 42 ===
476                      KEITH BURNS and AMIE WILKINSON

    As in [PS00], the cu-juliennes have a quasiconformality property: they are
approximately preserved by holonomy along the stable foliation.
    PROPOSITION 5.3. Let x; x0 2 Ws.p; 1/, and let hs W Wcu.x/ ! Wcu.x0/
                                 b      b
be the holonomy map induced by the stable foliation Ws. Then the sequences
hs.OJncu .x// and OJncu .x0/ are internested.

     Next, there are estimates on the volumes of unstable and center-unstable juli-
ennes.

    PROPOSITION 5.4. There exist ı > 0 and c   1 such that the following hold
for all x 2 Ws.p; 1/, all q; q0 2 Sn, and all n   0:
         1   Omu.OJnu .q//                  1   mu.Jnu .q//      c                              c;          c          u        c;                 u             Omu. OJn .q0//                mu.Jn .q0//
      Omu.OJnC1.q//u         ı Omu.OJnu .q//;     Omcu.OJnC1.x//cu        ı Omcu.OJncu .x//:

    The ﬁnal crucial property of the cu-juliennes is that, for the sets that appear in
the proof of Theorem 5.1, Lebesgue density points are precisely cu-julienne density
points.
    PROPOSITION 5.5. Let X be a measurable set that is both Ws-saturated and
essentially Wu-saturated. Then x 2 Ws.p/ is a Lebesgue density point of X if and
only if limn!1 Omcu.X W OJncu .x// D 1:
Note the asymmetry that X is fully Ws-saturated but only essentially Wu-saturated.
This is because X will be an essential stable saturate of a bi-essentially saturated
set.

    Remark. Pugh and Shub show that Lebesgue almost every point of any mea-
surable set is a cu-julienne density point. In their argument they prove a Vitali
covering lemma for their juliennes. This argument accounts for their deﬁnition
of cu-juliennes as a foliation product and for the stronger bunching hypothesis in
their main result. We do not know whether their result, speciﬁcally [PS00, Th. 7.1],
still holds under our weaker bunching hypothesis, or whether it holds at all in the
absence of dynamical coherence (although it can be shown that in the dynamically
coherent, symmetrized setting that they consider, their hypothesis can be weakened
          2C2= 0         1=from <          to <      0).

    The proof of Proposition 5.3 is essentially contained in [PS00]. For complete-
ness we reproduce the argument in the next section. Proposition 5.4 is proved in
Section 7, and the proof of Proposition 5.5 is in the ﬁnal section. We now use these
three propositions to prove the main result.

     Proof of Theorem 5.1. As we noted above, it sufﬁces to show that the Lebesgue
density points of A are Ws-saturated; to see that the Lebesgue density points of A


=== PDF PAGE 29 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           477

are Wu-saturated, just consider f  1 instead of f . Let As be an essential Ws-sat-
urate of A. Since m.A  As/ D 0, the Lebesgue density points of A are precisely
the same as those of As. Fix p 2 M, and suppose that x 2 Ws.p; 1/ is a Lebesgue
density point of As. Proposition 5.5 implies that x is a cu-julienne density point
of As.
    To ﬁnish the proof, we show that every x0 2 W s.p; 1/ is a cu-julienne density
point of As. Then by Proposition 5.5, every x0 2 W s.p; 1/ is a Lebesgue density
point of As. The Lebesgue density points of As, and hence of A, are therefore
Ws-saturated.
     Let hs W Wcu.x/ ! Wcu.x0/ be the holonomy map induced by the stable
        b      b    cufoliation Ws. The sequence hs.OJn .x//  Wcu.x0/ nests at x0. Transverse absolute                       b
continuity of hs with bounded Jacobians implies that
   lim Omcu.As W OJncu .x// D 1 ”    lim Omcu.hs.As/ W hs.OJncu .x/// D 1: n!1                   n!1
Since As is s-saturated, we then have
     lim Omcu.As W OJncu .x// D 1 ”    lim Omcu.As W hs.OJncu .x/// D 1:   n!1                   n!1
Since we are assuming that x is a cu-julienne density point of As, we thus have
                       lim Omcu.As W hs.OJncu .x/// D 1:              n!1
Working inside of Wcu.x0/, we will apply Lemma 2.1 to the sequences hs.OJncu .x//          b
and OJncu .x0/, which both nest at x0. Proposition 5.3 implies that these sequences
are internested. Proposition 5.4 implies that OJncu .x0/ is regular with respect to the
induced Riemannian measure Omcu on Wcu.x0/. Lemma 2.1 now tells us that                     b
     lim Omcu.As W hs.OJncu .x/// D 1 ”    lim Omcu.As W OJncu .x0// D 1;  n!1                      n!1
and so x0 is a cu-julienne density point of As. It follows from Proposition 5.5 that
x0 is a Lebesgue density point of As, and thus of A.

                           6. Julienne quasiconformality

   We adapt the proof of [PS00, Th. 4.4] to prove Proposition 5.3. It will sufﬁce
to show that k can be chosen so that
(15)               hs.OJncu .x//    OJncu k.x0/   for all n   k,
whenever x and x0 satisfy the hypotheses of the proposition. The hypotheses of
the proposition treat x and x0 symmetrically, so we can then reverse their roles to
                                 for all n   k, where hs W Wcu                                                                                              is the                                                                loc.x0/ ! Wcu.x/obtain hs.OJncu .x0//    OJncu k.x/                              b                                     b
holonomy induced also by the stable foliation Ws. Since hs and hs are inverses,
                                          for all n   k.we then obtain OJncu .x0/   hs.J ncu k.x//


=== PDF PAGE 30 OF 42 ===
478                      KEITH BURNS and AMIE WILKINSON

    To prove that k can be chosen so that (15) holds, we need two lemmas.

   LEMMA 6.1. There exists a positive integer k1 such that
                                Ohs.bBc                               n.x//   bBcn k1.x0/
for all x; x0 2 Ws.p/ and all n   k1, where Ohs W Wcloc.x/ ! Wc.x0/ is the local Ws                          b     b         b
holonomy.
     Proof. Proposition 3.1 implies Ohs is L-Lipschitz for some L   1. Hence the
image of Wc.x;  n/ under Ohs is contained in Wc.x0; L n/   Wc.x0;  n k1/ for      b                   b        b
any k1 large enough so that   k1 > L.

   LEMMA 6.2. There exists a positive integer k2 such that the following holds
for every integer n   k2. Suppose q 2 Sn,  q0 2 Sn  k1, and q0 2 Ws.q/. Let                                      b
y 2 OJnu .q/, and let y0 be the image of y under Ws holonomy from Wculoc.q/ to                                       b
Wcu.q0/.        Then y0 2 OJ                                for some z0 2 Wc.q0;                                                            k2/.                   n                   u k2.z0/b                         b      n
    Remark. Note that two types of holonomy maps appear in Lemma 6.2: the
point q0 is the image of q under Ws holonomy between Wculoc.x/ and Wcu.x0/,                   b            b      b
whereas y0 is the image of y under Ws holonomy.

     Proof of Lemma 6.2. Let z0 be the unique point in Wu.y0/ \ Wc.q0/.  It
                                 b     b
is not hard to see that z0j 2 N for j D 0; : : : ; n  1 and that z0n is the unique
point in Wu.y0n/ \ Wc.q0n/.  It will sufﬁce to prove that d.y0n; z0n/ D O. n/ and     b     b
d.q0; z0/ D O. n/.
   We have d.qn; yn/    n because y 2 f n.bWu.qn;  n//. By Proposition 3.1
(iii.a), we also have d.qn; q0n/ D O. n/ and d.yn; y0n/ D O. n/, since d.q; q0/ and
d.y; y0/ are both O.1/. Note that qn and z0n are respectively the images of yn and
y0n under Wu-holonomy between Wcsloc.yn/ and Wcs.qn/. Since  <   , uniform      b            b       b
transversality of the foliations Wu and Wcs implies that
                 b    b
             d.y0n; z0n/ D O.maxfd.qn; yn/; d.yn; y0n/g/ D O. n/:
   We next show that d.q0; z0/ D O. n/. By the triangle inequality,
          d.q0n; z0n/   d.q0n; qn/ C d.qn; yn/ C d.yn; y0n/ C d.y0n; z0n/:
All four of the quantities on the right side are easily seen to be O. n/. Since q0n
and z0n lie in the same Wc-leaf at distance O. n/, Proposition 3.1 now implies
             b1that d.q0; z0/ D O.. n/                             n/. But  and   were chosen so that <      . Hence
. n/ 1 n <  n and d.q0; z0/ D O. n/, as desired.

     Proof of Proposition 5.3. As noted above,  it sufﬁces to prove the inclu-
                                                     by Lemma 6.1.sion (15). For q 2 bBcn.x/, let q0 D Ohs.q/.                                  Then q0 2 bBcn k1.p0/


=== PDF PAGE 31 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           479


                                                                         y′
                                                     Ws                                                                 bWu              y
                                                                           z′                     bWu

                           p       Sn               q                                                                             q′                                                                        bWs
                                bBcn(p)                                           n           bWc                                        f




                           yn

                              O(νn)          y′n                 O(τn)
                                                        z′n                     qn

                         O(νn)            q′n


               Figure 3. Picture for the proof of Lemma 6.2.



Hence q; q0 2 Sn  k1, and we can apply Lemma 6.2 to obtain

                        hs.OJncu .x// [ OJnu k2.z/;
                              z2Q

where Q D S                                    For k   k2, we have               q02            bBcn  k2.q0/.                   bBcn k1.x0/

         [ OJnu k2.z/ [ OJnu k.z/:
                   z2Q         z2Q


                                                       This latter inclusionIt therefore sufﬁces to ﬁnd k   k2 such that Q   bBcn k.x0/.
holds if  n k1 C  n k2    n k, which is obviously true for all n   k, if k is
sufﬁciently large.


=== PDF PAGE 32 OF 42 ===
480                      KEITH BURNS and AMIE WILKINSON

                                7. Julienne measure

   We next prove Proposition 5.4. Continuity of Wu implies that there exists a
                            b
C1  1 such that
                  Omu.bWu.qn; n//   C1;   for all q; q0                                       2 Sn.(16)      C1 1
                                  n; n//                  Omu.bWu.q0
    Let Es; Ec, and Eu be the tangent distributions to the leaves of Ws; Wc,
     b b    b                          b b
and Wu, respectively. They are H¨older continuous by Proposition 3.1, part (vi).
   b                                                                                 jFurthermore, the restrictions of these distributions to Tn are invariant under Tf
for j D         1; : : : n.           We                    next                        observe                                     that                                      the Jacobian Jac.Tf njbEu/ is nearly constantwhen       restricted                 to                  the                        set                                     precisely:                       Tn. More

   LEMMA 7.1. There exists a C2  1 such that

              Jac.Tf
                                            for all n  1 and all y; y0 2 Tn.     C2 1        njbEu/.y/   C2
              Jac.Tf njbEu/.y0/
     Proof. By the chain rule, these inequalities follow from Lemma 4.4 with
˛ D Jac.Tf jbEu/.
    Let q 2 Sn, and let X    OJnu .q/ be a measurable set (such as OJnu .q/ itself).
         D R          n.X//Then     Omu.f               X Jac.Tf njbEu/.x/ d Omu.x/. From this and Lemma 7.1 wethen obtain         more                  estimates:

   LEMMA 7.2. There exists C3 > 0 such that, for all n   0, any q; q0 2 Sn, and
any measurable sets X    OJnu .q/ and X0    OJnu .q0/, we have
                  1 Omu.f n.X//   Omu.X/      Omu.f n.X//
         C3                      C3                   Omu.f n.X0//   Omu.X0/      Omu.f n.X0//:
     Recall that f n.OJnu .q// D Wu.qn;  n/, for q 2 Sn. The ﬁrst conclusion of                  b
Proposition 5.4 now follows from (16) and Lemma 7.2 with X D OJnu .q/ and X0 D
OJnu .q0/.
    The second conclusion is proved similarly.
   We next show that there exists a ı > 0 such that
         Omu.OJnC1.q//u               ı Omu.OJnu .q//   ı   for all n  0 and all q 2 Sn.
To prove this, we will apply Lemma 7.2 with q D q0, X D OJnC1.q/,u      and X0 D
OJnu .q/. This gives
                   Omu.OJnC1.q//u             1 Omu.f n.OJnC1.q///u
                       u                      C3           u          :                    Omu.OJn .q//         Omu.f n.OJn .q///


=== PDF PAGE 33 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           481
But f n.OJnC1.q//u   D f 1.bWu.qnC1; nC1// and f n.OJnu .q// D Wu.qn;b        n/, and
hence
             Omu.f n.OJnC1.q///u          Omu.f   Wu.qnC1; nC1///                 D      1.b                         :
                                                       n//              Omu.f n.OJnu .q///        Omu.bWu.qn;
This ratio is uniformly bounded below away from 0, since f  1 is a diffeomorphism,
the leaves of Wu are uniformly smooth, and the ratio nC1= n D .pn/ is uniformly       b
bounded below away from 0.
    To prove the ﬁnal claim, we begin by observing that, considered as a subset
of Wcu.x/,                                     n.x/ with Wu-ﬁbers                                                                             OJnu .q/. We have            the set OJncu .x/ ﬁbers over bBc  b                             b
just proved that these ﬁbers are c-uniform. Since  nC1= n D  .pn/ is uniformly
bounded away from 0, the ratio

                                          nC1/                         Omc.bBcnC1.x// D Wc.x;b
                           Omc.bBcn.x//     Wc.x;b      n/
is bounded away from 0, uniformly in x and n. Thus the sequence of bases bBcn.x/
of OJncu .x/ is regular in the induced Riemannian volume Omc. Proposition 3.1, part
7. implies that Wu C 1 subfoliates Wcu; in particular, considered as a subfoliation
        b          b
of Wcu.x/, Wu is absolutely continuous with bounded Jacobians. Proposition 2.4
  b       b
implies         that the              sequence OJncu .x/ is regular, with respect to the induced Riemannian
measure Omcu. This proves the ﬁnal claim.

                                 8. Julienne density

   We now come to the proof of Proposition 5.5. We must show that if a mea-
surable set X is both Ws-saturated and essentially Wu-saturated, then a point
x 2 Ws.p; 1/ is a Lebesgue density point of X if and only if
                         lim Omcu.X W OJncu .x// D 1:               n!1

   We will establish the following chain of equivalences:
   x is a Lebesgue density point of X ” lim m.X W Bn.x// D 1
                          n!1
  ” lim m.X W Cn.x// D 1 ” lim m.X W Dn.x// D 1
        n!1              n!1
  ” lim m.X W En.x// D 1 ” lim m.X W Fn.x// D 1
        n!1              n!1
  ” lim m.X W Gn.x// D 1 ” lim Omcu.X W OJncu .x// D 1:        n!1              n!1

Before deﬁning the sets Bn.x/ through Gn.x/, we outline the general scheme of
the proof. After verifying the ﬁrst equivalence, we prove that Bn.x/ is regular,
and that Bn.x/ and Cn.x/ are internested. The second equivalence then follows
from Lemma 2.1. The sets Cn.x/ and Dn.x/ both ﬁber over the same base in Wcs,                                          b


=== PDF PAGE 34 OF 42 ===
482                      KEITH BURNS and AMIE WILKINSON

with c-uniform ﬁbers in Wu, so the third equivalence follows from Proposition 2.5.
We prove that the sets Dn.x/, En.x/, Fn.x/, and Gn.x/ are all internested, and
that Gn.x/ is a regular sequence. Equivalences 4–6 then follow from Lemma 2.1.
Finally, Gn.x/ ﬁbers over OJncu .x/, with c-uniform Ws-ﬁbers, and so the ﬁnal equiv-
alence follows from Proposition 2.7. This ﬁnal step uses Ws-saturation of X.
    The sets Bn.x/ through Gn.x/ are deﬁned as follows. The set Bn.x/ WD
B.x;  n/ is a Riemannian ball in M. The sets Cn.x/, Dn.x/, and En.x/ will ﬁber
over the same base Dcsn .x/, where
                   Dcsn .x/ D  [     bBcn.x0/:
                              x02bWs.x; n/
Proposition 3.1, part (iv), implies that Dcsn .x/ is contained in the C 1 submanifold
Wcs.x/; the sequences Dcsn .x/ and Wcs.x;  n/ are internested. Letb                   b
       Cn.x/ D [  Wu.q;  n/  and  Dn.x/ D [  J nu .q/:
                q2Dcsn .x/                         q2Dcsn .x/

The set En.x/ is nearly identical to Dn.x/, with the crucial difference that the
Jnu -ﬁbers are replaced with OJnu -ﬁbers:
    En.x/ D [    OJnu .q/ D  [     OJncu .x0/ D  [     OJncu .x0/:
                                              n/           x02Ws.x; n/             q2Dcsn .x/        x02bWs.x;
The rightmost equality follows from the fact that Ws.x;  n/ D Ws.x;  n/ for all                            b
x 2 Ws.p; 1/ (see Proposition 3.1, part (v).)
   We deﬁne Fn.x/ to be the foliation product of OJncu .x/ and Ws.x;  n/:
                 Fn.x/ D  [    Ws.q/ \ Wcu.q0/:                            b
                           q2 OJncu .x/
                              q02Ws.x; n/

This deﬁnition makes sense since the foliations Wcu and Ws are transverse. Finally,
                         b
let
                    Gn.x/ D [  Ws.q;  n/:
                              q2 OJncu .x/
It is in the transition from Dn to En that the exchange between the measure-
theoretically useful foliation Wu and the geometrically useful foliation Wu takes
                                       b
place. The deﬁnition of Fn is where we ﬁrst use the foliation Ws.
     Figure 4 is a schematic illustration of the relationship between the sets En.x/,
Fn.x/, and Gn.x/. All three sets contain OJncu .x/ and Ws.x;  n/. The set En.x/
ﬁbers over Ws.x;  n/ with ﬁbers of the form OJncu .  /. The set Gn.x/ ﬁbers over
OJncu .x/ with ﬁbers of the form Ws.  ;  n/. The foliation product Fn.x/ of OJncu .x/
and Ws.x;  n/ is, in some sense, intermediate between En.x/ and Gn.x/.


=== PDF PAGE 35 OF 42 ===
          ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           483


                                            ˆJcun (x)


                                             Gn(x)      En(x)

                                bWs(x, σn)
             x                                   x





                          Ws


                         Fn(x)





                               x
                  bWcu





      Figure 4. Comparison between En.x/, Fn.x/ and Gn.x/.



We now prove these equivalences, following the outline described above.


=== PDF PAGE 36 OF 42 ===
484                      KEITH BURNS and AMIE WILKINSON

      First, recall that Bn.x/ is a round ball about x of radius  n. The forward im-
plication in the ﬁrst equivalence is obvious from the deﬁnition of Bn.x/. The back-
ward implication follows from this deﬁnition and the fact that the ratio  nC1= n D
 .pn/ of successive radii is less than 1, and is bounded away from both 0 and 1
independently of n. From this we also see that Bn.x/ is regular.
    The set Cn.x/ ﬁbers over Dcsn .x/, with ﬁber Wu.x0;  n/ over x0 2 Dcsn .x/.
The sequence Dcsn .x/ internests with the sequence of disks Wcs.x;  n/, by conti-                                 b
nuity and transversality of the foliations Wc and Ws. Continuity and transversality
                      b    b
of the foliations Wu and Wcs then imply that Cn.x/ and Bn.x/ are internested.              b
    To prove the equivalence

           lim m.X W Cn.x// D 1 ”    lim m.X W Dn.x// D 1;
      n!1                 n!1

we note that Cn.x/ and Dn.x/ both ﬁber over Dcsn .x/, with Wu-ﬁbers. Since X
is essentially Wu-saturated, Proposition 2.5 implies that it sufﬁces to show that the
ﬁbers of Cn.x/ and Dn.x/ are both c-uniform. The ﬁbers of Cn.x/ are easily seen
to be uniform, because they are all comparable to balls in Wu of ﬁxed radius  n.
The ﬁbers of Dn.x/ are the unstable juliennes J nu .x0/ for x0 2 Dcsn .x/. Uniformity
of these ﬁbers follows from Proposition 5.4.

   LEMMA 8.1. The sequences Dn.x/ and En.x/ are internested.

     Proof. Recall that

        Dn.x/ D [  Jnu .x/  and  En.x/ D [    OJnu .x/:
                  q2Dcsn .x/                      q2Dcsn .x/

Internesting of the sequences Dn.x/ and En.x/ means that there is a k  0 such
that Dn.x/  En k.x/ and En.x/  Dn k.x/. for all n   k. We will show that
there is a k for which the ﬁrst inclusion holds. Reversing the roles of Wu and Wu
                                           b
in the proof gives the second inclusion.
    Suppose now that y 2 Dn.x/. Then y 2 Jnu .q/ D f  n.Wu.qn; n// for some
q 2 Dcsn .x/; in particular,

(17)                       d.yn; qn/ D O. n/:

Let Oq be the unique point of intersection of Wu.y/ with Wcs.x/. We will show
                         b      b
that y 2 En k.x/ for some k that is independent of n. To do this, it sufﬁces to
show that Oq 2 Dcs              n k.x/ and y 2 OJnu k.Oq/ D                                f  .n k/.bWu.Oqn k; n  k//.
     In order to prove that Oq 2 Dcs         it will sufﬁce to show that                             n k.x/

(18)                          d.q; Oq/ D o. n/


=== PDF PAGE 37 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           485

(in fact, O. n/ would sufﬁce, but the argument gives o. n/). In order to prove that
y 2 OJ u          it will sufﬁce to show that     n  k.Oq/,
(19)                       d.yn; Oqn/ D O. n/:
     Equation (18) follows easily from (19). Since yn and Oqn lie in the same Wu                                           b
leaf, Proposition 3.1 and (19) imply that

(20)                   d.y; Oq/ D O.O n n/ D o. n/;

since O <   . Similarly, Proposition 3.1 and (17) imply that

(21)                        d.y; q/ D o. n/:

Applying the triangle inequality to (20) and (21) gives (18).
       It remains to prove (19). Recall from the construction of the fake foliations
in Proposition 3.1 that, at any point z in the neighborhood N of the orbit of p in
which the fake foliations are deﬁned, the tangent space Tz Wu.z/ lies in the "-cone                               b
about TzWu.z/ D Eu.z/. Also, the angle between Tz Wcs.z/ and either Tz Wu.z/                             b           b
or TzWu.z/ is uniformly bounded away from 0. Note Oqn is the unique point in
Wu.yn/ \ Wcs.xn/ and qn is the unique point in Wu.yn/ \ Wcs.xn/; combiningb     b                          b
this with (17) gives

                    d.yn; Oqn/ D O.d.yn; qn// D O. n/:

   LEMMA 8.2. En.x/ and Fn.x/ are internested, as are Fn.x/ and Gn.x/.

     Proof. The sets En.x/ and Fn.x/ both ﬁber over the same base Ws.x;  n/.
The ﬁbers of En.x/ are the cu-juliennes OJncu .x0/ for x0 2 Ws.x;  n/. bThe ﬁbers                                b
of Fn.x/ are images of OJncu .x/ under Ws-holonomy from Wcu.x/ to Wcu.x0/ for                                b     b
x0 2 Ws.x;  n/. It follows immediately from Proposition 5.3 that the sequences   b
En.x/ and Fn.x/ are internested.
    To see that Fn.x/ and Gn.x/ are internested, suppose that q0 lies in the bound-
ary of the ﬁber of Fn.x/ that lies in Ws.q/ for some q 2 OJncu .x/. Then q0 2 OJncu .x0/
for a point x0 that lies in the boundary of Ws.x;  n/. The diameters of OJncu .x/ and
OJncu .x0/ are both O. n/, and d.x; x0/ D  n. Hence, if k is large enough, we will
have  nCk   d.q; q0/    n k. Thus all points on the boundary of the ﬁber of
Fn.x/ in Wsloc.q/ lie outside Ws.q; nCk/ and inside Ws.q;  n  k/.

   We now know that any two of Dn.x/, En.x/, Fn.x/, and Gn.x/ are in-
ternested. As discussed above, to prove the fourth through sixth equivalences, it
now sufﬁces to show this:

   LEMMA 8.3. The sequence Gn.x/ is regular for each x 2 Ws.p; 1/.


=== PDF PAGE 38 OF 42 ===
486                      KEITH BURNS and AMIE WILKINSON

     Proof. The set
                    Gn.x/ D [  Ws.q;  n/
                               q2OJncu .x/
ﬁbers over OJncu .x/, with Ws-ﬁbers Ws.q;  n/. Since Ws is absolutely continu-
ous, Proposition 2.4 implies that regularity of Gn.x/ follows from regularity of
the base sequence and ﬁber sequence. Proposition 5.4 implies that the sequence
OJncu .x/ is regular in the induced measure Omcu. As we remarked above, the ratio
 nC1= n D  .pn/ is uniformly bounded below away from 0. Hence the ratio
ms.Ws.q; nC1//=ms.Ws.q;  n// is bounded away 0, uniformly in x; q, and n.
The regularity of Gn.x/ now follows from Proposition 2.4.
    To prove the ﬁnal equivalence, we use the fact that Gn.x/ ﬁbers over OJncu .x/
with c-uniform ﬁbers and apply Proposition 2.7. Here we use the fact that X is
Ws-saturated. This completes the proof of Proposition 5.5.

                                  9. Open questions

   We do not know whether the bunching assumption can be dropped in Theorem
0.1. It seems clear that a signiﬁcant new idea would be required to remove this
hypothesis. Indeed, even if we impose stringent assumptions on the regularity of
the center, center-stable, and center-unstable foliations, as Brin and Pesin do in their
1974 work, every known ergodicity result requires the center bunching hypothesis
(5). Thus a ﬁrst step in answering this question might be to answer the following
question:

     Question. Suppose that f  is C 2, volume-preserving, and partially hyperbolic.
If Ec is absolutely continuous (Lipschitz, smooth, and so on) and f  is (essentially)
accessible, is f then ergodic?
    As a concrete example, consider a diffeomorphism f   W T2  T2 ! T2  T2
of the form f .x; y/ D .A.x/; g .y//, where A W T2 ! T2 is the linear Anosov
diffeomorphism given by
                                 2 1  2
                   A D
                                 1 1
and g   W T2 ! T2 is a standard map of the form

           g .z; w/ D z C w; w C  sin.2  .z C w//=.2  / :

A straightforward calculation shows that there is an interval ƒ  R containing
.  4; 4/ such that, if  2 ƒ, then f   is partially hyperbolic with respect to the
standard (ﬂat) metric on T2  T2. It is also not difﬁcult to show, by examining the
spectrum of Tf  at the ﬁxed point .0; 0/, that f  is center bunched if and only if
 2 .  1; 1/.


=== PDF PAGE 39 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           487

     This example appears in [SW00b], where it is shown that there is a function
' W T2 ! T2 with '.0/ D 0 and an interval E D .0; 0/ such that, for all .  ;  / 2
ƒ E, the map f  ; .x; y/D.A.x/; g .y/C '.x// is both partially hyperbolic and
stably accessible. Also, f  ;  is center bunched if and only if .  ;  / 2 .  1; 1/  E.
For all .  ;  / 2 ƒ  E, the center bundle Ec is tangent to the ﬁbers fxg  T2 and
is C 1. The foliations Wcu and Wcs are also C 1.
    Theorem 0.1 implies that f  ;  is stably ergodic for all .  ;  / 2 .  1; 1/  E.
We do not however know whether f  ;  is ergodic for even a single value of .  ;  / 2
.ƒ n .  1; 1//  E.
   We also do not know whether or not the center bunching inequalities (5)
imply that a partially hyperbolic diffeomorphism is dynamically coherent. The
simplest non-dynamically coherent examples are linear Anosov diffeomorphisms
of nilmanifolds, viewed as partially hyperbolic by grouping the weaker stable and
unstable directions together to form a center distribution.  It was ﬁrst observed
in [Wil98] that this construction can provide examples that are not dynamically
coherent. However, these examples were not center bunched. Hammerlindl has
recently given examples that are center bunched but not dynamically coherent.
     In [BW06a], we discuss a family of examples of this type. They are Anosov
diffeomorphisms of a compact quotient of the product of the Heisenberg group
with itself induced by linear maps whose eigenvalues are
                 aCb    b    a     a      b      a  b;

where b  a   0 and  > 1 is the larger of the two eigenvalues of a hyperbolic
matrix in SL.2; Z/ whose eigenvalues are > 1 >   1 > 0. When a > 0, the map
can be viewed as a partially hyperbolic diffeomorphism with Eu and Es being
the left invariant one-dimensional distributions corresponding to the eigenvalues
 aCb        a     and       b, respectively. The center is the left invariant four-dimensional
distribution corresponding to the sum of the eigenspaces for the other eigenvalues.
    The linear map is chosen so that ŒEc; Ec is spanned by Eu ˚ Es. This
ensures that distribution Ec is not integrable. Consequently these examples are
not dynamically coherent.
    For these examples we have D O D   a b and D O D    b. Thus center
bunching fails, because D O D   a b      2b D   O .
    Note that center bunching only just fails in the extreme case when a D b.
Gourmelon has constructed perturbations of the a D b example that are robustly
ergodic, but neither Anosov nor dynamically coherent.  It does not seem to be
known whether or not his perturbations are center bunched.

    Acknowledgments. We thank Marcelo Viana for useful comments and for tell-
ing us about the proof of absolute continuity of stable foliations in the pointwise
partially hyperbolic setting in [AV08]. We thank Charles Pugh and Mike Shub for


=== PDF PAGE 40 OF 42 ===
488                      KEITH BURNS and AMIE WILKINSON

very useful comments. A proof of Theorem 0.1 under the additional hypothesis of
dynamical coherence appeared earlier in the unpublished preprint [BW05].

                              References

[AV08]       F. ABDENUR and M. VIANA, Flavors of partial hyperbolicity, Preprint, 2008. Avail-
              able at http://www.impa.br/opencms/pt/publicacoes/index.html

[Ano67]    D. V. ANOSOV, Geodesic ﬂows on closed Riemannian manifolds of negative curva-
               ture, Trudy Mat. Inst. Steklov. 90 (1967), 209, In Russian. MR 36 #7157 Zbl 0176.
           19101
[AS67]     D. V. ANOSOV and J. G. SINA˘I, Certain smooth ergodic systems, Uspehi Mat. Nauk
           22 (1967), 107–172, In Russian. MR 37 #370 Zbl 0177.42002
[BP74]     M. I. BRIN and J. B. PESIN, Partially hyperbolic dynamical systems, Izv. Akad. Nauk
          SSSR Ser. Mat. 38 (1974), 170–212, In Russian, translated in Math. USSR Izv. 8
              (1974), 177–218. MR 49 #8058 Zbl 0304.58017

[BW05]     K. BURNS and A. WILKINSON, Better center bunching, Preprint, 2005. Available at
             http://www.math.northwestern.edu/~wilkinso/papers/papers.html
[BW06a]                     , Dynamical coherence and center bunching, Preprint, 2006. Available at http://
             www.math.northwestern.edu/~wilkinso/papers/papers.html
[BW06b]                    , A note on unstable holonomy between centers, Preprint, 2006. Available at
             http://www.math.northwestern.edu/~wilkinso/papers/papers.html
[BPSW01]  K. BURNS, C. PUGH, M. SHUB, and A. WILKINSON, Recent results about stable
               ergodicity, in Smooth Ergodic Theory and its Applications (Seattle, 1999) (A. KATOK,
            R. DE LA LLAVE, Y. PESIN, and H. WEISS, eds.), Proc. Sympos. Pure Math. 69, Amer.
            Math. Soc., 2001, pp. 327–366. MR 2002m:37042 Zbl 1012.37019

[Dol04]     D. DOLGOPYAT, On differentiability of SRB states for partially hyperbolic systems,
               Invent. Math. 155 (2004), 389–449. MR 2005h:37070 Zbl 1059.37021
[GPS94]    M. GRAYSON, C. PUGH, and M. SHUB, Stably ergodic diffeomorphisms, Ann. of
            Math. 140 (1994), 295–329. MR 95g:58128 Zbl 0824.58032
[Has94]     B. HASSELBLATT, Periodic bunching and invariant foliations, Math. Res. Lett. 1
              (1994), 597–600. MR 95h:58097 Zbl 0845.58043
[HPS77]    M. W. HIRSCH, C. C. PUGH, and M. SHUB, Invariant manifolds, Lecture Notes in
            Math. 583, Springer-Verlag, New York, 1977. MR 58 #18595 Zbl 0355.58009

[Hop39]     E. HOPF, Statistik der geodätischen Linien in Mannigfaltigkeiten negativer Krüm-
           mung, Ber. Verh. Sächs. Akad. Wiss. Leipzig 91 (1939), 261–304. MR 1,243a Zbl
            0024.08003 JFM 65.1413.02
[KH95]     A. KATOK and B. HASSELBLATT, Introduction to the Modern Theory of Dynamical
             Systems, Encycl. Math. Appl. 54, Cambridge Univ. Press, Cambridge, 1995. MR 96c:
           58055 Zbl 0878.58020
[Mil97]       J. MILNOR, Fubini foiled: Katok’s paradoxical example in measure theory, Math. In-
               telligencer 19 (1997), 30–32. MR 98g:28005 Zbl 0883.28004
[Pes04]      Y. B. PESIN, Lectures on Partial Hyperbolicity and Stable Ergodicity, Zur. Lect. Adv.
             Math., European Mathematical Society, Zürich, 2004. MR 2005j:37039 Zbl 1098.
           37024
[PS72]      C. PUGH and M. SHUB, Ergodicity of Anosov actions, Invent. Math. 15 (1972), 1–23.
       MR 45 #4456 Zbl 0236.58007


=== PDF PAGE 41 OF 42 ===
              ON THE ERGODICITY OF PARTIALLY HYPERBOLIC SYSTEMS           489


[PS97]      C. PUGH and M. SHUB, Stably ergodic dynamical systems and partial hyperbolicity,
                 J. Complexity 13 (1997), 125–179. MR 98e:58110 Zbl 0883.58025

[PS00]                        , Stable ergodicity and julienne quasi-conformality, J. Eur. Math. Soc. (JEMS)
           2 (2000), 1–52. MR 2003j:37046 Zbl 0964.37017

[PSW97]    C. PUGH, M. SHUB, and A. WILKINSON, Hölder foliations, Duke Math. J. 86 (1997),
            517–546. MR 97m:58155 Zbl 0877.58045

[PSW00]                    , Correction to [PSW97], Duke Math. J. 105 (2000), 105–106. MR 2001h:
           37057

[RHRHU08] F. RODRÍGUEZ HERTZ, M. A. RODRÍGUEZ HERTZ, and R. URES, Accessibility and
              stable ergodicity for parially hyperbolic diffeomorphisms with 1D-center bundle, In-
               vent. Math. 172 (2008), 353–381. MR 2390288 Zbl 1136.37020

[SSS92]       J. SCHMELING and R. SIEGMUND-SCHULTZE, Hölder continuity of the holonomy
           maps for hyperbolic basic sets, I, in Ergodic Theory and Related Topics, III (Güstrow,
            1990) (U. KRENGEL, K. RICHTER, and V. WARSTAT, eds.), Lecture Notes in Math.
             1514, Springer-Verlag, New York, 1992, pp. 174–191. MR 93j:58104 Zbl 0766.58043

[SW00a]    M. SHUB and A. WILKINSON, Pathological foliations and removable zero exponents,
               Invent. Math. 139 (2000), 495–508. MR 2001c:37030 Zbl 0976.37013

[SW00b]                    , Stably ergodic approximation: Two examples, Ergodic Theory Dynam. Sys-
            tems 20 (2000), 875–893. MR 2001d:37032 Zbl 0970.37022

[Wil98]     A. WILKINSON, Stable ergodicity of the time-one map of a geodesic ﬂow, Ergodic
            Theory Dynamical Systems 18 (1998), 1545–1587. MR 99m:58129 Zbl 0915.58079

                        (Received October 11, 2005)
                         (Revised March 31, 2006)


E-mail address: burns@math.northwestern.edu
DEPARTMENT OF MATHEMATICS, NORTHWESTERN UNIVERSITY, 2033 SHERIDAN ROAD,
EVANSTON IL 60208-2730, UNITED STATES
http://www.math.northwestern.edu/people/facultyProﬁles/keith.burns.html

E-mail address: wilkinso@math.northwestern.edu
DEPARTMENT OF MATHEMATICS, NORTHWESTERN UNIVERSITY, 2033 SHERIDAN ROAD,
EVANSTON, IL 60208-2730, UNITED STATES
http://www.math.northwestern.edu/~wilkinso/


=== PDF PAGE 42 OF 42 ===


===== SOURCE: holonomy-note-2005.txt =====


=== PDF PAGE 1 OF 11 ===
 A note on stable holonomy between centers

                Preliminary version

             Keith Burns and Amie Wilkinson

                     June 11, 2005



   There is a slight disparity in smooth ergodic theory, between Pesin the-
ory and the Pugh-Shub partially hyperbolic theory. Both theories assume
a weakened form of hyperbolicity and conclude ergodicity. While the two
theories have considerable overlap, Pesin theory assumes a C1+δ hypothesis,
whereas the Pugh-Shub stable ergodicity theorem assumes a C2 hypothesis.
The purpose of this paper is to close the gap between C1+δ and C2.
   Let f  : M →M be a partially hyperbolic diﬀeomorphism of a compact
manifold M. Partially hyperbolic means the following. There is a nontrivial
splitting of the tangent bundle, TM = Eu ⊕Ec ⊕Es, that is invariant under
the derivative map Tf. Further, there is a Riemannian metric for which we
can choose continuous positive functions ν, ˆν, and ˆγ with

                    ν, ˆν < 1  and   ν < γ ≤ˆγ−1 < ˆν−1                   (1)

such that, for a unit vector v ∈TpM,

                    ∥Tfv∥≤ν(p),           if v ∈Es(p),              (2)
                 γ(p) ≤∥Tfv∥≤ˆγ(p)−1,       if v ∈Ec(p),              (3)
                  ˆν(p)−1 ≤∥Tfv∥,                     if v ∈Eu(p).              (4)

  A partially hyperbolic diﬀeomorphism is dynamically coherent  if there
are foliations Wcs and Wcu tangent to Ec ⊕Es and Ec ⊕Eu respectively.
In this case there is also a foliation Wc tangent to Ec whose leaves are the
intersections of the leaves of Wcs and Wcu. Each leaf of Wcs is foliated by
leaves of Wc and Ws; leaves of Wcu have the analogous property.


                                 1


=== PDF PAGE 2 OF 11 ===
Theorem 0.1 ([PSW], Theorem B) Let f  : M →M be a C2, partially
hyperbolic, dynamically coherent diﬀeomorphism of a compact manifold M
satisfying
                              ν < γˆγ.
Then the stable distribution is a uniformly C1 bundle when restricted to any
leaf of the center-stable foliation.  Consequently, the stable holonomy map
between center transversals is C1.

     If the assumption that f  is C2 is relaxed, then the ﬁrst conclusion of
Theorem 0.1 does not hold; that is, the stable bundle may no longer be
a C1 bundle over every leaf of the center-stable foliation.  This is easy to
see in the case where the unstable bundle is trivial. Let F  : T2 →T2 be
a linear Anosov diﬀeomorphism of the 2-torus, and let f = ϕ ◦F ◦ϕ−1,
where ϕ : T2 →T2 is a C1+δ diﬀeomorphism that is not C2. Then f is an
Anosov diﬀeomorphism, with Esf = ϕ∗(EsF) and Euf = ϕ∗(EuF). We regard f
as a partially hyperbolic diﬀeomorphism, with stable bundle Esf and center
bundle Euf . Then the center-stable foliation has one leaf, the whole 2-torus,
and the stable bundle Esf = ϕ∗(EsF) is merely H¨older continuous, since Tϕ is
H¨older-continuous, but not C1. Note that the hypothesis ν < γˆγ is satisﬁed
by this example, since the center bundle of f is 1-dimensional.
   While the ﬁrst conclusion of Theorem 0.1 does not hold for this example,
the second one does; the stable holonomy for f between center transversals is
C1, and in fact, C1+δ. This is because the stable holonomy for f is conjugate
via ϕ to the stable holonomy for F.
   Note also that, even under the hypothesis that f  is C2, the techniques
behind the proof of Theorem 0.1 do not give a C1+β condition on the stable
holonomy; for this, a C2+δ hypothesis on f  is required. Our main result
shows that we may relax the C2 assumption to C1+δ and still obtain that
the stable holonomy for f between center transversals is C1. The price of
relaxing the diﬀerentiability of f is that we require a more stringent bunching
condition, a condition that is nonetheless satisﬁed by many examples. As an
added beneﬁt, we obtain that the holonomy is C1+β, for some β > 0.

Theorem 0.2 Let f : M →M be a C1+δ, partially hyperbolic, dynamically
coherent diﬀeomorphism of a compact manifold M satisfying
                                 νθ < γˆγ,

where θ ∈(0, δ) is a H¨older exponent for Ec. Then there exists β > 0 such
that the stable holonomy map between center transversals is C1+β.


                                 2


=== PDF PAGE 3 OF 11 ===
   As far as we know, this result is new, although there are related results
for codimension-1 hyperbolic systems [PR], in the nonuniform setting when
Ec is a hyperbolic Lyapunov subspace in an Oseledec decomposition [BPS],
and for Lie group cocycles [PW] .
   For any partially hyperbolic diﬀeomorphism, the stable and unstable sub-
bundles Es and Eu are tangent to foliations, which we denote by Ws and Wu
respectively [BP]. These foliations induce an equivalence relation on M: we
say that p ∼us q if there is a sequence of points p = p0, . . . , pk = q such that
any two consecutive points in the sequence lie in the same Ws-leaf or the
same Wu-leaf. A partially hyperbolic diﬀeomorphism has the accessibility
property if there is only one ∼us-class, i.e. if p ∼us q for any p, q ∈M. It has
the essential accessibility property if a set that is measurable (with respect
to the volume) and is a union of ∼us-classes must have 0 or full measure. In
[BW], we use Theorem 0.2 to prove the following:

Theorem 0.3 Let f be C1+δ, volume preserving, and partially hyperbolic.
Suppose that f satisﬁes the strong center bunching condition:

                       max{ν, ˆν}θ < γˆγ,                             (5)

where θ ∈(0, δ) satisﬁes the inequalities:

                   νγ−1 < µθ,      ˆνˆγ−1 < ˆµθ.                        (6)

    If f is essentially accessible, then f is ergodic, and in fact has the Kol-
mogorov property.

   This has the corollary:

Corollary 0.4 Let f  : M →M be a C1+δ, volume-preserving, partially
hyperbolic, dynamically coherent diﬀeomorphism with 1-dimensional center
bundle. If f is accessible, then f is ergodic.

   Note that Theorem 0.2 can equally be applied to the stable foliation of
C1+δ Anosov diﬀeomorphism by regarding the unstable bundle as a center
bundle. By doing so, we obtain, for example, the following:

Corollary 0.5 Let f be a C1+δ conformal Anosov diﬀeomorphism. Then for
every β < δ, there exists a neighborhood U of f in Diﬀ1+δ(M), such that, for
every g ∈U, the stable and unstable holonomies for g are C1+β.


                                 3


=== PDF PAGE 4 OF 11 ===
1  Proof of Theorem 0.2

The proof follows similar lines to the proof of absolute continuity of the
stable foliation for partially hyperbolic systems. The main diﬀerence is that,
unlike the Jacobian map x 7→Jacxf, the multiplicative cocycle x 7→Txf
over f is not abelian. We exploit instead the fact that the map v 7→∥Tfv∥
is an abelian multiplicative cocycle over the projectivized tangent map f∗:
P(TM) →P(TM). This map is projective in the ﬁbers, with Lipschitz norm
(γˆγ)−1. The inequality νθ < γˆγ allows us to prove that the stable holonomy
hs has a projectivized derivative hs∗. Once this is established, a distortion
estimate using the v 7→∥Tfv∥cocycle gives us that hs is diﬀerentiable.
  We denote by hs the stable holonomy between two center leaves, omitting
reference to domain and codomain. We denote by πs a ﬁxed, uniformly C1+θ,
local approximation to hs. By this we mean that there is a constant C > 0
such that for any p, q ∈M with q ∈W s(p, 1) we have a C1+θ diﬀeomorphism

                         πsp,q : W loc(p)c  →W c(q)

such that:

   1. d(πsp,q(p), q) ≤Cd(p, q);

   2. d(Tπsp,qv, v) ≤Cd(p, q)θ, for all v ∈P(Ec(p));

   3.  if Wcloc(p) ∩Wcloc(p′)̸ = ∅, and the codomains of πsp,q and πsp′,q′ also
      intersect nontrivially, then πsp,q = πsp′,q′ on Wcloc(p) ∩Wcloc(p′).

   This can be accomplished by ﬁxing a C1+θ normal bundle to each leaf of
the center foliation Wc. Since the leaves of Wc vary continuously in the C1+θ
topology, this choice of normal bundle can be made to vary continuously from
leaf to leaf. The ﬁbers of this bundle foliate a uniform neighborhood of each
leaf. The holonomy of this local foliation between a local center leaf Wcloc(p)
and a neighboring center leaf Wcloc(q) gives the map πsp,q. By rescaling the
metric if necessary, we may assume that πsp,q is well-deﬁned on the domain
Wcloc(p), for all q ∈W s(p, 1). We will sometimes drop the reference to p and
q and write πs for short.
   Fix p, q ∈M with q ∈W s(p, 1), and let

                          hsn = f −n ◦πspn,qn ◦f n.


                                 4


=== PDF PAGE 5 OF 11 ===
Note that hsn  : Wcloc(p) →Wc(q) is well-deﬁned, for all n ≥0, since πspn,qn
extends uniquely to the domain f n(Wcloc(p)).
   The proof now proceeds in four steps alluded to above. In the ﬁrst step,
we show that hsn converges in the C0 topology to hs.  This ﬁrst step uses
only partial hyperbolicity and dynamical coherence. In the second step, we
show that the projective linear bundle maps hsn∗: P(Ec) →P(Ec) converge
in the C0 topology to a projective linear bundle map hs∗: P(Ec) →P(Ec).
In Step 3, we show that the derivatives Thsn : Ec →Ec converge in the C0
topology to a bundle isomorphism Ths : Ec →Ec. It follows that hs is C1,
with derivative Ths. Finally, we show that Ths is H¨older continuous. The
last three steps use the bunching hypothesis νθ < γˆγ.

Step 1: the sequence hsn converges to hs in the space of all contin-
uous maps from Wcloc(p) to Wc(q).

   Invariance of Ws under f implies that f −nhspn,qnf n = hs, where the stable
holonomy hspn,qn maps from f n(Wcloc(p)) to Wc(qn), sending pn to qn.
   Note that since f n contracts distances in Ws(x, 1) by a factor of νn(x),
we have that d(xn, hspn,qn(xn)) ≤νn(x), for all x ∈Wcloc(p) and n ≥1. By
property 1. of πs, we then obtain that

                    d(πs(xn), hspn,qn(xn)) ≤Cνn(x),                      (7)

for all x ∈Wcloc(p) and n ≥1.
  Now ﬁx x ∈Wcloc(p) and n ≥1. We have:

             d(hsn(x), hs(x)) =  d(f −nπs(xn), f −nhspn,qn(xn))
                 ≤  γ−n(x) d(πs(xn), hspn,qn(xn))
                 ≤  Cγ−n(x)ν−n(x).

Partial hyperbolicity implies that there exists κ < 1 such that the function
γ−nνn is uniformly bounded by κn. Hence hsn converges uniformly to hs.⋄

Step 2: the sequence hsn∗: P(TWcloc(p)) →P(TWc(q)) is Cauchy.

  We ﬁx some notation. For w ∈P(TM) and k ∈Z, let wk = f∗(w)k   ∈
P(TM). We ﬁrst prove some lemmas.



                                 5


=== PDF PAGE 6 OF 11 ===
Lemma 1.1 There exists δ > 0 such that, for every v ∈P(Ec(x)) and
w ∈P(Ec(y)):

   1.  if d(vk, wk) ≤δ, for k = 0, . . . , n −1, then

                         d(vn, wn) ≤(γˆγ)−n(xn)d(v, w)θ;

   2.  if d(v−k, w−k) ≤δ, for k = 0, . . . , n −1, then

                        d(v−n, w−n) ≤(γˆγ)−n(x)d(v, w)θ.

In particular, if d(v, w) ≤δ(γˆγ)n(x), then the conclusions in 1. hold, and if
d(v, w) ≤δ(γˆγ)n(xn), then the conclusions in 2. hold.


Proof. It suﬃces to prove the lemma in the case n = 1; the other cases are
proved inductively. In writing f∗(v) for v ∈S(TxM), we deliberately suppress
the basepoint x. For the following calculation, the basepoint is relevant, so
we shall write v = (x, ζ) and w = (y, ξ). Recall that the Lipschitz norm of f∗
restricted to the ﬁber P(Ec(x)) is γˆγ(x)−1 = (γˆγ)−1(x1), and the Lipschitz
norm of f∗−1 restricted to the ﬁber P(Ec(x)) is γˆγ(x−1)−1 = (γˆγ)−1(x). We
estimate:

  d(v1, w1) =  d(f∗(x, ζ), f∗(y, ξ))
       ≤  d(f∗(x, ζ), f∗(x, ξ)) + d(f∗(x, ξ), f∗(y, ξ))
       ≤  (γˆγ)−1(x1)d(ζ, ξ) + Hd(x, y)δ
       ≤  (γˆγ)−1(x1)  d(ζ, ξ)1−θ + Hd(x, y)δ−θ  sup{d(x, y), d(ζ, ξ)}θ.

If d(v, w) = sup{d(ζ, ξ), d(x, y)} is suﬃciently small, then

                        d(ζ, ξ)1−θ + Hd(x, y)δ−θ < 1,

and we have
                         d(v1, w1) ≤(γˆγ)−1(x1)d(v, w)

This completes the proof of 1, in the case n = 1. The case 2.  is proved
similarly.⋄

   The next Lemma will also be useful in Step 3.


                                 6


=== PDF PAGE 7 OF 11 ===
Lemma 1.2 For n ≥1, and k ≤n, let

                                   v′k = f∗−n+k π∗(vn),
where v ∈P(TxWcloc(p)). Then, for n suﬃciently large, we have:
                             d(vk, v′k) ≤Cνk(x)θ,

for all k ≤n.

Proof. Property 2. of πs and uniform contraction of Ws(p, 1) by ν together
imply that

                      d(vn, v′n) =  d(vn, πs∗(vn))
                  ≤  Cd(xn, hspn,qn(xn))θ
                  ≤  Cνn(x)θ.

Let δ > 0 be given by Lemma 1.1. The bunching hypothesis νθ < γˆγ implies
that if n is suﬃciently large, we have:

                           d(vn, v′n) ≤δ(γˆγ)n(x).

Assume n is this large.
  We prove Lemma 1.2 by backward induction on k. We have shown that

                    d(vn, v′n) ≤min{Cνn(x)θ, δ(γˆγ)n(x)}.

Lemma 1.1 now implies that for all 1 ≤k ≤n, we have

                    d(vk, v′k) =  d(f∗−n+k (vn), f∗−n+k (v′n))
                ≤  (γˆγ)k−n(x)d(vn, v′n)
                ≤  δ(γˆγ)k(x)

Suppose now that for some k we have

                             d(vk, v′k) ≤Cνk(x)θ.
Then, since d(vk, v′k) ≤δ, we have
                    d(vk−1, v′k−1) =  d(f∗(vk),−1    f∗(v′−1  k))
                   ≤  (γˆγ)−1(xk)d(vk, v′k)
                   ≤  (γˆγ)−1(xk)Cνk(x)θ
                   ≤  Cνk−1(x)θ,

                                 7


=== PDF PAGE 8 OF 11 ===
since νθ < γˆγ. ⋄

   Returning to the proof of Step 2, we ﬁx v ∈P(TWcloc(p)), and n ≥1.
Then

d(hn∗(v), hn+1∗(v)) =  d(f∗πs−n  ∗(vn), f∗−n−1 πs∗f∗(vn))
           ≤  (γˆγ)−n(x) d(πs∗(vn), f∗πs−1 ∗f∗(vn))
           ≤  (γˆγ)−n(x) h d(πs∗(vn), (vn)) + d(f∗f∗(vn),−1      f∗πs−1 ∗f∗(vn)i
           ≤  (γˆγ)−n(x)Cνn(x)θ + (γˆγ)−n−1(x)Cνn+1(x)θ

Since νθ < γˆγ, the sequence is Cauchy.
   Let hs∗= limn→∞hsn∗.

Step 3: the sequence Thsn : TWcloc(p) →TWc(q) is Cauchy.

   Since we have already found h∗, it suﬃces to show that the sequence of
functions v 7→∥Thnv∥, for v ∈P(TxWcloc(p)), is Cauchy.  This is uses a
standard C1+θ distortion estimate.

Lemma 1.3 There exists a K ≥1, such that, for all v ∈P(TxWcloc(p)) and
n ≥1, we have:
                  K−1 ≤Πn−1k=0∥Tfvk∥                           Πn−1k=0∥Tfv′k∥≤K.

Proof. For w ∈P(TM), let ϕ(w) = log ∥Tfw∥. Note that ϕ is a δ-H¨older
continuous function. Fix v ∈P(TxWcloc(p)) and n ≥1. Then

                  Πn−1k=0∥Tfvk∥       n−1X                          | log         =   |    ϕ(vk) −ϕ(v′k)|                  Πn−1k=0∥Tfv′k∥|       k=0
                                       n−1
                    ≤ X Hd(vk, v′k)δ
                                        k=0
                                         n−1
                    ≤ H X Cδνk(x)θδ
                                           k=0
                    ≤ HCδ(1 −νθδ)−1,

where ν = supz∈M ν(z) < 1. Setting K = exp(HCδ(1 −νθδ)−1) completes
the proof. ⋄


                                 8


=== PDF PAGE 9 OF 11 ===
   Using this lemma, we ﬁx v ∈P(TxWcloc(p)) and n ≥1 and estimate:
                        Πn−1k=0∥Tfvk∥
  |∥Thsnv∥−∥Thsn+1v∥| =                                 −1)vn∥|                        Πn−1k=0∥Tfv′k∥|∥Tπvn∥−∥T(fπf
              ≤ C∥Tπvn −T(fπf −1)vn∥
              ≤ C∥Tπvn −vn∥+ ∥T(ff −1)vn −T(fπf −1)vn∥
              ≤  C′νn(x)θ

Hence the sequence ∥Thsn∥: P(TWcloc(p)) →R  is Cauchy.  Let Th =
limn→∞Thsn.

Step 4: the function Ths : TWcloc(p) →TWc(q) is H¨older continuous.

  We ﬁrst show that hs∗: P(TWcloc(p)) →P(TWc(q)) is H¨older continuous,
uniformly in p ∈M and q ∈W c(p, 1). Choose ε > 0 so that νθ < (γˆγ)1+ε,
and let ρ = (γˆγ)(2+ε)/θ. Suppose that p′ ∈W loc(p).c     Then there exists an
n ≥1 such that min{d(p, p′), 1} ∈[ρn(p), ρn−1(p)).
   Our previous calculations show that there exists a K > 0 such that

                       dC0(hs∗, hsn∗) ≤Kνθn(γˆγ)−n.
Let β = 2+ε,θε  and note that

                                νθ(γˆγ)−1 < ρβ,

for all n. Thus it suﬃces to show that

                    d(hn∗(p, ξ), hn∗(p′, ξ)) ≤O(ρn(p)β),

for all n ≥1, p ∈M and p′ ∈W c(p, ρn(p)).
  Lemma 1.1 implies that

          d(f∗(p,n    ξ), f∗(p′,n    ξ)) ≤(γˆγ)−n(pn)ρn(p)θ < (γˆγ)n(p)1+ε.

Now property 2. of πs implies that

d(πsf∗(p,n    ξ), πsf∗(p′,n    ξ)) ≤  d(πsf∗(p,n    ξ), f∗(p,n   ξ)) + d(f∗(p,n    ξ), f∗(p′,n    ξ))
                        +d(f∗(p′,n    ξ), πsf∗(p′,n    ξ))
              ≤  Cd(pn, πs(pn))θ + d(f∗(p,n    ξ), f∗(p′,n    ξ)) + Cd(p′n, πs(p′n))θ
              ≤  Cνn(p)θ + (γˆγ)n(p)1+ε + Cνn(p′)θ
              ≤  O((γˆγ)n(p)1+ε)

                                 9


=== PDF PAGE 10 OF 11 ===
Again applying Lemma 1.1, we obtain that

           d(hn∗(p, ξ), hn∗(p′, ξ)) ≤  (γˆγ)−n(p′n)O((γˆγ)n(p)1+ε)
                   ≤  O((γˆγ)n(p)ε)
                    =  O(ρn(p)β).

  We now show that log ∥Ths∥: P(TWcloc(p)) →R+ is H¨older continuous.
Let v = (p, ξ) and w = (p′, ξ), and suppose as above that min{d(p, p′), 1} ∈
[ρn(p), ρn−1(p)). We showed in Step 3 that

                      ∞
                    log ∥Ths(v)∥= X ϕ(vi) −ϕ(hs∗(vi)),
                                        i=0

where ϕ = log ∥Tf∥. From this it follows that

                         ∞             ∞
|log ∥Ths(p, ξ)∥−log ∥Ths(p′, ξ)∥| = X (ϕ(vi) −ϕ(hs∗vi)) − X (ϕ(wi) −ϕ(hs∗wi))
                                             i=0                       i=0
                                      n−1                 n−1
                    ≤ X |ϕ(vi) −ϕ(wi)| + X |ϕ(hs∗vi) −ϕ(hs∗wi)|
                                             i=0                     i=0
                          ∞             ∞
                      + X |ϕ(vi) −ϕ(hs∗vi)| + X |ϕ(wi) −ϕ(hs∗wi)|
                                            i=n                     i=n
                                         n−1            n−1
                    ≤ O( X d(vi, wi)δ + X d(hs∗vi, hs∗wi)δ
                                                i=0              i=0
                          ∞         ∞
                      + X d(vi, hs∗vi)δ + X d(wi, hs∗wi)δ)
                                            i=n               i=n
                                         n−1              n−1
                    ≤ O( X ((γˆγ)−iρθn)δ + X ((γˆγ)−iρθn)βδ
                                                i=0                i=0
                          ∞
                      + X νθδi )
                                            i=n
                                                                βδ
                    ≤ O   (γˆγ)−nρθn)  + νθδn

Recall that we have chosen n so that

       d(vn, wn) ≤(γˆγ)−n(pn)ρn(p)θ < (γˆγ)n(p)1+ε = ρn(p)θ(1+ε)/(2+ε).

We also have that
                               νθ < (γˆγ)1+ε


                                10


=== PDF PAGE 11 OF 11 ===
From these facts we conclude that

               |log ∥Ths(p, ξ) −log ∥Ths(p′, ξ)∥| ≤  O(ρn(p)β′),

where
                             θ(1 + ε)   δθε(1 + ε)
                        β′ = δβ     =                  .
                            2 + ε      (2 + ε)2

References

[BPS]    Barreira, L., Pesin, Ya., and J. Schmeling, Dimension and product
         structure of hyperbolic measures, Ann. of Math. (2) 149 (1999), no.
          3, 755–783.

[BP]     Brin, M, and Ja. Pesin, Partially hyperbolic dynamical systems,
        Math. USSR Izvestija 8 (1974), 177–218.

[BW]    Burns, K. and A. Wilkinson, On the ergodicity of partially hyperbolic
         systems, preprint.

[PR]     Pinto, A. A. and D. A. Rand, Smoothness of holonomies for codi-
        mension 1 hyperbolic dynamics. Bull. London Math. Soc. 34 (2002),
         no. 3, 341–352.

[PW]    Pollicott, M., and P. Walkden, Livˇsic theorems for connected Lie
         groups, Trans. Amer. Math. Soc. 353 (2001), no. 7, 2879–2895.

[PSW]  Pugh, C., Shub, M., and A. Wilkinson, H¨older foliations, Duke
        Math. J., 86 (1997), no. 3, 517–546





                                11