Bull Braz Math Soc, New Series (2023) 54:42

https://doi.org/10.1007/s00574-023-00345-9

Coexistence Phenomena in the Hénon Family

Michael Benedicks1 · Liviana Palmisano1

Received: 3 May 2021 / Accepted: 1 May 2023

© The Author(s) 2023

Abstract
We study the classical Hénon family f a,b : (x, y) → (1 − ax 2 + y, bx), 0 < a < 2,
0 < b < 1, and prove that given an integer k ≥ 1, there is a set of parameters E k of
positive two-dimensional Lebesgue measure so that f a,b , for (a, b) ∈ E k , has at least
k attractive periodic orbits and one strange attractor of the type studied in Benedicks
and Carleson (Ann Math (2) 133(1):73–169, 1991). A corresponding statement also
holds for the Hénon-like families of Mora and Viana (Acta Math 171:1–71, 1993),
and we use the techniques of Mora and Viana (1993) to study homoclinic unfoldings
also in the case of the original Hénon maps. The final main result of the paper is the
existence, within the classical Hénon family, of a positive Lebesgue measure set of
parameters whose corresponding maps have two coexisting strange attractors.

Keywords Dynamical systems · Attractors · Hénon maps

1 Introduction

1.1 History

In 1976, the French astronomer and applied mathematician M. Hénon made a famous
computer experiment where he numerically detected but did not rigorously prove the
existence of a non-trivial attractor for a two-dimensional perturbation of the one-
dimensional quadratic map, f a,b : R2 → R2 defined by
   
x 1 − ax 2 + y
f a,b =
y bx

with a = 1.4 and b = 0.3, see Hénon (1976). Since then, several studies, both
numerical and theoretical, have been conducted with the aim of understanding this

B Michael Benedicks
michaelb@kth.se

1 KTH Royal Institute of Technology: Kungliga Tekniska Hogskolan, Stockholm, Sweden

0123456789().: V,-vol 123

 42 Page 2 of 42 M. Benedicks, L. Palmisano

family of maps which is now known as Hénon family. The complete understanding of
Hénon maps is still quite far from being achieved.
In his experiments Hénon also verified that attractive periodic orbits do indeed
occur for other parameter values from the same family. In view of this and of the result
of S. Newhouse, Newhouse (1974), stating that periodic attractors are generic, there
were no reason, at the time, to eliminate the possibility that the attractor observed by
Hénon was just a periodic orbit with a very high period.
However in 1991, L. Carleson and the first author proved the existence of the
attractor observed by Hénon for a positive Lebesgue measure set of parameter values
near a = 2 and b = 0, see Benedicks and Carleson (1991). More precisely, in the
paper it was shown that if b > 0 is small enough, then for a positive measure set of
a-values near a = 2, the corresponding maps f a,b exhibit a strange attractor.
To define what we mean by a strange attractor we first recall that a trapping region
for a map f is an open set U such that

f (U ) ⊂ U .

An attractor in the sense of Conley for a map f which has a trapping region is the set
∞
 ∞

= f j (U ) = f j (U ).
j=0 j=0

The attractor is topologically transitive if there is a point with a dense orbit. In

Benedicks and Carleson (1991) it was proved for a positive two-dimensional Lebesgue
measure set of parameters A in the (a, b) space, that there is a point z 0 (a, b) such that
z 1 = f a,b (z 0 ) satisfies the Collet–Eckmann condition,1 i.e. that there is a constant
κ > 0 such that
 
|D f n (z 1 ) 01 | ≥ eκn , for all n ≥ 0.

It is fairly easy to see that the attractor  for this set of parameters can be identified as
W u (ẑ), where ẑ is the unique fixed point of f a,b in the first quadrant, Benedicks and
Viana (2001). Moreover, the fact that the Collet–Eckmann conditions are satisfied leads
to topological transitivity, see Benedicks and Carleson (1991), and the combination
of  = W u (ẑ) and topological transitivity makes it appropriate to call the attractor
strange.
The techniques used in Benedicks and Carleson (1991) are a non trivial general-
izations of the ones presented in Benedicks and Carleson (1985) by the same authors
for the one-dimensional quadratic family. Those techniques opened the way for the
understanding of a new class of non-hyperbolic dynamical systems.
Further results have been achieved for Hénon maps by using and developing the
techniques in Benedicks and Carleson (1991). In Mora and Viana (1993) the results of
Benedicks and Carleson (1991) are obtained for a general perturbation of the family of
1 A quadratic map q (x) = 1 − ax 2 satisfies the Collet-Eckman condition if |(q j ) (1)| ≥ Ceκ j for all
a a
j ≥ 0 and some positive constants κ and C.

123
Coexistence Phenomena... Page 3 of 42 42

quadratic maps on the real line, called Hénon-like family. The statistical properties, the
existence of a Sinai–Ruelle–Bowen (SRB) measure, exponential decay of correlation
and a central limit theorem were studied in Benedicks and Young (1993) and Benedicks
and Young (2000). Furthermore the metric properties of the basin of attraction of the
strange attractor was studied in Benedicks and Viana (2001). In that paper it was
proven that Lebesgue almost all points in the topological basin for the attractor
∞

B= f − j (U ),
j=0

are generic for the SRB measure. Here U is the trapping region as above.
Other more recent approaches to generalizations of this class of dissipative attractors
were given by Wang and Young in Wang and Young (2001), Wang and Young (2008)
and by Berger in Berger (2019).
In the present paper in Theorem 1.4, we show that coexistence of periodic attractors
and strange attractors occur in the Hénon family for a positive Lebesgue measure set
of parameters. Our proof is mainly based on the techniques in Benedicks and Carleson
(1991). However the construction of the periodic attractors is inspired by Thunberg
(2001), where H. Thunberg proved the existence of attractive periodic orbits for one-
dimensional quadratic maps for parameters that accumulate on the ones corresponding
to the quadratic maps with absolutely continuous invariant measures of Benedicks
and Carleson (1985) and Benedicks and Carleson (1991). A result similar to that of
Thunberg (2001) has been obtained for Hénon maps in Ures (1995).
After the completion of this paper it was pointed out to the authors by P. Berger
that there is an alternative approach to Theorem 1.3 using the method of Newhouse
(1974), Newhouse (1979), in particular the version of the Newhouse theory for one-
dimensional families of map presented in Robinson (1983). The present approach is
however gives a different, more constructive, approach to the phenomena of Newhouse.
In particular Baire Category arguments are avoided.
Furthermore this constructive method allows us to we prove the existence of a
positive two-dimensional Lebesgue measure set of parameters in the Hénon family
for which there exist two coexisting strange attractors. This result stated as Theorem
1.6 confirms the numerical experiments of Cu Curry (1979) and is the main result of
the present paper.
The next section contains more details about our main results.

1.2 Statement of the Results

We now present our main results. We first give the definition of Hénon-like families
as in Mora and Viana (1993).
Definition 1.1 An a-dependent one-dimensional parameter family of maps Fa is called
a Hénon-like family if
 
1 − ax 2
Fa (x, y; b) = + ψ(a, x, y; b),
0

123
 42 Page 4 of 42 M. Benedicks, L. Palmisano

and we have the following properties:

(i) ψ satisfies the condition

||ψ||C 3 ≤ K bt .

(ii) Let A, B, C, D be the matrix element of

 
A B
D Fa = ,
C D

and assume A, B, C, D, satisfies the conditions stated in Theorem 2.1 of Mora

and Viana √ (1993), √ √ √
(a) |A| ≤ K , b/K ≤ |B| ≤ K b, b/K ≤ |C| ≤ K b, b/K ≤ | det D Fa | ≤
K b, ||D Fa a|| ≤ K and ||D Fa−1 a|| ≤ K /b.
(b) ||D(a,x,y) A|| ≤ K , ||D(a,x,y) B|| ≤ K 1/2+t , ||D(a,x,y) C|| ≤ K 1/2+t , ||D(a,x,y) D||
≤ K 1+2t . Moreover ||D(a,x,y) (det D Fa )|| ≤ K b1+t and ||D 2 Fa || ≤ K .
(c) ||D(a,x,y)
2 A|| ≤ K bt , ||D(a,x,y)
2 B|| ≤ K b1/2+2t , ||D(a,x,y)
2 C|| ≤ K b1/2+2t ,
||D(a,x,y)
2 D|| ≤ K b1+3t . Finally ||D(a,x,y)
2 (det D Fa )|| ≤ K b1+2t and ||D 3 Fa || ≤
t
Kb .

Remark 1.2 The original Hénon family corresponds to

 
√ y
ϕ(x, y; b) = b .
x

Theorem 1.3 Suppose Fa (., .; b) is an a-dependent Hénon-like family as in Definition

1.1. Then there is a b0 > 0 so that for all k ≥ 1, and all 0 < b < b0 , there is a
set of a-parameters Ak,b ( with fixed b) which has positive one-dimensional Lebesgue
measure, i.e. |Ak,b | > 0 and such that for all a ∈ Ak,b , Fa (., .; b) has at least k
attractive periodic orbits and at least one strange attractor of the type constructed in
Benedicks and Carleson (1991) and Mora and Viana (1993).

The method introduced to prove Theorem 1.3 gives also the following result.
Theorem 1.4 Suppose Fa (., .; b) is a Hénon-like family as in Definition 1.1. If b0 > 0
is sufficiently small, then for all 0 < b < b0 and for all a in some set A∞,b , Fa (., .; b)
has infinitely many coexisting attractive periodic orbits (the Newhouse phenomenon).
Theorem 1.3 and Theorem 1.4 hold for the original Hénon family.
Theorem 1.5 Consider the original Hénon family f a,b , 0 < a < 2, 0 < b < 1.
(a) There is a set of positive two-dimensional Lebesgue measure of parameters with
at least k ≥ 1 attractive periodic orbits and one Hénon-like strange attractor.
(b) There are parameters in the Hénon family for which there are infinitely many
attractive periodic orbits.

123
Coexistence Phenomena... Page 5 of 42 42

The existence of Hénon and Hénon-like maps in one-parameter families with

infinitely many sinks has already been established in Robinson (1983), Gonchenko
et al. (2008) and Gavrilov and Silnikov (1972). In difference to the previous
approaches, the present methods of proof are completely constructive. In particu-
lar, the methods avoid Baire category arguments, the Newhouse thickness criterion
and the persistence of tangencies is not used.
Our method allows also to obtain a stronger result about the coexistence of two
chaotic, non-periodic attractors. The following can be considered as the main theorem
of the paper.
Theorem 1.6 There is a positive two-dimensional Lebesgue measure set of parameters
A, such that for (a, b) ∈ A, the maps of the Hénon family f a,b have two coexisting
strange attractors.
Our results can be viewed as some steps in the Palis program, see Palis (2000), aim-
ing to describe coexistence phenomena for dissipative surface maps. Other coexistence
results has been obtained in e.g. [BMP, Be1, Pal].

2 Overview of Results and Methods on Hénon and Hénon-like Maps

In this section we collect definitions and constructions by Benedicks and Carleson

(1991) and Mora and Viana (1993) which will be used in the sequel. We briefly review
the construction of Collet–Eckmann maps in the quadratic family and the Hénon
family of Benedicks and Carleson (1985), Benedicks and Carleson (1991), and the
corresponding construction in Mora and Viana (1993). For more details we refer to
the original papers.

2.1 The One-dimensional Case

j
Let us first consider the quadratic family qa (x) = 1−ax 2 and we write ξ j (a) = qa (0),
j ≥ 0. We start with an interval ω0 = [a  , a  ] ⊂ (0, 2) and very close to 2. We partition
  2 −1
(−δ, δ) = |r |≥rδ Ir , where Ir = (e−r , e−r +1 ), Ir = −Ir and Ir = r =0 Ir , , where
the intervals Ir , are disjoint and of equal length. The definition is similar for negative
r :s. We do an explicit preliminary construction of the first free return so that it satisfies

ξn 1 (ω) = Irδ , ,

i.e. a parameter interval ω is mapped by the parameter dynamics a → ξn 1 (a) to a

parameter interval in the partition {Ir , }. Here r is chosen so that e−r ≥ e−αn(ω) , and
therefore Assertion 4, (ii), in Sect. 2.2 is satisfied. This condition is called the basic
assumption (BA) in Benedicks and Carleson (1991).
We give a brief description of the constructions in Benedicks and Carleson (1985),
Benedicks and Carleson (1991). At the n:th stage of the construction, we have a
partition Pn and for ω ∈ Pn , when n = n k is a free return, we have

ξn (ω) ⊂ Ir ∪ Ir −1 if r > 0.

123
 42 Page 6 of 42 M. Benedicks, L. Palmisano

(The case r < 0 is analogous.) We define the bound period at a free return as the
maximum integer p so that

|ξn+ j (a) − ξ j (a  )| ≤ e−β j ∀a, a  ∈ ω, ∀ j ≤ p. (2.1)

After the bound period there is a free period of length L, during which the corre-
sponding iterates are called free, and at time n + p + L we have a return, at which

ξn+ p+L (ω) ∩ (−δ, δ) = ∅.

This corresponds to a new free return to an interval Ir , which can either be essential,
i.e. the image covers a whole Ir , -interval or it is contained in the union of two adjacent
such intervals. The latter case is called an inessential free return. If we have an essential
return the part of ω ∈ Pn−1 , which is mapped to (−e−αn , e−αn ) is deleted and we
define the partition Pn by pulling back the intervals {Ir , } to the parts of ω that remain
after deletions. The union of  the partition elements of the parameter space that remain
at time k is written as Ak = ω∈Pk . The numbers α and β are small and positive. In
the one-dimensional case one can choose α = 400 1
and β = 100
1
. Define ρk = |rk |,
k = 0, . . . , rs . Then (ρ0 , . . . , ρs ) is an itinerary, which essentially determine the
derivative expansion that from free return time n k to free time n k+1 is always

e−3βρk
≥ . (2.2)
e−ρk
A combinatorial argument shows, see Sect. 2.2 in Benedicks and Carleson (1991),
that there are escape situations for partition elements ω at times Ẽ(ω). The definition
of an escape situation is somewhat arbitrary but let us define it as a pair (ω, Ẽ), ω ∈ P Ẽ
which is defined so that ω, under the parameter dynamics, is mapped to an interval of
size ≥ 10 1
at time Ẽ.
The escape time Ẽ has a distribution depending essentially on the itineraries
(ρ0 , ρ1 , . . . , ρs ) of the subintervals of ω ∈ Pn 0 . By Sect. 2.2 of Benedicks and Car-
leson (1991) we have
• the total time T spent in an itinerary (ρ0 , ρ1 , . . . , ρs ) satisfies

s
T ∼ ρj
j=0

• z n i , at the return times n i , i = 1, 2, . . . , s, can be viewed as almost independent

random variable,
• the distribution of the escape times after the parameter selection satisfies

|{a ∈ ω0 | E(a) > t}| ≤ C |ω0 | e−γ t

with γ , C > 0.
This is known as the large deviation argument.

123
Coexistence Phenomena... Page 7 of 42 42

2.2 The Two-dimensional Case

By perturbing the quadratic family interpreted as an endomorphism (x, y) → (1 −

ax 2 , 0), where a is close to 2, we obtain a Hénon-like map of the type given in
Definition 1.1.
If the map is orientation reversing it has a fixed point ẑ ≈ ( 21 , 0) in the first quadrant.
For small b, the unstable eigenvalue λu is approximately equal to −2 and the product of
the stable and unstable eigenvalues λu and λs , i.e. λu ·λs = d̂, where d̂ = det(D Fa (ẑ)).
One of the main new ingredients in the two-dimensional theory is that the critical
point 0 of the one-dimensional map in the n:th stage of the induction is replaced by a
critical set Cg , g ≤ Cn/ log(1/b). There is also a special set of critical points  N ⊂ Cg
on which the induction is carried on, and which is increased as the induction index n
grows. (Note that the critical set  N in the construction is only changed for a special
sequence {Nk } of times n. The induction on n is done for n satisfying Nk ≤ n ≤ Nk+1 .)
In the case of Hénon-like maps it is most natural to define instead of the critical point,
the critical value. The unstable manifold W u (ẑ) of the fixed point has a sharp turn
close to x = 1. The critical value z 1 has the property that there is κ > 0 so that

 

D F j (z 1 ) 01
≥ eκ j for all 0 ≤ j ≤ n. (2.3)

The first approximation of z 1 is defined as the tangency point between the vector field
defined by the most contracting direction of D F(z) close to (1, 0). Successively the
equation (2.3) is verified by induction for higher and higher n and this allows most
contracting directions of higher orders to be defined. This makes better and better
approximations of the critical value. This allows us to define the image z 2 of the
critical value z 1 under the maps F, and also the critical point z 0 as z 0 = F −1 (z 1 ).
The critical point z 0 will play a crucial role in our construction. Note that all this is
defined for an interval ω ∈ Pn and all points a of ω have equivalent z 0 , z 1 and z 2 . An
arbitrary point a ∈ ω can be used for the definitions.
We now define for a ∈ ω the first generation G 1 of W u (ẑ) as the segment of
W u (ẑ) from z 1 to z 2 . We also make the notation W1 = G 1 and inductively define
Wk+1 = Fa (Wk ) and then G k = Wk+1 \Wk for k ≥ 1.
The induction proceeds by using information of the critical points  N (and cor-
responding critical values) defined on segments of W u (ẑ) of generation ≤ g =
C N / log(1/b), where C is a numerical constant. One can consider  N as the set
of “precritical points”. A succesive modification procedure at the times Nk will make
the “precritical points” converge to the final critical points.
We require the following:
Consider a free return time n of the induction, and for all ω ∈ Pn all critical values
z 1 associated with  N satisfy
Assertion 4 of Benedicks and Carleson (1991), equation (12b), p.42. in Mora and
Viana (1993)
There

is a constant κ > 0 so that

 1 

(i)
D Fa (z 1 ) 0
≥ eκ j
j
∀ j ≤ n;
(ii) disth (Fa (z 1 ),  N ) ≥ e−α j
j
∀ j ≤ n.

123
 42 Page 8 of 42 M. Benedicks, L. Palmisano

The formal definition of disth (Fai (z 0 ),  N ), denoted by di in Benedicks and Carleson

(1991), is given in Assertion 1, p. 127, in that paper and this quantity at returns satisfies

3|z i − z̃ 0(i) | ≤ di (z 0 ) ≤ 5|z i − z̃ 0(i) |,

(i)
where z i is at returns, by construction located horizontally to its binding point z̃ 0 ∈
 N . The condition (ii) is called the Basic Assumption (BA) in Benedicks and Carleson
(1985), Benedicks and Carleson (1991). Roughly speaking, a binding point is chosen
at a suitable horizontal location so that the splitting argument, and the bound period
distorsion estimates of the corresponding wν∗ -vectors will be valid, see Sect. 2.3 below.

2.3 Splitting Algorithm

Now we recall the splitting algorithm for expanded vectors as in Benedicks

 and Car-
leson (1991), and Mora and Viana (1993) p. 40–41. Let wν = D F ν (z 0 ) 01 , and we
write

wν = E ν + wν∗ .

E ν corresponds to the part of wν that is in a folding situation, i.e. there are various
terms in E ν that come from a splitting at a previous return. In particular if ν is outside
of all bound periods wν = wν∗ .
We now summarize an essential part of Assertion 4 concerning distorsion of the
vectors wν∗ during the bound period, which has an analogous definition to that in the
one-dimensional case given in (2.1).
There are constants C0 and C, such that for all critical points z 0 ∈  N
(a) If p is the binding time for ζ0 to z 0

||wν∗ (ζ0 )||

C −1 ≤ ≤ C, 0 ≤ ν ≤ p.
||wν∗ (z 0 )||

(b) Let z 0 ∈  N , let ζ0 and ζ0 be two points bound to z 0 during time [0, p] and let n
be the first free return n ≥ p. Furthermore let wν∗ (ζ0 ) and wν∗ (ζ0 ) be the associated
vectors of the splitting algorithm. We write the vectors in polar coordinates, where
Mν (·) denotes the absolute value and θν (·) the argument, and measure the distance
between the orbits using

i (ζ0 , ζ0 ) = max |ζ j − ζ j |.

0≤ j≤i

Then there is a constant C0 such that, if

k
j 1
≤ , and k ≤ min(n, N ),
d j (z 0 ) C0
j=1

123
Coexistence Phenomena... Page 9 of 42 42

then if ν ≤ k
⎧ ⎫
Mν (ζ0 ) ⎨ ν
j ⎬
≤ exp C0 , (2.4)
Mν (ζ0 ) ⎩ d j (z 0 ) ⎭
j=1

and

|θν (ζ0 ) − θν (ζ0 )| ≤ 2b1/4 ν . (2.5)

Very similar estimates appear in Lemma 10.2, in Mora and Viana (1993). Their
estimate in the Modulus equation (2.4) is better with the quantity

ν

k = k (ζ0 , ζ0 ) = b(s−ν)/4 |ζs − ζs |,
s=1

instead of i (ζ0 , ζ0 ) = max0≤ j≤i |ζ j − ζ j |.

We have written (2.5) with the constant 2b1/4 as in Mora and Viana (1993) instead
of 2b1/2 as in Benedicks and Carleson (1991) since our estimates are required to work
also in the more general setting of Hénon-like maps.

2.4 Derivative Estimates and C2 (b) Curves for Hénon-Like Maps

We also need at several places that uniform expansion of the x-derivative of the n:th
iteration of a function F(x; a) automatically gives a uniform comparison of a and
x-derivatives of the iterated function. In the one-dimensional case this is formulated
abstractly in Lemma 2.1 in Benedicks and Carleson (1991). The corresponding esti-
mate in the two-dimensional case is Benedicks and Carleson (1991) lemmas 8.1 and
8.4 and Mora and Viana (1993) Lemma 11.3, which we formulate as a distorsion result
for the wν∗ vectors of the splitting algorithm.

Lemma 2.1 We consider the critical orbit z ν (a) as a function of the parameter a. We
denote its derivative with respect to a by ż ν (a). Then the following holds
For all 2 ≤ ν ≤ n and a ∈ Pν−1 (ω) ⊂ E ν−1 (z 0 ) we have
(i)

1 ||ż ν (a)||
≤ ≤ 100.
100 ||wν∗ (a)||

Moreover if ν is a free iterate then

(ii) | angle(ż ν (a), wν∗ )| ≤ bt/2 .

We also need a statement about distorsion for the tangent vectors of the parameter
dependent curves a → z ν (a), which can be formulated as follows.

123
 42 Page 10 of 42 M. Benedicks, L. Palmisano

Corollary 2.2 There is a constant C(K , α, β, δ), so that if ν is a free return then if
ω ∈ Pν−1 (z 0 ) then for all a, a  ∈ ω

||ż ν (a  )||
≤C and angle(ż ν (a  ), ż ν (a)) ≤ 10b1/4 .
||ż ν (a)||

In several places, in particular for parameter dependent curves and pieces of unsta-
ble manifolds, it is relevant that the corresponding curves segments are C 2 (b)-curves
which in the setting of the Hénon-like maps of Mora and Viana (1993), has the fol-
lowing definition.
Definition 2.3 A curve γ (x) = (x, h(x)), x1 ≤ x ≤ x2 is called a C 2 (b)-curve if the
curve is C 2 , and there is a constant C so that |h  (x)| ≤ Cbt and |h  (x)| ≤ Cbt for
x1 ≤ x ≤ x2 . The constant t > 0 appears in the definition of the Hénon-like maps.

2.5 Stable and Unstable Manifold

We also need some geometric information on the attractor. A reference is Mora and
Viana (1993), Sect. 4, but we will also need two quantitative statements on the stable
and unstable manifolds of the fixed point formulated in lemmas 2.4, 2.5 and 2.6 below.

Lemma 2.4 Let γas , a ∈ ω̃0 , be the first leg of the stable manifold of ẑ(a) pointing√in
the negative y direction. Then γas at all points has slope bounded below by K / b
where K is a numerical constant. Moreover γ s has a C 1 dependence on a. Also the
downwards pointing leg γas of W s (ẑ) intersects W u (ẑ) at a homoclinic point ẑ  .

Proof We consider the orientation reversing case when the fix point (x̂, ŷ) satisfies
ŷ > 0.
By the C 1 -version of the stable manifold theorem, there is a small segment of the
γs -leg pointing down. Note that we do not have control of the size of this leg. It depends
on a0 , the middle point of ω̃0 , and b. By C 1 continuity of the stable manifold we can
choose a sufficiently small segment 0 so that its slope is close to the slope at the fixed
point. As in Mora and Viana (1993) the derivative of the map is defined as
 
A B
D Fa (x, y) = (a, x, y).
C D

The stable direction at the fixed point has approximate slope s0 , where

−2a x̂
s0 = ,
B
and by continuity this is true also for points of 0 . Now define inductively n+1 =
Fa−1 (n ) for n ≤ n 0 , where n 0 is determined so that (x, y) ∈ n for n ≤ n 0 should
satisfy y ≥ 78 ŷ. Note that we have strong expansion of the inverse map Fa−1 and n 0 is
finite.

123
Coexistence Phenomena... Page 11 of 42 42

Next we verify that the cone defined by

1
|s − s0 | ≤ |s0 |
10

is invariant under D Fa−1 . For this we use the derivative estimates of A, B, C, D and
the determinant AD − BC in Mora and Viana (1993), Theorem 2.1. This will hold
for the sequence of curve segments {n }, n ≤ n 0 . The length of n 0 , will be greater or
equal to 18 ŷ > 0. We now do two final iterates and conclude that n 0 +2 has a subcurve
√
with vertical slope ≥ K / b and length ≥ C ŷb−1 . It follows that we have the required
homoclinic intersection ẑ  , compare Lemma 3.3. 

Lemma 2.5 Consider a family of Hénon-like maps Fa (., .; b) which is area reversing.
Let a time ν be given and let a parameter interval of a-values, ω ∈ Pν . For a ∈ ω
there is a critical point z 0 and a critical orbit z 1 , z 2 , z 3 located on W u (ẑ). Let γu be the
segment of W u (ẑ) from z 2 to z 3 . Then for a suitable choice of δ0 , the curve segment

γ1u = γu ∩ {(x, y) : x ≥ −1 + δ0 }

is an approximate parabola and the two segments

γ1u ∩ {(x, y) : x ≤ 1 − δ0 }

are two C 2 (b) curves.

Sketch of proof. For the first part of the proof we follow Mora and Viana (1993),
Sect. 7. In formula (2), p.30, they state that the unstable manifold restricted to G 0 ∩
{|x| ≤ 1 − δ0 } can be viewed as the graph y(x) = yϕ (a, x) with

||yϕ ||C 2 ≤ const bt ,

If we iterate the unstable manifold once it follows that it folds to a parabola. From a
curvature argument, see Mora and Viana (1993) Lemma 9.3, it follows that the curve
is C 2 (b). 

We will later need information on the structure of the stable manifold of the fixed
point ẑ.
Lemma 2.6 There is an approximate equidistribution of pieces of the stable manifold
W s (ẑ), with a definite slope s, |s| ≥ Const. δ that intersect {(x, y) : |x| ≥ δ}. The
interspacing of the the legs of W s (ẑ) is ∼ π2 · 3·21 k .
Proof Consider the tent map ξ → 1 − 2|ξ |. It has a fixed point ξ = 13 . The preimages
of this fixed point are located at
ν
ξν,k = , ν = −3 · 2k + 1, . . . , 3 · 2k − 1
3 · 2k

The corresponding points for the quadratic map x → 1 − 2x 2 are given by xν,k =
sin π2 ξν,k . This means that the interspacing of the legs of W s (ẑ) is as required. 


123
 42 Page 12 of 42 M. Benedicks, L. Palmisano

2.6 The Stable Foliation and Its Properties

The stable foliation of order n for different values of n will play an important role in
the following, in particular in the capturing argument in Sect. 4 and in the construction
of the sink in Sect. 3. This construction of the stable foliation appears in Benedicks
and Carleson (1991), but we will use the version in Mora and Viana (1993), Sect. 6.
We will need some lemmas about the expansion properties of the maps. Because of
the dissipative properties of the maps these will lead also to the existence of contractive
vector fields and a corresponding stable foliation.
Let F be a Hénon-like map and denote by M ν (z) = D F ν (z). Let u 0 be a tangent
vector of W u (ẑ) near ẑ. Let ζ0 = (ξ0 , η0 ) be a point on the unstable manifold, satisfying
|ξ0 | ≥ δ and for any 1 ≤ ν ≤ n, M ν (ζ0 )u 0  ≥ κ ν . We get an expansive behavior of
horizontal vectors, compare Corollary 6.2 in Mora and Viana (1993). Here κ < 1 is
allowed. We need a condition similar to partial hyperbolicity relating b and κ such as
√  4
b ≤ κ/10K 2 , compare the hypothesis of Lemma 2.9 below.
Lemma 2.7 Assume that ζ0 = (ξ0 , η0 ) is a point on the unstable manifold satisfying
|ξ0 | ≥ δ and
 ν 
 M (ζ0 )u 0  ≥ κ ν , 1 ≤ ν ≤ n. (2.6)

Then all 1 ≤ ν ≤ n and for all unit vector v0 with | slope(v0 )| ≤ 1

10 ,

 ν   
 M (ζ0 )v0  ≥ 1  M ν (ζ0 ) .
2
We will also need Lemma 6.3 in Mora and Viana (1993) which implies estimates
of the norms and angles of the expanded vectors.
Lemma 2.8 Let ζ0 and norm 1 vectors u, v satisfying

|ζ0 − ζ0 | ≤ σ n and u − v ≤ σ n

 2
κ
with σ ≤ 10K 2
, then
ν
(a) ≤ M
1
2
(ζ0 )u
M ν (ζ0 )v 
≤ 2,

 ν 
√ 2n−ν √ n
(b)
angle M (ζ0 )u, M ν (ζ0 )v
≤ σ ≤ σ .
Observe that, by Lemma 2.7, the conclusions of Lemma 2.8 are verified for all
unit vectors u, v such that u − v ≤ σ n and |slope(u)| ≤ 10 1
. Similarly, because by
construction, ζ0 = (ξ0 , η0 ), with |ξ0 | > δ is κ-expanding up to time n and therefore
we can apply Lemma 6.4 of Mora and Viana (1993), that in our setting becomes:
√
Lemma 2.9 Let ζ0 be such that |ζν − ζν | ≤ σ ν for every 1 ≤ ν ≤ n with b ≤ σ ≤
 4
κ/10K 2 . Then
ν
(a) 1
2 ≤ M (ζ0 )u
M ν (ζ0 )v 
≤ 2,

123
Coexistence Phenomena... Page 13 of 42 42

 
 2 √ ν+1
(b)
angle M ν (ζ0 )u, M ν (ζ0 )v
≤ K κ σ

for any 1 ≤ ν ≤ n and any norm 1 vectors u, v with | slope(u)| ≤ 1

10 and | slope(v)| ≤
1
10 .
The above result combined with results at the end of Sect. 6 and Sect. 7C in Mora
and Viana (1993) gives the following lemma on the existence of the stable vector field
e(n) and the corresponding stable foliation which will be instrumental for the capture
argument, Sect. 4, and also for the construction of the sink, Sect. 3.
Lemma 2.10 Let ζ0 satisfy equation (2.6) and let s be a segment of W u (ẑ) centered in
ζ0 = of length σ 2n . The stable vector field e(n) through s can be integrated from s to
G 1 = F(G 0 ). Let s1 be the arc of end points obtained on G 1 , then
(a) dist (F n (s), F n (s1 )) = K κ n ,

 
 2 √ 4
(b)
angle M n (ζ  )u, M n (ζ  )v
≤ K σ ,
0 0 κ

where ζ0 ∈ s, ζ0 ∈ s1 , u = τ (ζ0 ) and v = τ (ζ0 ).

We also need Lemma 6.1. from Mora and Viana (1993).
Lemma 2.11 If eν (z) is the most contractive direction, then for 1 ≤ μ ≤ ν ≤ n


   K b μ
(a)
angle(eμ (z), eν (z))
≤ 3K ,
 κ2 μκ
2
 
(b) D f μ (z)eν (z) ≤ 4K κ
K b
κ2
.

We consider the integral curves of the vector field

 
ẋ
= e1 (z).
ẏ

Since
 
−1 1 D −B
D F(z) =
detD F(z) −C A
√ √
and A = −2ax + O(bt ), C1 b ≤ |B| ≤ C2 b, it is easy to see that

A 2ax
slope e1 (z) = − ≈ √ .
B b

As a conclusion we get that the integral curves of the stable vector field e(1) are
approximate parabolas. At the critical value z 1 , the expansive property (2.6) is valid
and we obtain the following result, see Figure 1.
Lemma 2.12 Suppose that F satisfies the assumption of Lemma 2.10. Then there is a
quadrilateral containing the critical value, which
 n is
 completely foliated with leaves
that are integral curves of ek (z) given that k = 10 .

123
 42 Page 14 of 42 M. Benedicks, L. Palmisano

Fig. 1 Stable foliation at the critical value

Proof This is a small variation of Lemma 5.8 in Benedicks and Carleson (1991), which
we are going to pursue in the following with more detail. The idea is to successively
define smaller and smaller quadrilaterals Q n which are foliated by integral curves of
the most contractive vector field ek (z) of D F k (z).
We know that for the point z̃ 0 = z 1

 

D F(z 1 ) 1
≥ eκ̃ν , ν = 1, . . . , n.
0
n
Moreover we will only use this estimate in the range 1 ≤ ν ≤ k, k = 10 . We will
inductively define a sequence {γi } of integral curves of ei (z) through z = z 1 . We start
by defining γ1 as the integral curve of e1 (z) through z 1 . We now pick z̃ 0 = z 1 . Suppose
γi is defined and stretches from y = −1, y = 1. Pick a point ζ0 ∈ γi . Then by Lemma
6.1 (b) in Mora and Viana (1993),
  j
4K K 2b
d(ζ j , z̃ j ) ≤ .
κ κ2

 4K   K b i
Let ζ0 be on the horizontal segment containing ζ0 at distance κ κ2
,

  j  i
4K K 2b Kb
d(ζ j , z̃ j ) ≤ +5 j
κ κ2 κ2
  j
8K K 2b
≤ .
κ κ2

Define
  i 
Kb
i = z | disth (z, γi ) ≤ 16K .
κ2

Then the integral curves of ei+1 (z) are defined in i and do not leave i . We define
i+1 by the restrictive condition

123
Coexistence Phenomena... Page 15 of 42 42

  i+1 
Kb
i+1 = z | disth (z, γi+1 ) ≤ 16K .
κ2

We proceed in this way by induction. Finally we can vary the point z̃ 0 on a horizontal
line segment s through z, providing that |s| ≤ cn (for a suitably chosen c). 


3 Construction of a Sink

In the following we work in the Hénon-like setting. Let z 0 ∈  E be the critical point on
the left leg of W u (ẑ), see Sect. 2.2. One can choose z 0 uniquely for all a ∈ ω0 ∈ P E ,
see Sect. 5 in Mora and Viana (1993) or Sect. 6 in Benedicks and Carleson (1991).
We now fix E 0 to be such that z E 0 (ω0 ) is in an escape situation as defined in the end
of Sect. 2.1.

3.1 Construction of a Long Escape Situation

The aim of this section is to prove that long escape situations occur. In these situations
we can guide the dynamics to behave in the direction we wish, in particular, we can
create attractive periodic orbits.
Definition 3.1 We say that z E (ω), ω ⊂ P E , is in a long escape situation at time E if
z E (ω) is a C 2 (b) curve2 such that
 
3 5
π1 z E (ω) ⊃ , ,
8 8

where π1 is the projection on the first coordinate, i.e. if γ (t) = (γ1 (t), γ2 (t)) then
π1 γ (t) = γ1 (t).
Lemma 3.2 There exist ω̃0 ⊂ ω0 and a time E such that z E (ω̃0 ) is in a long escape
situation.
Proof This proof is purely one-dimensional, since b is small and the dynamics is
outside of (−δ, δ)×R. We use an argument very similar to that in Thunberg (2001). By
Mora and Viana (1993), there is a √ time n and an interval ω0 ∈ Pn so that π1 z n (ω0 ) ∩
(−δ, δ) = ∅ and |π1 z n (ω0 )| ≥ δ. Consequently,
√ one of the components, L n of
π1 z n (ω0 )\ (−δ, δ) has length bigger than δ/3. Let ω = [a1 , a2 ] be defined by the
relation

π1 z n (ω ) = L n = [π1 u, π1 v] ,

where u and v are the end points of the curve z n (ω ). Consider then the future iterates
z n+i (ω ), i = 1, 2, . . . , under the parameter dynamics. Observe that π1 z n+2 (ω ) is

2 See Definition 2.3

123
 42 Page 16 of 42 M. Benedicks, L. Palmisano

located at
    2 
π1 Fa21 (u), π1 Fa22 (v) = 1 − a1 1 − a1 δ 2
+ O(b ), π1 Fa2 (v)
t 2

    4 
= 1 − a1 + O δ 2 + O(bt ), 1 − a2 +  δ 3 ,

where the function (x) satisfies c1 x ≤ (x) ≤ c2 x for some

 numerical constants  c1
and c2 . Observe that Fa21 (u) and Fa22 (v) and consequently π1 Fa21 (u), π1 Fa22 (v) are
located near the saddle fixed point close to (−1, 0) where the dynamics is expanding
in the x-direction by a factor bigger than 3 as long as

3
π1 Fa2+i
2
(v) ≤ − (3.1)
4

Denote by i 0 the last i for which (3.1)

 is4 verified.
 Then π1 Fa2+i
1
0
(u) is still close to
−1; its distance to −1 is of order O δ 2− 3
. After 2 more iterates

  3 5
π1 Fa4+i
1
0 (u), π F 4+i 0 (v) ⊃
1 a2 , .
4 4




To the fixed point (x̂, ŷ) there is a symmetric point on W u (x̂, ŷ), (xˆ1 , yˆ1 ), located
approximately at (−x̂, ŷ). The leg of W s (ẑ) in the negative y-direction crosses this
homoclinic point and the slope s of √ the curve segment of γs joining the two points
(x̂, ŷ) and (xˆ1 , yˆ1 ) satisfies s ≥ C/ b on all points of γs , see Lemma 2.4. We choose
the intersection with the preimage to ensure that at the next iterate when the curve
segment intersects the stable manifold, the distance to the fixed point ẑ is defined by a
high accuracy and is very close to the width of the parabola at this x-coordinate. This
is needed to make the time E  , which will appear later, well defined, see Lemma 3.6.
Lemma 3.3 There is a subinterval ω̃0 ⊂ ω̃0 such that, for all a ∈ ω̃0 , the stable leg of
W s (ẑ) pointing downwards, denoted by γas , intersects the middle half of z E (ω̃0 ).

Proof Let ã0 be the midpoint of ω̃0 and let p1 = γãs0 ∩ z E (ω̃0 ). Let ã0 be the preimage
of p1 in ω̃0 . Observe that γãs intersects z E (ω̃0 ) at p2 . By Lemma 2.1,
0

| p1 − p2 | ≤ K |ω̃0 | ≤ K e−cE ,

where K is a positive constant. We choose now a subinterval ω̃0 ⊂ ω̃0 having midpoint
ã0 and such that z E (ω̃0 ) has length e−cE . Then ω̃0 has the required property, i.e. for
all a ∈ ω̃0 , γas intersects z E (ω̃0 ) in its middle half. 


The following lemma allows us to control the dynamics so that part of the parameter
interval returns close to a critical point with a controlled geometry, see Fig. 3. This
will create an attractive periodic orbit for all selected parameters.

123
Coexistence Phenomena... Page 17 of 42 42

Fig. 2 Stable foliation at the fixpoint

Lemma 3.4 There is a subinterval ω̃0 ⊂ ω̃0 , with midpoint ã0 and a time N so that,
z N (ω̃0 ) has the following properties:
(i) z N (ω̃0 ) is a C 2 (b) curve,
(ii) |z N (ω̃0 )| = 1001 1
DN ,
 
(iii) dist π1 z 0 (ã0 ), z N (ω̃0 ) ≤
 1 1
50 D N ,
where D N = |w N |.
The proof of Lemma 3.4 consists of several steps, formulated in a sequence of
lemmas.
Consider the phase curve γ = z E (ω̃0 ) and denote by ã0 the midpoint of ω̃0 . We
recall the λ-lemma, see e.g. Palis et al. (1982), Lemma 7.1.

Lemma 3.5 Let 0 be a saddle fixed point of a C 2 map. Let V = B u ×B s be the cartesian
product of an unstable and stable ball at the fixed point 0, let q ∈ W s (q)\{0} and
let D u be a disk transverse to W s intersecting W s in q. Let Dnu be the connected
component of F n (D u ) ∩ V to which F n (q) belongs. Given ε > 0 there exists n 0 ∈ N
such that if n > n 0 , then Dnu is ε > 0 C 1 close to B u .

In our present setting we can obtain a quantative version of the λ-lemma adapted
to our situation. In the following we refer to Fig. 2.
Lemma 3.6 Suppose a C 2 (b)-curve γ of size e−κ E crosses the leg of W s (ẑ) in the

negative y-direction. Then after E  iterates where E  ∼ E, FaE (γ ) will be a C 2 (b)
curve stretching along W u (ẑ) and across the ordinate axis x = 0 to x = − 41 . Close

to x = 0 the vertical distance between W u (ẑ) and FaE (γ ) can be estimated as

1 
≤ const. (λs ) 10 E . (3.2)

and the angles between points with the same x-coordinate satisfies
1 
≤ const. (λs ) 40 E . (3.3)

123
 42 Page 18 of 42 M. Benedicks, L. Palmisano

Fig. 3 The capturing argument

Proof We apply the construction of the stable foliation in lemmas 2.6 and 2.12. For
each point of ζ0 ∈ γ we connect it to a corresponding point ζ0 on W u (ẑ). It is then
possible to apply Lemma 2.9 with z̃ 0 = ζ0 , z̃ 0 = ζ0 and κ = (1 + ε)λs , for a suitable
ε > 0. We conclude that the estimates of (3.2) and (3.3) hold. 


Remark 3.7 Note that λu · λs = det D Fa (ẑ) and that the factor 10
1
comes from the
comparison between κ and log |λu |, where log 2 − ε ≤ log |λu | ≤ log 2, and where ε
depends on 2 − a.

Proof of Lemma 3.4. For the following we refer to Fig. 3.

(i) We apply Lemma 3.5 to a fixed parameter ã0 ∈ ω̃0 from Lemma 2.10, (b), to

γ with fixed parameter ã0 . At a certain time E  ∼ E, FãE (ω̃0 ) stretches along
  0
W u (ã0 ) covering its x-projection − 41 , 41 .
(ii) By the comparability of x and a derivatives, see Corollary 2.2, during the time
from E to E + E  and the fact that |ω̃0 | ∼ e−2cE , one can check that z E+E  (ω̃0 )
 
covers the x-projection − 18 , 18 . Now restrict ω̃0 to a subinterval ω̃0 with mid-
point ã0 so that for N = E + E  , |z N (ω̃0 )| = 100
1
D N −1 .
(iii) Note that, as in Mora and Viana (1993), Sect. 7, z N (ω̃0 ) is a C 2 (b) curve and
 
dist π1 z 0 (ã0 ), z N (ω̃0 ) ≤ 50
1
D −1
N and we also obtain by Lemma 2.10, (b), (3.3)
that the angle θ between the points of z N (ω̃0 ) with the same x-coordinate on the
first leg of W u (ẑ) satisies

1 
θ ≤ const. (λs ) 40 E . (3.4)

Here we again have to use the comparasion of parameter and phase derivatives,
Lemma 2.1 and the distorsion of the the a-derivative within a partition interval,
see Corollary 2.2.

123
Coexistence Phenomena... Page 19 of 42 42

Fig. 4 Stable foliation at the critical point

3.2 Construction of an Invariant Contractive Region

In this section we prove the existence of an invariant contractive region around the
critical point. We pick an arbitrary a ∈ ω̃0 , with ω̃0 as in Lemma 3.4. We refer to
Fig. 4.
Associated to a there is a critical point z 0 (a) located on the first left leg of W u (ẑ),
see Sect. 2.2. We fix now a curve γ : (−ρ  , ρ) → R2 on this left leg so that γ (0) = z 0 ,
where ρ = 10 1
D N −1 . and ρ  will be choosen as follows.
Close to the critical value z 1 there is, by Lemma 2.12, a quadrilateral foliated by
leaves of the stable vector field e[N /10] . The leave γ3 of e[N /10] through F(γ (ρ)) hits
W u (ẑ) in another point ζ  and ρ  is defined so that F(−ρ  ) = ζ  . The pullback of the
stable leave γ3 by F is denoted by γ3 .  
We define DN  as the domain bounded by f γ 
 ,ρ) and the stable leave γ . Let
|(−ρ
  3
D N be the pullback under F, namely D N = F −1 DN . We will prove that DN and
hence also D N are invariant under FaN for all a in ω̃0 .
Consider the tangent vector τ1 (s) of γ1 (s) = Fa (γ (s)) and write it, following
Lemma 9.6 in Mora and Viana (1993) as

τ1 (s) = α(s)e E−1 (s) + β(s)w1 ,

1
with 23 a|s| ≤ |β(s)| ≤ 25 a|s| and w1 = 0 . Observe that, at time E,
   
 
 D FaE−1 e E−1  = O b E−1 .

Denote by γ E1 and γ E2 the two sub-curves of γ defined by restricting the arclength

to (−ρ  , 0) and (0, ρ) respectively. For the image of these curves the tangent vector
decomposes as

τ E (s) = α(s)D F E−1 e E−1 (s) + β(s)w E−1 .

123
 42 Page 20 of 42 M. Benedicks, L. Palmisano

Since, by the induction, w E  ≥ eκ E , we conclude that

1

α(s)D F E−1 (e E−1 (s))
≤ O(b E−1 ) ≤ |s|w E 
2

and since slope(w E ) = O(bt ), it follows that γ E1 \γ̃ E1 and γ E2 \γ̃ E2 are C 2 (b) curves.
The curves γ̃ E1 and γ̃ E2 correspond to the subsegments close to z E , which are still in
fold periods of the initial binding to z 0 , and those segments are of size (Cb) E . The
curve γ E3 = F E (γ3 ) has, by Lemma 2.11 (b), length |γ E3 | ≤ (Cb) E .
There is, by Lemma 2.11, a stable vector field e E  defined in a vertical region
containing the curves γ E1 , γ E2 and γ E3 . By Benedicks and Carleson (1991) the curves
   
F E (γ E1 ), F E (γ E2 ) and F E (γ 3 ) are located below γ and at distance O(b E ). By the
angle estimate (3.4) it follows that except for the points still in fold period to z 0 at time
 
N = E + E  , the slopes of points of the curves γ  = F E (γ E1 ) and γ̃  = F E (γ E2 )

with the same x-coordinates is ≤ (Cb) E /40 .
 
The curve F E (γ 3 ) has diameter ≤ 2 · 5 E · (Cb) E , and it is located close to z N . At
this point we choose ρ  so that F(γ (ρ)) and F(γ (−ρ  )) are on the same stable leave
of e E close to ẑ. The curve segment F N (γ 1 ) has length
 ρ  ρ
length(F N (γ 1 )) ≤ |β(s)|w N (s)ds + O(b N )dρ
0 ρ 0

≤ 4s D N ds + O(ρb ) = 2ρ 2 D N + O(ρb N )
N
0
 2
1 −1 3 1
≤3 D · DN = .
10 N 100 D N

The length of F N (γ 2 ) is estimated similarly. Finally

 2 1
diam(F N (γ 3 )) ≤ 5 E (Cb) E ≤ .
100 D N
  −1
It follows that F N −1 DN has diameter ≤ 5
100 D N and it is at distance O(b N −1 ) to
γ . Since D FC 1 ≤ 5, then
 
F N DN ⊂ DN .

The discussion above can be summarized in the following lemma (see Fig. 5).
Lemma 3.8 For all a ∈ ω̃0 , there exists a domain D N (a) around the critical point
z 0 (a), so that
N
Fa,b (D N (a)) ⊂ D N (a).

A corresponding statement holds for the region DN (a) close to the critical value
Fa (z 0 )

123
Coexistence Phenomena... Page 21 of 42 42

Fig. 5 The invariant region at the critical point

Lemma 3.9 There exists an integer k such that, for all a ∈ ω̃0 , Fa,b
N k contracts.

Proof Take an arbitrary point z ∈ DN (a) and as in Sect. 2.3, consider the unit vector

v = α0 en (z) + β0 w1 ,
  N
where w0 = 01 and en (z) is the contracting direction of order n = 10 at z. Consider
the decomposition of D F N (z)v as

D F N (z)v = α0 D F N (z)en (z) + β0 w N +1 .

Observe that, at the first return time N , en (z) is mapped to D F N (z)en (z) with
 
 
D F N (z)en (z) ≤ 5 N −n bn . (3.5)

Let us decompose α0 D F N (z)en (z) as

 
α0 D F N (z)en (z) = α1s en F N (z) + β1s w1 ,

where, by (3.5), |α1s |,|β1s | ≤ 5 N−n bn |α0 |.

Observe now that  D F N w1  = D N . As a consequence
 
D F N (z)β0 w0 = α1u en F N (z) + β1u w0 ,

where |α1u | ≤ D N |β0 | and |β1u | ≤ 5 1

10 D N D N |β0 |. Using the notation αν = (ανu , ανs ),
βν = (βνu , βνs ), it follows that


|α1 | ≤ |α1s | + |α1u | ≤ 5 N −n bn |α0 | + D N |β0 |,
|β1 | ≤ |β1s | + |β1u | ≤ 5 N −n bn |α0 | + 10
5
|β0 |.

123
 42 Page 22 of 42 M. Benedicks, L. Palmisano

z0

z0

Fig. 6 Capturing of the second critical point

Let A be the matrix

Observe that A has spectral radius at most 21 . Finally we choose k > 0 such that
 1 k 2
2 D N < 1. Then Ak is a contraction and therefore also D F N k is a contraction. 


4 Capturing of a New Critical Point

The next step in the construction is to create a new attractor for the same parameter
values of maps with a sink, see Sect. 3. This attractor can be another sink or a strange
attractor. In order to do so, we need to select another critical point and follow its
evolution for the same parameter values as those of the first sink constructed in the
previous section.
It is important that we can use the binding critical points for the initial critical point.
By choosing its distance appropriately z ν (ω) will follow the initial critical point and
the new critical point will still be bound to the first at its first return time N . At this time
there will be a secondary bound period after which the secondary critical point again is
bound. After the third bound period we will essentially be in a situation corresponding
to the initial inductive situation in Benedicks and Carleson (1991), Mora and Viana
(1993). Using the machinery of Benedicks and Carleson (1991), we will prove that
the new critical point also will reach an escape situation. At this point we will be
able to choose parameters which go through an unfolding of a homoclinic tangency.
Following (Palis and Takens 1993) and Mora and Viana (1993), this will allow to
create a new Henon-like family and to consequently set up the inductive procedure.
More precisely, to this new Henon-like family, one could apply Sect. 3 to create a new
sink or Mora and Viana (1993) to create a strange attractor.
Our aim is first to capture a new critical point z 0 at a specific distance to z 0 . We
will show that the critical point z 0 and the segment W u (ẑ) are accumulated by leaves
of W u (ẑ) which contain other critical points. Fix a ∈ ω = ω̃0 and let z 0 = z 0 (ω) be a
critical point. We select a segment L of the unstable manifold of length 2σ n 1 around
ẑ  , see Lemma 2.4, where n 1 is a prescribed integer. By Lemma 2.10 and Lemma 2.11
it follows that the image F n 2 (L) has length ≈ 2σ n 1 · (2a)n 2 . By adjusting n 1 and n 2 ,

123
Coexistence Phenomena... Page 23 of 42 42

we obtain a sequence of long leaves γ j which accumulate on the first leg of W u (ẑ)
restricted to − 21 ≤ x ≤ 21 .
This is formulated in the next lemma, where distv (ẑ 0 , z 0 ) denotes the vertical dis-
tance between the leaves of the unstable manifold containing the critical points ẑ 0 and
z0 .
Lemma 4.1 There are constants C1 , C2 such that for all j ≥ 16 there is a critical
point zˆ0 and a corresponding segment γ̂ u containing zˆ0
  j+1  j
d̂ d̂
C1 ≤ distv (ẑ 0 , z 0 ) ≤ C2 (4.1)
2a 2a

where d̂ = det D F(ẑ).

Proof The exact estimates of (4.1) is obtained since most of the time is spent in the
linearization domain of the saddle point ẑ where the eigenvalues are ∼ 2a and ∼ d̂/2a



4.1 The New Critical Point

Observe that, for each n, γn and F ps intersects in a unique point, z 0 and that p depends
on n. Pick n so that the vertical distance

1
dv (γu , γn ) = dn = η
DN

for a suitable η satisfying 1 < η < 2 to be chosen later. Moreover, by Lemma 2.11,
(b), there exists a constant K close to 1 so that

1 maxπ1 γn |h u (x) − h n (x)|

≤ ≤K
K minπ1 γn |h u (x) − h n (x)|

where h u and h j are the graphs of γu and γn and π1 γn is the projection of h n on the
x-axe.
Lemma 4.2 Suppose that the horizontal distance satisfies

dh (γu , γn ) = dn ,

then
(n)

dh (z 0 , z 0 ) ≤ dn

Proof This is a reformulation of Lemma 5, Sect. 2.3.1 of Benedicks and Young (1993)
and the same proof applies also in our setting. 


123
 42 Page 24 of 42 M. Benedicks, L. Palmisano

Lemma 4.3 At time N , z N (ω) = F N (z 0 ) is located in horizontal position to z 0 .

Moreover there exists a constant K close to 1 so that
1
dh (z 0 , z N ) ≤ dh (z N , z N ) ≤ K dh (z 0 , z N ).
K
Furthermore
1 1−η
≤ dh (z 0 , z N ) ≤ K 1 D N
1−η
D
K1 N

for some constant K 1 close to 1.

Proof Let 0 be a curve joining z 0 and z 0 and let 1 be its image joining z 1 and z 1
close to the critical value. On 0 , using Sect. 2.3, we decompose the tangent vector
as
1
τ (z) = α(y)e N (z) + β(y) 0

with z = (x, y) ∈ 0 . Consider now the vertical segment from z 0 to γn and let yn , yn
be the y-coordinates of its end points. Then
 yn
1
dn ≤ β(y)dy ≤ K dn
K yn

 
with K a constant close to 1. Use the notation w j = D F j (z 0 ) 01 and apply the
distortion estimates during the bound period for w j , see Lemma 10.2 in Mora and
Viana (1993), which gives

1
D N ≤ w N  ≤ K D N .
K
Furthermore
1 1 1
≤ dn ≤ K η .
K D ηN DN

This proves the last inequality of the lemma. 


Observe now that, by Corollary 5.7 in Benedicks and Carleson (1991), w N and the
tangent vector τ N are aligned with γu forming an angle smaller than dn4 . Note that
Lemma 5.5 and Corollary 5.7 in Benedicks and Carleson (1991) do not depend on
the special form of the map and applies also in our context. As final remark, one can
notice that the distortion during the bound period are stated in the case of phase space
dynamics. Moreover they are valid also in the parameter dependent setting because of
the uniform comparison between the x and a-derivatives, see Corollary 2.2.
The second bound period from time N to time 2N . Note that, for η close to 2,
 (ω) will still be bound to z and that z  (ω) is located in horizontal position with
z 2N N N

123
Coexistence Phenomena... Page 25 of 42 42

respect to z 0 . We repeat the same procedure as in Lemma 4.3. Join z 0 and z N (ω) by
a curve 0 and decompose the tangent vector of 1 = F(0 ) as
1
τ (s) = A(s)e N (s) + B(s) 0 ,

2 s ≤ B(s) ≤ 2 s, see Lemma 9.6 in Mora and Viana (1993)

where B(s) satisfies 3a 5a

and Assertion 4(c) in Benedicks and Carleson (1991). Again by the bound distortion
 (ω)) and d(z , z  (ω))
lemma in Mora and Viana (1993) (Lemma 10.2), d(z N , z 2N 0 2N
can be estimated from below and above using

 s 

1 2

s DN ≤
B(t)dt w N

≤ K s 2 D N
K 0

 (ω)). A similar statement for points in horizontal position appear

where s = d(z 0 , z 2N
in Benedicks and Carleson (1991), Assertion 4, (b) and (c) and in Mora and Viana
(1993), Corollary 10.7. We conclude that
 2
 (ω)) is comparable with a fixed constant to D 1−η
(a) d(z 0 , z 2N DN = DN ,
3−2η
N
 (ω)| is comparable to |z  (ω)|D
(b) |z 2N
1−η 1−η
N N D N , which is comparable to D N .

Let us now study the period when z 2N +ν (ω), ν ≥ 0, is bound to z 0 (ω).
We define the preliminary binding period p1 as the maximal integer so that, for all
ν ≤ p1 ,



z
≤ e−βν .
2N +ν (ω) − z ν

In principle p1 could be infinite, but this is not the case.

Lemma 4.4 The preliminary binding period p1 < ∞.

Proof The proof of this fact will follow after the proof of Lemma 4.5. 




Lemma 4.5 Let ρ =

z 2N +ν (ω) − z ν
. If ν ≥ ν0 is outside of all folding periods, then

3a 2

5a 2
ρ wν  ≤
z 2N

+ν (ω) − z ν ≤ ρ wν  , (4.2)

2 2
where wν = D F ν (z 0 )w0 .

Proof We introduce an horizontal curve 0 joining z 0 and z 2N with tangent vector
ν
τ (s). The length of ν = F (0 ) is equal to
 ρ  
 D F ν (0 (s))τ0 (s) ds.
0

We decompose
1
τ1 (s) = A(s)eν−1 + B(s) 0 ,

123
 42 Page 26 of 42 M. Benedicks, L. Palmisano

and then
1
τν (s) = A(s)D F ν−1 (0 (s))eν−1 + B(s)D F ν−1 (0 (s)) 0 ,

where, by Lemma 4.2

3a 5a
s ≤ |B(s)| ≤ s, (4.3)
2 2

see Sect. 8 in Mora and Viana (1993). We apply the splitting algorithm from Sect. 8,
(i) − (v) in Mora and Viana (1993) to D F ν−1 (0 (s)). If v is outside of it follows
from (4.3) and integrating that
 ρ
3 2 5 2
aρ wv  ≤ τv (s) ds ≤ aρ wv  .
4 0 4

We conclude that Lemma 4.5 holds. 



Proof of Lemma 4.4. By the basic assumption which is part of the induction, see
Assertion 4 (ii) in Sect. 2.2,

d(z v (a), C) ≥ e−αv ,

and ρ = d(z v (a), C). Since by the induction ||wν || ≥ eκν , ν = 1, 2, . . . , n, it follows
that p1 < ∞. 

Suppose now at the time p1
 
  
z 2N + p1 +1 (ω) − z 2N + p1 +1 (ω) ≥ e−β( p1 +1) .

We follow an argument from Benedicks and Carleson (1991), Subsection 6.2. It

follows from the basic assumption, see Assertion 4 (ii) in Sect. 2.2, that

d(z v (a), C) ≥ e−αv

that the deepest and longest bound period for z j satisfies p̃1 ≤ 4α p1 . The next level
bound period satisfies p̃2 ≤ 4α p̃1 . As consequence the length of the combined bound
period of z p1 will be less than

4α
p̃v ≤ 4α p1 + (4α)2 p1 + · · · = p1 .
ν
1 − 4α

This means that at the time p,

  1
3ρ 2 w p  ≥ e−β p1 .
44α p1 (1−4α)

123
Coexistence Phenomena... Page 27 of 42 42

 
But p1 ≤ p ≤ 1 + 4α
1−4α p1 . If we chose β = 10α as in Benedicks and Carleson
(1991) we obtain
 
3ρ 2 w p  ≥ e− 4 β p1
3
(4.4)

and also
 
3ρ 2 w p  ≥ ρ 2 e−β p . (4.5)

We can choose β1 satisfying

3
β ≤ β1 ≤ β
4

so that we have the estimate

 
ρ 2 w p  ≥ C −1 e−β1 p .
 
Let us also denote D p = w p . This means that with p as in 4.5

 
1−η 2
C −1 e−β1 p ≤ D p D N ≤ e−β1 p .

On the other hand

e(c1 −α) p ≤ D p ≤ ec1 p

so we obtain that
 
1−η 2
C −1 D −β
p
2 ≤ D
p DN ≤ C D −β
p ,
2

β1 β1
where c1 ≤ β2 ≤ c1 −α . Hence

2(η−1) 2(η−1)
C −1 D N
1+β2 1+β2
≤ D p ≤ C DN .

Note that the estimate

C −1 D −β
p
2 ≤ ρ 2 D ≤ C D −β2
p p

implies that

− 1 β2 1
− 1 β2
C −1/2 D p 2 ≤ ρ D p2 ≤ C 1/2 D p 2

123
 42 Page 28 of 42 M. Benedicks, L. Palmisano

and we obtain that


1 1
− β2

z 2N + p (ω)
∼
z 2N

(ω)
2ap D p ∼ 2a D N D p2 2 .
1−η

We now choose η = 3
2 + . This means that
 
 2 1


1 1 2 +


− 1 − 1 − 1 β2 − 1 − 2 − 2 β2 1+β2

z 2N + p (ω)
≥ 2a D N 2 D p2 2 = 2a D N 2 D N .

β2 β2
− − −β
If  = β22 we obtain that 2a D N 2 2 = 2a D N 2 .
We then follow the segment until the next return 2N + p + and


−β

z 2N + p+ (ω)
≥ const D N 2 .

Since D N ≥ eκ N , we obtain

z 2N + p+ (ω)
≥ const e−κβ2 N

and the free period satisfies ≤ β2 κκ1−1 N , where κ1 is the Lyapunov exponent
associated to the dynamics outside of (−δ, δ). Moreover, the time 2N + p + is less
than or equal to 3N . We can now relax the condition of the basic assumption, see
Sect. 2.2 and apply the machinery to a subinterval ω ⊂ ω which is chosen so that


1





z 2N + p+ (ω )
≥
z 2N + p+ (ω)
.
4
As a consequence

z 2N + p+ (ω )
≥ const’ e−κβ2 N .

The corresponding bound period for a return time to a position at horizontal distance

e−r with r  ≤ β2 N has length smaller than or equal to 4β2 N < N . In particular, we
can

use that the

induction is valid up to time N and we can repeat
the argument for

 

 

z 2N + p+ (ω )
. At the expiration time of the new bound period p1 ,
z 2N + p+ + p1 (ω )

satisfies



 

z 2N + p+ + p1 (ω )
≥ const er (1−3β)
z 2N + p+ (ω )
,

see (2.2). After a finite number of steps s, at time n s and for a parameters interval ω(s) ,
we have

 

1

z n s ω(s)
≥ .
10

123
Coexistence Phenomena... Page 29 of 42 42

Fig. 7 Long escape situation for the second critical point

We are then in an escape situation and the argument in Sect. 3 applies.

5 Construction of a Tangency

We aim to construct a non-degenerate quadratic tangency at the long escape time Ñ .

We consider a parameter interval ω̃. For each a in ω̃ there is a critical point z̃ 0 and
a fixed point ẑ a For each fixed a ∈ ω̃ a segment γua ⊂ W u (ẑ a ) which contains z̃ a
and ẑ a . We aim to prove that FaÑ (γua ) has very high curvature near FaÑ (z̃ 0 (a)). It is
advantageous to study the curvature of FaÑ −1 (γ̃ua ) where γ̃ua is the curve Fa (γua ) which
is located close to the critical value.
We decompose the tangent vector τ (u) along γ̃ua as

τ (u) = A(u)E Ñ −1 (u) + B(u)W Ñ −1 (u)

where
⎧
⎨ E Ñ −1 (u) = e Ñ −1 (ζ1 (u)) 
1
⎩W Ñ −1 (u) = D FaÑ −1 (ζ1 (u))
0

and ζ1 (u) is the parametrization of γ̃ua by arclength.

We have
 ρ
 
ζ Ñ (ρ) − z̃ Ñ = A(u)E Ñ −1 (u) + B(u)W Ñ −1 (u) du. (5.1)
0

Here ζ1 = ζ1 (ρ) is an arbitrary point on γ̃au at arclength ρ from z̃ 1 (a) and ζ Ñ (ρ) =
FaÑ −1 (ζ1 (ρ)).
Differentiating (5.1) twice, we get

ζ Ñ (ρ) = A(ρ)E Ñ −1 (ρ) + B(ρ)W Ñ −1 (ρ) (5.2)

123
 42 Page 30 of 42 M. Benedicks, L. Palmisano

and

ζ Ñ (ρ) = A (ρ)E Ñ −1 (ρ) + A(ρ)E Ñ −1 (ρ) + B  (ρ)W Ñ −1 (ρ) + B(ρ)W Ñ −1 (ρ)
(5.3)

Lemma 5.1 For all ρ > 0

W Ñ −1 (ρ)
≤ 25 Ñ −1 .

Proof Observe that

1
W Ñ −1 (ρ) = D F(x Ñ −1 , y Ñ −1 ) . . . D F(x1 , y1 ) 0 .

By differentiating with respect to ρ and taking the matrix norm, one gets,
⎛ ⎞


! 


⎝  D F(x j , y j )⎠ Pi 

W Ñ −1 (ρ)
=
i j=i

where
 
d −2axi + ∂x ψ1 ∂ y ψ1
Pi = .
dρ ∂x ψ2 ∂ y ψ2

Since the C 2 norms of ψ1 and ψ2 have the bound Cbt/2 , see Mora and Viana (1993),
Sect. 7A, we get
$ %


 9  Ñ −2  9 i



W Ñ −1 (ρ)
≤ ·3 ≤ 25 Ñ
2 2
i

 9 i
where we used that Wi  < 2 (since D F < 29 ). 


E (0)

Proposition 5.2 Let |ρ0 | =  Ñ  ,

W (0) then for a suitably chosen ρ0 , and for all ρ0 ≤
Ñ 
|ρ| ≤ ρ0 , the curvature of ζ Ñ (ρ), κ ζ Ñ (ρ) satisfies the following:

|W Ñ (ρ)|   C1 |W Ñ (ρ)|
2C1 ≥ κ ζ Ñ (ρ) ≥
|E Ñ (ρ)| 2 2 |E Ñ (ρ)|2

with 2 ≤ C1 ≤ 4.

Remark 5.3 Observe that the numbers 21 and 2 appearing in the curvature estimates
above can be chosen arbitrarily close to 1, if b is sufficiently small.

123
Coexistence Phenomena... Page 31 of 42 42

Proof Recall that

ζ Ñ (ρ) × ζ Ñ (ρ)

κ(ρ) =

3 .



ζ Ñ (ρ)

We start by computing ζ  (ρ) × ζ  (ρ). We get

Ñ Ñ

ζ Ñ (ρ) × ζ Ñ (ρ) = A(ρ)A (ρ)E Ñ −1 (ρ) × E Ñ −1 (ρ) + A(ρ)2 E Ñ −1 (ρ) × E Ñ −1 (ρ)
+ A(ρ)B  (ρ)E Ñ −1 (ρ) × W Ñ −1 (ρ) + A(ρ)B(ρ)E Ñ −1 (ρ) × W Ñ −1 (ρ)
+ A (ρ)B(ρ)W Ñ −1 (ρ) × E Ñ −1 (ρ) + A(ρ)B(ρ)W Ñ −1 (ρ) × E Ñ −1 (ρ)
+ B(ρ)B  (ρ)W Ñ −1 (ρ) × W Ñ −1 (ρ) + B(ρ)2 W Ñ −1 (ρ) × W Ñ −1 (ρ)

and since E Ñ −1 (ρ) × E Ñ −1 (ρ) = W Ñ −1 (ρ) × W Ñ −1 (ρ) = 0

 
ζ Ñ (ρ) × ζ Ñ (ρ) = A(ρ)B  (ρ) − A (ρ)B(ρ) E Ñ −1 (ρ) × W Ñ −1 (ρ)
+ A(ρ)2 E Ñ −1 (ρ) × E Ñ −1 (ρ) + B(ρ)2 W Ñ −1 (ρ) × W Ñ −1 (ρ)
+ A(ρ)B(ρ)E Ñ −1 (ρ) × W Ñ −1 (ρ) + A(ρ)B(ρ)W Ñ (ρ) × E Ñ −1 (ρ).

In Benedicks and Carleson (1991), Sect. 7.5 there are estimates of A, A , B and B 
in the classical Hénon case. We have new similar estimates in the Hénon-like case as
follows:
Claim. There are constants γ1 and C1 = 2a/γ02 such that

2
ρ ≤ B(ρ) ≤ 4
γ02
ρ, (5.4)
γ02
2a dy
B  (ρ) = + O(b) = C1 + O(b), (5.5)
γ02 dρ
A(ρ) = 1 + O(ρ 2 ), (5.6)
A (ρ) = O(ρ). (5.7)

We prove now the previous Claim. Observe that for (x, y) close to the critical value
z̃ 1
   
−2ax + ψ1x ψ1y −2ax + α1 β1
D fa = =
ψ2x ψ2y γ1 δ1

and
 
1 δ1 −β1
D f a−1 = .
detD f a −γ1 −2ax + α1

123
 42 Page 32 of 42 M. Benedicks, L. Palmisano

It follows that
   
.0 1 −δβ1 − β1 (−2ax + α1 )
(D f a−1 )2 =
1 (detD f a )2 β1 γ1 + (−2ax + α1 )2

This means that the most contractive direction for (x, y) close to z̃ 1 (a) has slope

2ax − α1
s≈ .
β1

By the construction of the local stable manifold, see Benedicks and Carleson (1991)
pp. 110-111, it follows that there is a temporary stable foliation with slope of ≈ 2ax1 /β
with x1 ≈ 1.
We claim that the image of the leg of the unstable manifold near the critical value is
an approximate parabola. The unstable direction at  ẑ is given by the unstable direction
1/2
of the fixed point located approximately at . The slope of W u near z̃ 0 is given
0
by
    
−2ax−1 + α−1 β−1 −2ax−2 + α−2 β−2 −2ax−k + α−k β−k
... v
γ−1 δ−1 γ−2 δ−2 γ−k δ−k 0

where v0 is the slope of W u (ẑ) which is essentially horizontal. Observe that x−1 is
approximately given by 1 − ax−1 2 = 0. The slope of W u near z̃ is approximately
0
given by γ−1 /−2ax−1 .
By approximating the unstable manifold by a straight line

x = x0 + t
y = y0 + kt

where k = γ−1 /−2ax−1 we have that the image of W u is the curve (x1 (t), y1 (t) with
the derivative

x1 (t) = −2a(t + x0 ) + α0 + kβ0
y1 (t) = γ0 + kδ0

so The curvature is then given by κ(t) = |γ1 (t) × γ2 (t)|/|γ1 (t)|3 ≈ 2a/γ02 .
This means that, in a suitable almost orthogonal coordinate system (η1 , ξ1 ), one
can use a version of Hadamard’s lemma, see Lemma 8.7 in Benedicks and Carleson
(1991) to get that the image parabola looks approximately as

 2
η1
ξ1 = 1 − a .
γ0

123
Coexistence Phenomena... Page 33 of 42 42

For convenience of the reader we recall here Lemma 8.7, Benedicks and Carleson
(1991) which we just used. Let f ∈ C 2 (A, A + ) and suppose that

| f (a)| ≤ M0 , | f  (a)| ≤ M2 .

Then if

4M0 < 2

it hold that

| f  (a)| ≤ M0 (1 + M2 ).

This completes the proof of the Claim.

The following estimates hold.

 
3

A(ρ)B  (ρ) − A (ρ)B(ρ) E

Ñ −1 (ρ) × W Ñ −1 (ρ) ≥ 4 C 1 E Ñ −1 (ρ) × W Ñ −1 (ρ)

C1

≥ E Ñ −1 (ρ)

W Ñ −1 (ρ)
,
2
where we used the fact that the angle between W Ñ −1 and E Ñ −1 is very close to π/2l,
see formula (9), Sect. 6 in Mora and Viana (1993). By Lemma 6.8 in Mora and Viana
(1993), we get

E Ñ −1 (ρ) × E Ñ −1 (ρ)
≤
E Ñ −1 (ρ)

E Ñ −1 (ρ)

≤
E Ñ −1 (ρ)
(K 1 b) Ñ −4

with K 1 > 0. By Lemma 5.1 we have

42

B(ρ)2 W Ñ −1 (ρ) × W Ñ −1 (ρ)
≤
W Ñ −1 (ρ)
2 ρ 2 25 Ñ −1
γ0


2 

16
E (ρ0 )

≤
W Ñ −1 (ρ)

E Ñ −1 (ρ)
· 2
Ñ −1

2 25 Ñ −1
γ0
W (ρ )
|E (ρ)|
Ñ −1 0 Ñ −1
1

≤ W Ñ −1 (ρ)

E Ñ −1 (ρ)
,
100
We have used the distorsion estimate for W -vectors, see Mora and Viana (1993),
Lemma 10.2. to conclude that W Ñ −1 (ρ) and W Ñ −1 (ρ0 ) are comparable, the estimate

that |W Ñ −1 (ρ)| · |E Ñ −1 (ρ)| ≈ bt( Ñ −1) and the estimate

E Ñ −1 (ρ)
< (K b/κ) Ñ −1
By Lemma 6.8 in Mora and Viana (1993),

8

A(ρ)B(ρ)W Ñ −1 (ρ) × E Ñ −1 (ρ)
≤ 2 |ρ|
W Ñ −1 (ρ)
(K 1 b) Ñ −4
γ0

123
 42 Page 34 of 42 M. Benedicks, L. Palmisano

8
E Ñ −1 (ρ0 )

≤ 2

W Ñ −1 (ρ0 )
(K 1 b) Ñ −2

γ0 W Ñ −1 (ρ0 )

≤ E Ñ −1 (ρ0 )

W Ñ −1 (ρ0 )
.
100

By Lemma 5.1 we have

8

A(ρ)B(ρ)E Ñ −1 (ρ) × W Ñ −1 (ρ)
≤ 2 |ρ|
E Ñ −1 (ρ)
25 Ñ −1
γ0

E Ñ −1 (ρ0 )
Ñ −1

≤ 2 W Ñ −1 (ρ0 ) E Ñ −1 (ρ0 )

2 25
γ0
W (ρ0 )

Ñ −1
1

≤ W Ñ −1 (ρ0 )

E Ñ −1 (ρ0 )
,
100

2

2

where we used that |ρ|2 ≤ |ρ0 |2 =

E Ñ (ρ0 )
/
W Ñ (ρ0 )
and
E Ñ −1 (ρ)
<
(K b/κ) Ñ −1 , K , κ > 0, see formula (5) of Sect. 6 in Mora and Viana (1993).
The term that dominates is A(ρ)B  (ρ)E N −1 ˜ (ρ) × W N −1˜ (ρ). The proof of the
lemma is concluded by combining the previous five estimates. 


5.1 Quadratic Tangency

We prove that in a long escape situation a quadratic tangency appears.

Proposition 5.4 Let z E (ω) be a curve segment of critical values in an escape situation
that intersect γ s , the leg of W s (ẑ) pointing downwards. Then there exists a unique
a0 ∈ ω such that the tangency between γas0 and γau0 is quadratic.

Remark 5.5 Actually, the curvature of γas0 is close to zero while the curvature of γau0
|W N |
is close to its maximal which is 2 |E |2
within a factor close to 1.
N

√
Proof By Proposition 5.2, the ρ which makes the slope equal to −C/ b is roughly

|E N | √
ρ=− b.
2C|W N |

Observe that this ρ satisfies the estimate |ρ| ≥ ρ0 , so we avoid the exact tip of the
parabola like image of the unstable manifold. We use the bounds in Proposition 5.2
for the curvature and the angle between E Ñ (0) and W Ñ (0) is π2 . Using that ||W Ñ || ≤
√
25 Ñ (Lemma 5.1) and ||D E Ñ || ≤ C b (Mora and Viana (1993), Lemma 6.6), the
statement follows. 


123
Coexistence Phenomena... Page 35 of 42 42

Fig. 8 Quadratic tangency

6 Proof of Theorems 1.3, 1.4 and 1.5

The proof of theorems 1.3, 1.4 and 1.5 is done by induction. From Sects. 3 and 5 we
selected maps with a sink and a new tangency. We reapply now Sect. 3 to get a second
sink and Sect. 5 to get a new tangency. One could stop this process after k steps. At
this moment one would have k sinks and a new tangency. This tangency will then
be used to create a strange attractor using Mora and Viana (1993) and give the proof
of Theorem 1.3. Alternatively, one could continue the process infinitely many times
to get infinitely many sinks. This leads to the proof of Theorem 1.4. The inductive
procedure is formulated in the next proposition.
Proposition 6.1 There exists K > 0 such that, for all k = 0, 1, . . . , K , there are
parameters intervals ωk with ωk ⊂ ωk−1 , so that, for all a ∈ ωk , there is a C 2 (b)
curve γk (a) ⊂ W u (ẑ) with z k (a) ∈ γk (a). Moreover, for all k = 0, 1, . . . , K there are
regions D Nk (a) with D N j (a) ∩ D Ni (a) = ∅ for all i = j such that D Nk (a) is bounded
by γk (a) and parabolic leaves of Wloc s and it contains a unique sink.

Proof We proceed by induction and the case of one sink appears in Sect. 3. Assume that
we have already constructed k sinks and that a parameter interval ω(k) corresponding
(k+1)
to the critical point z 0 is in escape situation and intersects W s (ẑ). We now have an
unfolding of a homoclinic tangency as in Palis and Takens (1993) and Mora and Viana
(1993). We can then do the renormalization procedure associated to this unfolding as
in these papers and we obtain a new renormalized Hénon-like family. This allows us
to create a new sink as in Sect. 3, and we obtain also a new escape situation following
the argument in Sect. 4. 

Proof of Theorem 1.3. The proof is a small modification of that of Proposition 6.1.
The only difference is that, at the time k, instead of construct a new sink one can create
a strange attractor as in Mora and Viana (1993) at the homoclinic unfolding. 

Proof of Theorem 1.4. The proof is a minor modification of that of Theorem 1.3.
The only difference is that instead of switching to construction of a strange attractor
after k steps, we continue to construct more and more sinks. We obviously obtain
Newhouse parameters in the limit. Note that the renormalizations take parameters of

123
 42 Page 36 of 42 M. Benedicks, L. Palmisano

a specific Hénon-like family linearly to new renormalized parameters of the corre-

sponding Hénon-like family. For each renormalization of order k, we get a set Ak of
parameters in the renormalized Hénon-like family of maps with k sinks. We denote by
Ak the pullback of Ak containing parameters of the original Hénon-like family. Con-
sider now a non-empty closed subset of Ak , Bk and denote by Bk the push-forward of
Bk . We do at this point, another renormalization and we get a sequence of inclusions

A1 ⊃ B1 ⊃ · · · ⊃ Ak ⊃ Bk ⊃ Ak+1 ⊃ . . . .

The intersection
∞

Ak
k=1

is then non-empty and so is then the set of maps with infinitely many sinks. 

Proof of Theorem 1.5. This result is a direct consequence of Theorem 1.3 and
Theorem 1.4, since the Hénon family is a special example of a Hénon-like family. 


7 Construction of Two Coexisting Strange Attractors

In this section we prove the existence of two strange attractors for a parameter set of
positive Lebesgue measure within the classical Hénon family.
We first outline the proof. The idea is to find parameters with two coexisting homo-
clinic tangencies. To do this we consider two very close critical points which are
in escape situation simultaneously. We must chose them very carefully so that their
images are at suitable distance at the escape situation. To do this we have to chose
carefully their initial distance and the time they spend in the hyperbolic region out-
side of (−δ, δ). We will create one true tangency for the first critical point at the
point a0 and then we create a tangency for a second critical point but for a different
parameter value a0 . Both critical points will have associated parameter sets of positive
one-dimensional Lebesgue measure with different strange attractors. These parameter
sets will intersect if the parameters with the respective strange attractor are abundant
at the respective points.
We return to the construction of the first critical point z 0 and the corresponding
long escape situation of Sect. 3.1. We fix b < b0 and by Lemma 3.2 we see that there
is a subinterval ω̃0 such that z E (ω̃0 ) is in a long escape situation.
We now construct a second critical point z̃ 0 . The construction is similar to the
corresponding one in Sect. 4. The difference is that z̃ 0 will be chosen much closer to
z 0 vertically than z̃ 0 is to z 0 and its distance can be chosen exponentially well spaced,
see (4.1). From Lemma 4.1, choose j and the corresponding ẑ 0 so that j is the minimal
integer so that for all a ∈ ω̃0 at time E, z̃ E is still bound to z E .

123
Coexistence Phenomena... Page 37 of 42 42

7.1 Proof of Theorem 1.6

We start with the construction of the first critical point z 0 and follow it until the first
escape situation which appears at time T0 .
Close to z 0 we have a number of critical points which are inter-spaced as follows.
Proposition 7.1 Close to the critical point z 0 we have a sequence of critical points
( j)
z̃ 0 so that with λ1 denoting the unstable eigenvalue,
 j
( j) b
|z 0 − z̃ 0 | ≤ C
|λ1 |

and
b ( j) ( j−1) ( j+1) ( j) b ( j) ( j−1)
C |z̃ − z̃ 0 | ≤ |z̃ 0 − z̃ 0 | ≤ C  |z̃ − z̃ 0 |.
|λ1 | 0 |λ1 | 0

Proof The stable manifold at the fixed point ẑ intersects the first leg of the unstable
manifold in a homoclinic point z h . Segments around z h are captured towards ẑ and
by the Lambda lemma these segments are accumulated on the unstable manifold, in
particular at z 0 . The behavior is dominated by the behavior at the fixed point ẑ and is
dominated by the stable eigenvalue λ2 = b/λ1 . 

( j)
We want to chose a j so that z̃ 0 is a suitable distance to z 0 so that at an escape
( j)
time E booth the point z 0 and z̃ 0 escapes.
Consider a subinterval ω in the parameter space that escapes at time T0 . We can
( j)
accomplish that |z 0 (ω)| ∼ 10
1
. Now chose a second critical point z̃ 0 (a) for a ∈ ω.
( j)
Denote the initial distance between z 0 (a) and z̃ 0 (a) for a fixed a ∈ ω by d j . By
Proposition 7.1 it follows that
 j
b
dj = O .
|λ1 |

It follows that at an escape time T0

 j
( j) b
|z T0 (a) − z̃ T0 (a)| ∼ ||WT0 (a)||,
|λ1 |

where ∼ denotes that the quotients of the two sides are bounded above and below with
fixed constants.
We want to accomplish that at a suitable time T0 + L there are simultaneous inter-
( j)
sections of z T0 +L (ω ) and z̃ T0 (ω ) with different legs of the stable manifold W s (ẑ) for
all a ∈ ω . To achieve this we study the distribution of the vertical segments of W s .
We consider the tent map

y → 1 − 2|y|

123
 42 Page 38 of 42 M. Benedicks, L. Palmisano

which is conjugate to the full quadratic map

1 − 2x 2

The preimages of the fixed point y = 1

3 of the tent map are located in

k
yk,ν = , k = −3 · 2ν − 1, . . . , 3 · 2ν + 1.
3 · 2ν
These points correspond to xk,ν = sin π2 yk,ν and by continuity this is approximately
true for all a close to 2. To each xν,k there corresponds an almost vertical branch γk,ν
( j)
of W s and, by choosing j and L appropriately, we will have the situation that z̃ 0 (a0 )
is located between γk,ν and γk+1,ν for suitable chosen k and ν. This follows by the
following argument. Suppose that for a given orbit z E+ j (ω ), j ≥ 0, moves outside
(−δ, δ)×R. Note that ||D f || ≤ 5. By Lemma 4.5 in Benedicks and Carleson (1991) it
follows that the slope sT + j of wT0 + j satisfies |sT0 + j | ≤ b/δ if we restrict to a suitable
parameter interval ω .
After restricting ω further if necessary we can obtain that for some j = L, say
z T0 + j (ω ), stretches across one γk,ν and the stable leg of W s (ẑ). Denote the intersection
points by a0 and a0 respectively.
We will need information about the local behavior at the image. It follows from
Proposition 5.2 that there are points of tangencies for parameters ã0 , ã0 ] close to a0
respectively.
Let us consider the homoclinic tangencies that appears in Proposition 5.4 for the
parameters as = ã0 and as = ã0 at time T0 + L
Suppose that the common tangency occurs for a parameter a0 . We consider the
normalization argument in Palis and Takens (1993).
The curvature is given, by

|W N1 (ρ)|
Q1 =   2 3/2
|B(ρ)||W N1 |
|E 1N |2 1+ |E 1 |2N

The maps ϕμN are written in coordinates

(1 + x, y) → (0, 1) + (H1 (μ, x, y), H2 (μ, x, y))

with

H1 (μ, x, y) = v · x 2 + μ + wy + H̃1 (μ, x, y)

H2 (μ, x, y) = u · y + H̃2 (μ, x, y).

They define N dependent on reparametrization of the parameter μ and a μ-

dependent change of coordinates renormalizations. The parameter renormalization
is given by

123
Coexistence Phenomena... Page 39 of 42 42

μ = σ 2N · μ + w · κ N · σ 2N − σ N .

In the renormalized coordinates the parameter interval is [μ0 , 2]. We write


[μ0 , 2] = Jr = Jr, (7.1)
r ≥r0

with Jr, disjoint and |Jr, | = r12 |Jr |.

We do a similar decomposition of [μ0 , 2].

[μ0 , 2] = Jr = Jr, (7.2)
r ≥r0

To each of the decomposition’s (7.1) and (7.2) we get two decompositions



[ã0,0 , ã0 ] = ωr = ωr , , (7.3)
r ≥r0

and


[ã0,0 , ã0 ] = ωr = ωr, . (7.4)
r ≥r0

There is a uniform distorsion bound for the parameter maps a  → ν  and a  → ν 

For each ω = ωr , we do the parameter selection to create a strange attractor as in
Mora and Viana (1993).
Let n be an essential free return time and let E n (z 0 ) be the set kept at time n where
we take into account the parameter deletions because of the (BA) conditions and the
large deviation estimate. By formula (1) in Sect. 12 in Mora and Viana (1993) the
measure of the deleted set ω\E n (ω) satisfies

m(ω\E n (ω)) ≤ B0 e−α0 n m(ω), (7.5)

where B0 and α0 depend on K , α, β and δ but not on N and b.

We define
 

E n (ω) = E n−1 (ω)\ (ω\E n (z 0 )) . (7.6)
z0

The number of critical points is

 θn
K
#Cn ≤ 4
ρ0

by Sect. 12 in Mora and Viana (1993).

123
 42 Page 40 of 42 M. Benedicks, L. Palmisano

This leads to
 θn
K
m(E n−1 \E n ) ≤ 4B0 4 e−α0 n m(ω)
ρ0
  n
K θ −α0
= 4B0 e m(ω)
ρ0
≤ 4B0 e−α0 n/2 m(ω)

and
⎛ ⎞

m⎝ (E n−1 \E n )⎠ ≤ 4B0 e−α0 n/2 m(ω). (7.7)
n≥N n≥N

This means that for N sufficiently large only a small proportion of ω will be deleted.
We now turn to the construction of simultaneous attractors.
The first attractor will be constructed as above and the second attractor will be
chosen corresponding to intervals ω = ωr, .
We now prove the coexistence of two attractors. To the critical point z 0 there corre-
sponds a parameter interval 0 = [a0,0  , a  ] and to z̃ there corresponds the parameter
0 0
interval 0 = [a0,0 , a0 ]. The subintervals ωr  ,  and ωr ,  will have intersections for
  

suitable chosen r  ,  , r  ,  , even for a number of adjacent (r  ,  ) and ωr ,  . Because

of the estimate (7.7) the corresponding sets Er  ,  and Er ,  will have a nonempty
intersection.
We also have to verify that the two attractors are distinct. Suppose that U is a
trapping region for f m 1 i.e.

f m 1 (U ) ⊂ U ,

and correspondingly V is a trapping region for f m 2 , i.e.

f m 2 (V ) ⊂ V .

Then the two trapping cycles are

& 'm 1 −1
f i (U ) i=0 ,
& i 'm 2 −1
f (V ) j=0

The crucial point of the construction is that V is chosen to avoid

( )m 1 −1
f i (U )
i=0

123
Coexistence Phenomena... Page 41 of 42 42

This is possible because of the combinatorial structure of the leafs of W s on which

( )m 1 −1
f i (U )
i=0

is located. As a consequence, V ∩ f i (U ) = ∅ for all i ≥ 0. In fact if i ≥ m 1 ,

then there exists k ≥ 1 and j = 0, . . . m 1 − 1 such that i = j + km 1 . Hence,
V ∩ f i (U ) ⊂ V ∩ f j (U ) = ∅. It is left to prove that f j (V )∩ f i (U ) = ∅ for all i, j ≥ 1.
By contradiction, suppose that f j (V ) ∩ f i (U ) = W = ∅. Then f m 2 − j (W ) ⊂
V ∩ f m 2 − j+i (U ) = ∅ which contradicts that V avoids
( )m 1 −1
f i (U ) .
i=0

It follows that the two trapping orbits can be chosen to be disjoint.

We can now finish the proof of the main theorem of the section.
Proof of Theorem 1.6. Consider a b interval (b1 , b2 ), b2 > b1 > 0, and b2 suffi-
ciently small. For each b ∈ (b1 , b2 ) there is a set E b of positive Lebesgue measure so
that there are two strange attractors and the result follows by Fubini’s theorem.
Acknowledgements The first author was supported by the Swedish Research Council Grant 2016-05482.
The second author was supported by the Trygger Foundation, Project CTS 17:50 and the research was
partially supported by the NSF grant 1600554 and the IMS at Stony Brook University. The authors would
like to thank P. Berger, L. Carleson and J-P Eckmann for helpful discussions. The project was initiated at
Institute Mittag-Leffler during the program Fractal Geometry and Dynamics, September 04–December 15,
2017.

Funding Open access funding provided by Royal Institute of Technology.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence,
and indicate if changes were made. The images or other third party material in this article are included
in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If
material is not included in the article’s Creative Commons licence and your intended use is not permitted
by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

References
Benedicks, M., Carleson, L.: On iterations of 1 − ax 2 on (−1, 1). Ann. Math. (2) 122(1), 1–25 (1985)
Benedicks, M., Carleson, L.: The dynamics of the Hénon map. Ann. Math. (2) 133(1), 73–169 (1991)
Benedicks, M., Viana, M.: Solution of the basin problem for Hénon like attractors. Invent. math. 143,
375–434 (2001)
Benedicks, M., Young, L.-S.: Sinai–Bowen–Ruelle measures for certain Hénon map. Invent. math. 112,
541–576 (1993)
Benedicks, M., Young, L.-S.: Markov extensions and decay of correlations for certain Hénon maps,
Astérisque, 261, xi, 13-56 (2000)
Benedicks, M., Martens, M., Palmisano, L.: Newhouse laminations (2023). arXiv:1811.00617
Berger, P.: Abundance of non-uniformly hyperbolic Hénon like endomorphisms. Astérisque 410, 53–177
(2019)
Berger, P.: Zoology in the Hénon family: twin babies and Milnor’s swallows (2023). arXiv:1801.05628

123
 42 Page 42 of 42 M. Benedicks, L. Palmisano

Curry, J.H.: On the Hénon transformation. Comm. Math. Phys. 68(2), 129–140 (1979)
Gavrilov, N.K., Silnikov, L.P.: On the three dimensional dynamical system close to a system with a struc-
turally unstable homoclinic curve. I. Math. USSR Sbornik, 17,: 467–485. II. Math USSR Sbornik
19(1972), 139–156 (1972)
Gonchenko, S., Shilnikov, L., Turaev, D.: On dynamical properties of multidimensional diffeomorphisms
from Newhouse regions: I. Nonlinearity 21, 923–972 (2008)
Hénon, M.: A two dimensional mapping with a strange attractor. Comm. Math. Phys. 50(1), 66–77 (1976)
Mora, L., Viana, M.: Abundance of strange attractors. Acta Math. 171, 1–71 (1993)
Newhouse, S.: Diffeomorphisms with infinitely many sinks. Topology 13, 9–18 (1974)
Newhouse, S.: The abundance of wild hyperbolic sets and non-smooth stable sets for diffeomorphisms.
Publ. Math. IHES 50, 101–151 (1979)
Palis, J. Jr., de Melo, W., Manning, A.K.: Geometric theory of dynamical systems an introduction (1982)
Palis, J., Jr.: A global view of dynamics and a conjecture on the denseness of finitude of attractors. Astérisque
261, 335–347 (2000)
Palis, J., Takens, F.: Hyperbolicity and sensitive chaotic dynamics at homoclinic bifurcations. Camb. Stud.
Adv. Math. 35, x+234 (1993)
Palmisano, L.: Laminations of coexisting attractors, Annali della Scuola Superiore di Pisa (2023). https://
doi.org/10.2422/2036-2145.202106_020
Robinson, C.: Bifurcation to infinitely many sinks. Comm. Math. Phys. 90(3), 433–459 (1983)
Thunberg, H.: Unfolding of chaotic unimodal maps and the parameter dependence of natural measures.
Nonlinearity (2) 14, 323–337 (2001)
Ures, R.: On the approximation of Hénon-like attractors by homoclinic tangencies. Ergodic Theory Dyn.
Syst. 15, 1223–1229 (1995)
Wang, Q., Young, L.-S.: Strange attractors with one direction of instability. Comm. Math. Phys. 218, 1–97
(2001)
Wang, Q., Young, L.-S.: Towards a theory of rank one attractors. Ann. Math. 167, 349–480 (2008)

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.

123