##
**Induced expansion for quadratic polynomials.**
*(English)*
Zbl 0867.58048

The present paper deals with a class \(\mathcal F\) of maps defined as follows. A map \(f:[-1,1]\to [-1,1]\) is said to be unimodal if \(f(-1)=-1\) and there is an orientation-reversing diffeomorphism \(h:[0,1]\to [-1,a]\) (with \(0<a<1\)) such that \(f(x)=h(x^2)\). Further, let \(\eta>0\). We say that a unimodal map \(f(x)=h(x^2)\) belongs to the class \({\mathcal F}_\eta\) if \(h^{-1}\) has a conformal extension which maps the upper half-plane into the lower half-plane, and \(h\) has a real analytic extension to some open interval \(U\supset [0,1]\) as a diffeomorphism onto \((-1-\eta,1+\eta)\). Finally we define \({\mathcal F}=\bigcup_{\eta>0}{\mathcal F}_\eta\). All maps from \({\mathcal F}\) have non-positive Schwarzian derivative. In particular, \(\mathcal F\) includes the family of quadratic polynomials \(f_a(x)=a-(a+1)x^2\).

The main results of the paper are Theorems 1 and 2 below.

Theorem 1. Let \(f\in {\mathcal F}_\eta\) be non-renormalizable (that is, there is no proper subinterval \(I\) of \([-1,1]\) containing 0 with the properties \(f^n(I)\subset I\) and \(f^n(\partial I)\subset \partial I\) for some \(n>1\)). Then \(f\) is expansion inducing, that is, there are an open, dense and having full measure subset \(O\) of \((-q,q)\) and a map \(F:O\to (1-q,q)\) satisfying the following properties (here \(q\) is the only fixed point of \(f\) in \([0,1]\)):

(i) If \(O_i\) is a connected component of \(O\) then \(F|_{O_i}=f^{n_i}|_{O_i}\) for some \(n_i\) and \(F(O_i)=(1-q,q)\).

(ii) \(F\) is expanding and has bounded distortion, which means that there are constants \(K>1\) and \(D>0\) such that \(|F'(x)|>K\) and \(|(\log F')'(x)|<D\) for any \(x\in O\). Moreover, \(D\) depends only on \(\eta\).

Theorem 2. Let \(f\in {\mathcal F}_\eta\) be renormalizable and let \(I\ni 0\) be a maximal subinterval of \([-1,1]\) such that \(f^n(I)\subset I\) and \(f^n(\partial I)\subset \partial I\) for some \(n>1\). Define a point \(x\in [-1,1]\) to be almost parabolic with period \(m\) and depth \(k\) provided that the derivative of \(f^m\) at \(x\) is one, \(f^m\) is monotone between \(x\) and \(0\), and \(k\) consecutive images \(f^m(0),\dots,f^{km}(0)\) are between \(x\) and \(0\), and denote by \(k(n)\) the maximum of depths of almost parabolic points with periods less than \(n\). Specify a number \(D>0\). Then, for every given \(k\), there is a number \(N(\eta,D,k)\) not depending on \(f\) so that if \(n>N(\eta,D,k)\) and \(k(n)\leq k\), then \(f^n|_I\) is conjugate to a map from \({\mathcal F}_D\) via a linear transformation.

Theorem 1 implies that if \(f\in {\mathcal F}\) is non-renormalizable then the \(\omega\)-limit set of almost all points from \([-1,1]\) is the whole interval \([f^2(0),f(0)]\); this result was also previously proved (in the more general setting of maps with non-positive Schwarzian derivative whose only critical point \(c\) satisfies \(f''(c)\neq 0\)) by M. Lyubich [Ann. Math., II. Ser. 140, No. 2, 347-404 (1994; Zbl 0821.58014)]. Theorem 2 is used by the second author [‘Hyperbolicity is dense in the real quadratic family’, to appear in Ann. Math.], where it is shown that the set of parameters \(a\) such that the polynomial \(f_a(x)=a-(a+1)x^2\) has a hyperbolic periodic attractor is dense in \([-1,1]\). Both Theorems 1 and 2 are consequences of a technical Theorem C, where, roughly speaking, so-called “decaying of box geometry” is demonstrated for a special type of maps called box mappings. The proof uses a mixture of real and complex tools.

While the significance of the results of the paper is obvious, in the reviewer’s opinion they are presented in a somewhat involved way (besides, there are some misprints: in the paper it is written “upper half-plane” instead of “lower half-plane” in the definition of \({\mathcal F}_\eta\), and \(1/2\) instead of \(0\) in Theorem 2). Hence an interested but non-truly specialist reader may also wish to check the introduction to M. Jakobson and G. Świątek [Ergodic Theory Dyn. Syst. 14, No. 4, 721-755 (1994; Zbl 0830.58019)] (which is generalized by the paper under review) and Chapter V from the monograph by W. de Melo and S. van Strien [‘One-dimensional dynamics’. Berlin: Springer-Verlag (1993; Zbl 0791.58003)].

The main results of the paper are Theorems 1 and 2 below.

Theorem 1. Let \(f\in {\mathcal F}_\eta\) be non-renormalizable (that is, there is no proper subinterval \(I\) of \([-1,1]\) containing 0 with the properties \(f^n(I)\subset I\) and \(f^n(\partial I)\subset \partial I\) for some \(n>1\)). Then \(f\) is expansion inducing, that is, there are an open, dense and having full measure subset \(O\) of \((-q,q)\) and a map \(F:O\to (1-q,q)\) satisfying the following properties (here \(q\) is the only fixed point of \(f\) in \([0,1]\)):

(i) If \(O_i\) is a connected component of \(O\) then \(F|_{O_i}=f^{n_i}|_{O_i}\) for some \(n_i\) and \(F(O_i)=(1-q,q)\).

(ii) \(F\) is expanding and has bounded distortion, which means that there are constants \(K>1\) and \(D>0\) such that \(|F'(x)|>K\) and \(|(\log F')'(x)|<D\) for any \(x\in O\). Moreover, \(D\) depends only on \(\eta\).

Theorem 2. Let \(f\in {\mathcal F}_\eta\) be renormalizable and let \(I\ni 0\) be a maximal subinterval of \([-1,1]\) such that \(f^n(I)\subset I\) and \(f^n(\partial I)\subset \partial I\) for some \(n>1\). Define a point \(x\in [-1,1]\) to be almost parabolic with period \(m\) and depth \(k\) provided that the derivative of \(f^m\) at \(x\) is one, \(f^m\) is monotone between \(x\) and \(0\), and \(k\) consecutive images \(f^m(0),\dots,f^{km}(0)\) are between \(x\) and \(0\), and denote by \(k(n)\) the maximum of depths of almost parabolic points with periods less than \(n\). Specify a number \(D>0\). Then, for every given \(k\), there is a number \(N(\eta,D,k)\) not depending on \(f\) so that if \(n>N(\eta,D,k)\) and \(k(n)\leq k\), then \(f^n|_I\) is conjugate to a map from \({\mathcal F}_D\) via a linear transformation.

Theorem 1 implies that if \(f\in {\mathcal F}\) is non-renormalizable then the \(\omega\)-limit set of almost all points from \([-1,1]\) is the whole interval \([f^2(0),f(0)]\); this result was also previously proved (in the more general setting of maps with non-positive Schwarzian derivative whose only critical point \(c\) satisfies \(f''(c)\neq 0\)) by M. Lyubich [Ann. Math., II. Ser. 140, No. 2, 347-404 (1994; Zbl 0821.58014)]. Theorem 2 is used by the second author [‘Hyperbolicity is dense in the real quadratic family’, to appear in Ann. Math.], where it is shown that the set of parameters \(a\) such that the polynomial \(f_a(x)=a-(a+1)x^2\) has a hyperbolic periodic attractor is dense in \([-1,1]\). Both Theorems 1 and 2 are consequences of a technical Theorem C, where, roughly speaking, so-called “decaying of box geometry” is demonstrated for a special type of maps called box mappings. The proof uses a mixture of real and complex tools.

While the significance of the results of the paper is obvious, in the reviewer’s opinion they are presented in a somewhat involved way (besides, there are some misprints: in the paper it is written “upper half-plane” instead of “lower half-plane” in the definition of \({\mathcal F}_\eta\), and \(1/2\) instead of \(0\) in Theorem 2). Hence an interested but non-truly specialist reader may also wish to check the introduction to M. Jakobson and G. Świątek [Ergodic Theory Dyn. Syst. 14, No. 4, 721-755 (1994; Zbl 0830.58019)] (which is generalized by the paper under review) and Chapter V from the monograph by W. de Melo and S. van Strien [‘One-dimensional dynamics’. Berlin: Springer-Verlag (1993; Zbl 0791.58003)].

Reviewer: V.Jiménez López (Murcia)

### MSC:

37D99 | Dynamical systems with hyperbolic behavior |

37F99 | Dynamical systems over complex numbers |

37A99 | Ergodic theory |

37C70 | Attractors and repellers of smooth dynamical systems and their topological structure |

### References:

[1] | A. BLOKH and M. LYUBICH , Non-existence of wandering intervals and structure of topological attractors for one dimensional dynamical systems (Erg. Th. and Dyn. Sys., Vol. 9, 1989 , pp. 751-758). MR 91e:58101 | Zbl 0665.58024 · Zbl 0665.58024 · doi:10.1017/S0143385700005319 |

[2] | B. BRANNER and J. H. HUBBARD , The iteration of cubic polynomials , Part II : patterns and parapatterns (Acta Math., Vol. 169, 1992 , pp. 229-325). MR 94d:30044 | Zbl 0812.30008 · Zbl 0812.30008 · doi:10.1007/BF02392761 |

[3] | A. DOUADY and J. H. HUBBARD , On the dynamics of polynomial-like mappings (Ann. Sci. Ec. Norm. Sup. (Paris), Vol. 18, 1985 , pp. 287-343). Numdam | MR 87f:58083 | Zbl 0587.30028 · Zbl 0587.30028 |

[4] | J. GRACZYK , Ph. D. Thesis (Mathematics Department of Warsaw University ( 1990 ) ; also : Dynamics of non-degenerate upper maps , preprint of Queen’s University at Kingston, Canada, 1991 ). |

[5] | J. GRACZYK and G. ŚWIATEK , Critical circle maps near bifurcation (Stony Brook IMS preprint, 1991 , Proposition 2). |

[6] | J. GUCKENHEIMER , Limit sets of S-unimodal maps with zero entropy (Commun. Math. Phys., Vol. 110, 1987 , pp. 655-659). Article | MR 88i:58111 | Zbl 0625.58027 · Zbl 0625.58027 · doi:10.1007/BF01205554 |

[7] | J. GUCKENHEIMER and S. JOHNSON , Distortion of S-unimodal maps (Annals of Math., Vol. 132, 1990 , pp. 71-130). MR 91g:58157 | Zbl 0708.58007 · Zbl 0708.58007 · doi:10.2307/1971501 |

[8] | F. HOFBAUER , F. and G. KELLER , Some remarks about recent results on S-unimodal maps (Annales de l’Institut Henri Poincaré, Physique Théorique, Vol. 53, 1990 , pp. 413-425). Numdam | Zbl 0721.58018 · Zbl 0721.58018 |

[9] | M. JAKOBSON , Absolutely continuous invariant measures for one-parameter families of one-dimensional maps (Commun. Math. Phys., Vol. 81, 1981 , pp. 39-88). Article | MR 83j:58070 | Zbl 0497.58017 · Zbl 0497.58017 · doi:10.1007/BF01941800 |

[10] | M. JAKOBSON and G. ŚWIATEK , Metric properties of non-renormalizable S-unimodal maps (preprint IHES, no. IHES/M/91/16, 1991 ). |

[11] | M. JAKOBSON and G. ŚWIATEK , Quasisymmetric conjugacies between unimodal maps (Stony Brook preprint, Vol. 16, 1991 ). |

[12] | G. KELLER and T. NOWICKI , Fibonacci maps revisited (manuscript, 1992 ). |

[13] | O. LEHTO and K. VIRTANEN , Quasikonforme Abbildungen (Springer-Verlag, Berlin-Heidelberg-New York, 1965 ). MR 32 #5872 | Zbl 0138.30301 · Zbl 0138.30301 |

[14] | M. LYUBICH , Milnor’s attractors, persistent recurrence and renormalization , in (Topological methods in modern mathematics, Publish or Perish, Inc., Houston TX, 1993 ). MR 94e:58082 | Zbl 0797.58050 · Zbl 0797.58050 |

[15] | M. LYUBICH and J. MILNOR , The dynamics of the Fibonacci polynomial (Jour. of the AMS, Vol. 6, 1993 , pp. 425-457). MR 93h:58080 | Zbl 0778.58040 · Zbl 0778.58040 · doi:10.2307/2152804 |

[16] | M. MARTENS , Ph. D. thesis (Math. Department of Delf University of Technology, 1990 ; also : IMS preprint, Vol. 17, 1992 ). |

[17] | J. MILNOR , The Yoccoz theorem on local connectivity of Julia sets . A proof with pictures (class notes, Stony Brook, 1991 - 1992 ). |

[18] | W. DE MELO and S. VAN STRIEN , One-Dimensional Dynamics (Springer-Verlag, New York, 1993 ). MR 95a:58035 | Zbl 0791.58003 · Zbl 0791.58003 |

[19] | C. PRESTON , Iterates of maps on an interval (Lecture Notes in Mathematics, Vol. 999, Berlin, Heidelberg, New York : Springer, 1983 ). MR 85c:58058 | Zbl 0582.58001 · Zbl 0582.58001 |

[20] | D. SULLIVAN , Bounds, quadratic differentials and renormalization conjectures (to appear in American Mathematical Society Centennial Publications, Vol. 2, American Mathematical Society, Providence, R.I., 1991 ). Zbl 0936.37016 · Zbl 0936.37016 |

[21] | G. ŚWIATEK , Hyperbolicity is dense in the real quadratic family (preprint Stony Brook, 1992 ). |

[22] | O. TEICHMÜLLER , Untersuchungen über konforme und quasikonforme Abbildung (Deutsche Mathematik, Vol. 3, pp. 621-678). Zbl 0020.23801 | JFM 64.0313.06 · Zbl 0020.23801 |

[23] | J.-C. YOCCOZ , unpublished results. |

[24] | J. J. P. VEERMAN and F. M. TANGERMAN , Scalings in circle maps (1) (Commun. in Math. Phys., Vol. 134, 1990 , pp. 89-107). Article | MR 92h:58106 | Zbl 0723.58026 · Zbl 0723.58026 · doi:10.1007/BF02102091 |

This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.