# On buried points and phase transition points in the Julia sets concerning renormalization transformation

- Junyang Gao
^{1}Email author

**2014**:239

https://doi.org/10.1186/1687-1847-2014-239

© Gao; licensee Springer. 2014

**Received: **4 April 2014

**Accepted: **13 August 2014

**Published: **11 September 2014

## Abstract

Considering a family of rational maps ${T}_{n\lambda}$ concerning renormalization transformation, we give a perfect description of buried points and phase transition points in the Julia set $J({T}_{n\lambda})$. Furthermore, we prove that $J({T}_{n\lambda})$ contains an open interval where all points are buried points for some parameters *n* and *λ*, which is according to the problem that Curry and Mayer proposed.

**MSC:**37F10, 37F45.

## Keywords

## 1 Introduction

with two parameters $n\in \mathbb{N}$ ($n>1$) and $\lambda \in \mathbb{R}$. This physical model can be derived from the limit distribution of zeroes of the partition function in the famous Yang-Lee theory [8–11]. In fact, some interesting relationships among the phase transitions, the critical exponents, the critical amplitudes and the shape of the Julia sets were found in [12]. After this, many works have been devoted to the Julia sets of the renormalization transformation [13–17].

In [14], Qiao proved that ${\beta}_{n}\in (1,2)$ when *n* is an even number and ${\beta}_{n}\in (2,3)$ when *n* is an odd number. Moreover, he also proved that the equation ${(t-1)}^{2n-1}-2{(t-1)}^{n}-t+3=0$ has one unique real solution ${\gamma}_{n}\in (2,3]$ when *n* is an even number. Furthermore, in this paper we prove the following.

**Theorem 1**

*Suppose that*${T}_{n\lambda}$

*is defined as*(1),

*we have*

- (1)
*if**n**is an even number*,*then*$J({T}_{n\lambda})$*contains buried points if and only if*$\lambda \in ({\beta}_{n},{\gamma}_{n})$; - (2)
*if*$n=3$,*then*$J({T}_{n\lambda})$*contains buried points if and only if*$\lambda \in [3-\sqrt{2},{\beta}_{n}]$; - (3)
*if*$n>3$,*then*$J({T}_{n\lambda})$*contains buried points if and only if*$\lambda \in [{\lambda}_{\alpha},{\beta}_{n}]$,*where*${\lambda}_{\alpha}\in (1,2)$*is an absolute constant*.

In [7], Curry and Mayer proposed some questions about buried points. One of these questions is the following:

Is there the set of buried points whose components are not either points, circles, or the irrational points of the Sierpinski curve?

In this paper, considering the Julia sets $J({T}_{n\lambda})$, we prove that the set of buried points may contain an open interval of the real axis ℝ. In fact, we have the following.

**Theorem 2**

*Suppose that*$n>3$

*is an odd number*,

*then there exists*$\lambda \in (1,2)$

*such that*$J({T}_{n\lambda})$

*contains an open interval in which all points are buried points*(

*see Figure*1).

For this physical model, it is well known that there is a specific physical significance about the intersection points between the Julia set $J({T}_{n\lambda})$ and the positive real axis ${\mathbb{R}}^{+}$ when *λ* is a positive integer. In fact, these intersection points are the locations of phase transitions of this model. A natural problem is how to distribute on ${\mathbb{R}}^{+}$ the locations of phase transitions? In the following we will show that the number of this kind of intersection points may be 1, 2, 3, or ∞. Denote by $\mathrm{\#}(J({T}_{\lambda})\cap {\mathbb{R}}^{+})$ the number of the intersection points of $J({T}_{\lambda})$ and ${\mathbb{R}}^{+}$. Then we have the following.

**Theorem 3**

*If*

*λ*

*is a positive integer*,

*then*

- (1)
$\mathrm{\#}(J({T}_{n\lambda})\cap {\mathbb{R}}^{+})=\mathrm{\infty}$

*for any even number*$n\ge 2$*and*$\lambda =1$ (*see Figure*2); - (2)
$\mathrm{\#}(J({T}_{n\lambda})\cap {\mathbb{R}}^{+})=3$

*for any odd number*$n\ge 3$*and*$\lambda =1$ (*see the left figure in Figure*3); - (3)
$\mathrm{\#}(J({T}_{n\lambda})\cap {\mathbb{R}}^{+})=2$

*for any positive*$n\ge 2$*and*$\lambda \in [2,2+\frac{1}{\sqrt[n-1]{2}-1})$ (*see the immediate figure in Figure*3); - (4)
$\mathrm{\#}(J({T}_{n\lambda})\cap {\mathbb{R}}^{+})=1$

*for any positive*$n\ge 2$*and*$\lambda \in [2+\frac{1}{\sqrt[n-1]{2}-1},+\mathrm{\infty})$ (*see the right figure in Figure*3).

## 2 Some notations and preliminary results

Let *f* be a rational map with degree $d\ge 2$ from a complex sphere $\overline{\mathbb{C}}$ to itself. The notation ${f}^{k}$ means the *k* th iteration of *f*. A point *z* is called a *critical point* if ${f}^{\prime}(z)=0$. A point *z* is called a *periodic point* if ${f}^{k}(z)=z$ for some $k\ge 1$, the minimal of such *k* is called the *period* of *z*. For a periodic point ${z}_{0}$ with periodic *k*, the *multiplier* of ${z}_{0}$ is defined as ${({f}^{k})}^{\prime}({z}_{0})$. The periodic point ${z}_{0}$ is either *attracting*, *indifferent*, or *repelling* according to $|{({f}^{k})}^{\prime}({z}_{0})|<1$, $|{({f}^{k})}^{\prime}({z}_{0})|=1$, or $|{({f}^{k})}^{\prime}({z}_{0})|>1$. The *Julia set*, denoted by $J(f)$, is the closure of repelling periodic points. Its complement is called the *Fatou set*, denoted by $F(f)$. A component *D* of $F(f)$ is called completely invariant if $f(D)={f}^{-1}(D)=D$. Moreover, if *D* is a completely invariant component, then $J(f)=\partial D$. Let $P(f)$ be the *post-critical set* of *f*, *i.e.*, the closure of the forward orbits of critical points. It plays a crucial rule in the study of complex dynamics. For the classical results in complex dynamics, see [18, 19] and [20].

In order to prove our theorems, we need the following lemmas.

*Let* *f* *be a rational map* *f* *with* $degf\ge 2$ *and* $J(f)$ *connected and locally connected*, *then* $J(f)$ *contains buried points if and only if* ${f}^{2}$ *has no completely invariant Fatou component*.

**Lemma 2** ([17])

*If* $J({T}_{n\lambda})$ *is connected*, *then it is locally connected*.

**Lemma 3** ([17])

*If* $n\ge 3$ *is an odd integer*, *then* ${T}_{n\lambda}$ *has only four real fixed points* ${q}_{1},{q}_{2},1,{q}_{3}$ (${q}_{1}<{q}_{2}\le 0$, ${q}_{3}>1$) *for* $\lambda \in [1,2)$.

**Lemma 4** ([17])

*Suppose that*${q}_{2}$

*is the fixed point stated in Lemma*3.

*If*$n=3$

*and*$\lambda \in (1,2)$,

*then*${q}_{2}$

*is an attracting fixed point*.

*If*$n>3$

*is an odd integer*,

*then there exist two numbers*${\lambda}_{\alpha}$

*and*${\lambda}_{\beta}$ (${\lambda}_{\alpha}<{\lambda}_{\beta}$)

*in*$(1,2)$

*such that*

- (1)
${q}_{2}$

*is an attracting fixed point for*$\lambda \in (1,{\lambda}_{\alpha})\cup ({\lambda}_{\beta},2)$; - (2)
${q}_{2}$

*is a parabolic fixed point and*${T}_{n\lambda}^{\prime}({q}_{2})=-1$*for*$\lambda ={\lambda}_{\alpha}$*or*$\lambda ={\lambda}_{\beta}$; - (3)
${q}_{2}$

*is a repelling fixed point for*$\lambda \in ({\lambda}_{\alpha},{\lambda}_{\beta})$.

- (1)
*If*$n\ge 2$*is an even integer*,*then*${T}_{n\lambda}^{2}$*contains a completely invariant Fatou component if and only if*$\lambda \in (-\mathrm{\infty},{\beta}_{n}]\cup ({\gamma}_{n},+\mathrm{\infty})$.*Furthermore*, $J({T}_{n\lambda})$*is connected for*$\lambda \in ({\beta}_{n},{\gamma}_{n})$. - (2)
*If*$n>1$*is an odd integer*,*then*${T}_{n\lambda}^{2}$*contains a completely invariant Fatou component for*$\lambda \in (-\mathrm{\infty},1]\cup ({\beta}_{n},+\mathrm{\infty})$.*Furthermore*, ${T}_{n\lambda}^{2}$*contains at most three Fatou periodic*(*not completely invariant*)*cycles for*$\lambda \in [2,{\beta}_{n}]$.

## 3 Proof of Theorem 1

So ${T}_{n\lambda}$ has only four critical values 1, ∞, 0 and ${T}_{n\lambda}(-\lambda +1)={(-\lambda +1)}^{n}$.

By Lemma 1, Lemma 2 and Lemma 5, it is easy to see that $J({T}_{n\lambda})$ contains buried points if and only if $\lambda \in ({\beta}_{n},{\gamma}_{n})$ for even integer *n*. Similarly, if $n>1$ is an odd number, $J({T}_{n\lambda})$ contains no buried point for $\lambda \in (-\mathrm{\infty},1]\cup ({\beta}_{n},+\mathrm{\infty})$, $J({T}_{n\lambda})$ contains buried points for $\lambda \in [2,{\beta}_{n}]$. Hence we need only to investigate the case that $n>1$ is an odd integer and $\lambda \in (1,2)$.

respectively.

- (I)If $n=3$ and $\lambda \in (1,2-\frac{1}{\sqrt[n-1]{2}+1})$. It is easy to verify that $({q}_{1},{q}_{1}^{\ast})\subset {A}_{\lambda}({q}_{2})$ since ${T}_{3\lambda}(x)>x$ for $x\in ({q}_{1},{q}_{2})$. Hence $0\in {A}_{\lambda}({q}_{2})$. Next we define$\mathrm{\Gamma}(\lambda )=\{z:|z+\frac{\lambda}{2}-1|=\frac{\lambda}{2}\}\cap \{z:\mathbf{Re}\mathbf{z}\le \mathbf{0}\}.$

Note that $1-\lambda \in \mathrm{\Gamma}(\lambda )\cap ({q}_{1},{q}_{1}^{\ast})$, then $\mathrm{\Gamma}(\lambda )\subset {A}_{\lambda}({q}_{2})$. Since ${T}_{3\lambda}^{-1}(0)=\pm \sqrt{\lambda -1}i\in \mathrm{\Gamma}(\lambda )$, then ${A}_{\lambda}({q}_{2})$ is completely invariant. This implies that $J({T}_{3\lambda})=\partial {A}_{\lambda}({q}_{2})$ and thus $J({T}_{3\lambda})$ contains no buried point.

- (II)If $n\ge 5$. Let ${\lambda}_{\alpha}$ and ${\lambda}_{\beta}$ be two numbers stated in Lemma 4. Below we prove${\lambda}_{\alpha}<2-\frac{1}{\sqrt[n-1]{2}+1}$(7)

By the proof of Proposition 2 in [17], we know that $\varphi (t)$ is monotone decreasing from $\varphi (-1)=2$ to $\varphi (0)=1$ on $(-1,0)$. So we conclude that ${\lambda}_{\alpha}<{\lambda}_{{t}_{0}}$ since ${q}_{2}$ is a parabolic fixed point of ${T}_{n{\lambda}_{\alpha}}$ by Lemma 4.

*i.e.*, ${t}_{\ast}<\frac{2{n}^{2}-4n+3}{2({n}^{2}-1)}$. By a calculation, we have

So ${\lambda}_{{t}_{0}}<\frac{3}{2}$ and thus (7) is obvious.

Below we prove (8).

We get (8).

If $\lambda \in (1,{\lambda}_{\alpha})$. By (5) and (7), we know ${q}_{1}^{\ast}>0$. By the same discussion as used in the case $n=3$, we can deduce that ${A}_{\lambda}({q}_{2})$ is completely invariant for $\lambda \in (1,{\lambda}_{\alpha})$. This implies that $J({T}_{n\lambda})=\partial {A}_{\lambda}({q}_{2})$ and thus $J({T}_{n\lambda})$ contains no buried point.

If $\lambda \in [{\lambda}_{\alpha},{\lambda}_{\beta}]$, it is easy to see that $F({T}_{n\lambda})$ contains at most another *k*-periodic cycle except for ${A}_{\lambda}(1)$ and ${A}_{\lambda}(\mathrm{\infty})$. By (3) and (4), we can deduce that $k\ge 2$ since ${q}_{2}$ is a repelling fixed point by Lemma 4 (in fact, the periodic Fatou components lie on both sides of ${q}_{2}$). Then ${T}_{n\lambda}^{2}$ cannot contain a completely invariant Fatou component. This shows that $J({T}_{n\lambda})$ contains buried points.

If $\lambda \in ({\lambda}_{\beta},2)$. By (5), ${q}_{1}^{\ast}<0$. By (3), ${T}_{n\lambda}(0)\in {A}_{\lambda}(\mathrm{\infty})$. It shows that ${A}_{\lambda}({q}_{2})$ is not completely invariant since it contains only one critical point $z=1-\lambda $ and $deg({T}_{n\lambda}^{k}{|}_{{A}_{\lambda}({q}_{2})})=2$, and thus $J({T}_{n\lambda})$ contains buried points. It completes the proof of Theorem 1.

## 4 Proof of Theorem 2

In fact, let Γ be the circle centered at $1-\frac{1}{2}\lambda $ with radius $\frac{1}{2}\lambda $. Then Γ passes through critical points 1, $1-\lambda $ and $\pm i\sqrt{\lambda -1}$, and ${T}_{n\lambda}(\mathrm{\Gamma})=[{(1-\lambda )}^{n},1]\subset \overline{Int(\mathrm{\Gamma})}$. We conclude that ${A}_{\lambda}(\mathrm{\infty})\cap \overline{Int(\mathrm{\Gamma})}=\mathrm{\varnothing}$ and $\partial {A}_{\lambda}(\mathrm{\infty})\cap [1-\lambda ,1]=\mathrm{\varnothing}$. For $x\in ({q}_{1},1-\lambda )$, there exists a positive integer *k* such that ${T}_{n\lambda}^{k}(x)\in (1-\lambda ,{(1-\lambda )}^{n}]$. This implies $x\notin \partial {A}_{\lambda}(\mathrm{\infty})$. Hence $\partial {A}_{\lambda}(\mathrm{\infty})\cap \overline{\mathbb{R}}=\{{q}_{1},{q}_{3}\}$.

Let ${\mathrm{\Omega}}_{0}$ be the component of $\mathbb{C}\setminus {T}_{n\lambda}^{-1}([-\mathrm{\infty},0])$ containing 1. Then $1-\frac{1}{2}\lambda \in \partial {\mathrm{\Omega}}_{0}$ and ${A}_{\lambda}(1)\subset {\mathrm{\Omega}}_{0}$. Therefore $\overline{{A}_{\lambda}(1)}\cap [{q}_{1},{q}_{1}^{\ast}]=\mathrm{\varnothing}$ and $\partial {A}_{\lambda}(1)\cap \overline{\mathbb{R}}=\{{q}_{3}^{\ast},{q}_{3}\}$.

Obviously, ${T}_{n\lambda}^{-1}({A}_{\lambda}(\mathrm{\infty}))$ has only two components ${A}_{\lambda}(\mathrm{\infty})$ and $D(-\frac{\lambda}{2}+1)$. By ${T}_{n\lambda}:\mathbb{R}\to \mathbb{R}\mathrm{\setminus}({(1-\lambda )}^{n},1)$ and $\overline{D(-\frac{\lambda}{2}+1)}\cap \mathbb{R}=[{q}_{1}^{\ast},{q}_{3}^{\ast}]\subset ({(1-\lambda )}^{n},1)$, we know that $\overline{W}\cap \mathbb{R}=\mathrm{\varnothing}$ for every component *W* of ${T}_{n\lambda}^{-1}(D(-\frac{\lambda}{2}+1))$. Similarly, by (3) and (4), we can deduce that $\overline{{W}_{\ast}}\cap \mathbb{R}=\mathrm{\varnothing}$ for every component ${W}_{\ast}$ of ${T}_{n\lambda}^{-1}({A}_{\lambda}(1))$ except for ${A}_{\lambda}(1)$. Hence the closures of all pull-backs of *W* and ${W}_{\ast}$ are disjoint with ℝ since ${T}_{n\lambda}$ maps ℝ into $\mathbb{R}\cup \{\mathrm{\infty}\}$. By [17], we know that there exist at least two real numbers ${\lambda}_{1},{\lambda}_{2}\in (1,2)$ such that ${T}_{n{\lambda}_{1}}$ and ${T}_{n{\lambda}_{2}}$ are Feigenbaum-like maps. Obviously, ${T}_{n{\lambda}_{1}}$ and ${T}_{n{\lambda}_{2}}$ contain only two Fatou periodic domains ${A}_{\lambda}(1)$ and ${A}_{\lambda}(\mathrm{\infty})$. By the above analysis and (4), we conclude that $({q}_{1},{q}_{1}^{\ast})$ lies in the Julia sets $J({T}_{n{\lambda}_{1}})$ and $J({T}_{n{\lambda}_{2}})$, and each point of this open interval does not belong to the boundaries of any Fatou components. This implies that $({q}_{1},{q}_{1}^{\ast})$ is a buried interval of $J({T}_{n{\lambda}_{1}})$ and $J({T}_{n{\lambda}_{2}})$ (in fact, by a calculation, there exists ${\lambda}_{0}\in (1,2)$ such that $1-{\lambda}_{0}$ is strictly eventually periodic, then $({q}_{1},{q}_{1}^{\ast})$ is also a buried interval of $J({T}_{n{\lambda}_{0}})$). It completes the proof of Theorem 2.

## 5 Proof of Theorem 3

- (1)
If

*n*is an even number. We consider the real fixed points of ${T}_{n1}$,*i.e.*, the set $\{x\in \mathbb{R}|{T}_{n1}(x)={(\frac{{x}^{2}}{2x-1})}^{n}=x\}$. It is easy to see that ${T}_{n\lambda}(x)$ has no fixed point on $(-\mathrm{\infty},0)$ since ${T}_{n\lambda}(x)\ge 0$ for $x\in \mathbb{R}$. By (2), ${T}_{n1}(x)$ is monotone increasing on $(0,\frac{1}{2})$ from 0 to +∞. Note that ${T}_{n1}(\frac{1}{2})=+\mathrm{\infty}$ and $z=1$ is an attracting fixed point, then there exists at least one fixed point $\alpha \in (0,\frac{1}{2})$ such that ${T}_{n1}(x)>x$ for $x\in (0,\alpha )$ and $(0,\alpha )\subset {A}_{\lambda}(0)$.

Next we claim that *α* is the unique fixed point in $(0,\frac{1}{2})$.

Otherwise, if there exists another fixed point *γ* in $(0,\frac{1}{2})$. Without loss of generality, we assume $\alpha <\gamma $, then ${T}_{n1}(x)>x$ or ${T}_{n1}(x)<x$ on $(\alpha ,\gamma )$. Note that ${T}_{n1}(x)$ is monotone increasing on $(0,\frac{1}{2})$, then one of them (*α* or *γ*) is an attracting fixed point or a parabolic fixed point. Since ${T}_{n1}(0)=0$, ${T}_{n1}(\frac{1}{2})=+\mathrm{\infty}$, ${T}_{n1}(1)=1$ and ${T}_{n1}(\mathrm{\infty})=\mathrm{\infty}$, then the immediate basin contains no critical point. But the Sullivan theorem [18] says that each immediate basin contains at least one critical point. It is a contradiction. Hence *α* is the unique fixed point on $(0,\frac{1}{2})$. Similarly, we can deduce that there also exists a unique fixed point *β* in $(1,+\mathrm{\infty})$. Hence ${T}_{n1}$ has only four real fixed points 0, 1, *α*, *β*. *α* and *β* are two repelling fixed points.

*β*is a repelling fixed point.

- (2)If $n>1$ is an odd number. By Lemma 3, we know that ${T}_{n1}$ has only four fixed points ${q}_{1}<0$, ${q}_{2}=0,1$ and ${q}_{3}>1$. By a similar discussion as used in (3) in Theorem 1, we can deduce that there exist only two points ${q}_{1}^{\ast}\in (0,\frac{1}{2})$ and ${q}_{3}^{\ast}\in (\frac{1}{2},1)$ such that ${T}_{n1}({q}_{1}^{\ast})={q}_{1}$ and ${T}_{n1}({q}_{3}^{\ast})={q}_{3}$. Furthermore, we can deduce that$\begin{array}{c}({q}_{1},{q}_{1}^{\ast})\subset {A}_{n1}(0),\phantom{\rule{1em}{0ex}}\phantom{\rule{1em}{0ex}}({q}_{1}^{\ast},{q}_{3}^{\ast})\subset {A}_{\lambda}^{-1}(\mathrm{\infty}),\hfill \\ ({q}_{3},{q}_{3})\subset {A}_{\lambda}(1),\phantom{\rule{1em}{0ex}}\phantom{\rule{1em}{0ex}}({q}_{3},+\mathrm{\infty})\subset {A}_{\lambda}(\mathrm{\infty}).\hfill \end{array}$

*i.e.*, $\mathrm{\#}(J({T}_{n1})\cap {\mathbb{R}}^{+})=3$.

- (3)
We know that ${T}_{n\gamma}$ has only one fixed point $q\in (1,+\mathrm{\infty})$. Obviously,

*q*is a repelling fixed point and $(q,+\mathrm{\infty})\subset {A}_{\lambda}(\mathrm{\infty})$. In what follows we distinguish two cases to discuss. - (a)
If $n\ge 2$ and $\lambda =2$. ${T}_{n2}(z)={(\frac{{z}^{2}+1}{2z})}^{n}$, ${T}_{n2}$ has only six critical points ±1, ±

*i*, 0, ∞ and ${T}_{n2}(\pm 1)=1$, ${T}_{n2}(\pm i)=0$, ${T}_{n2}(0)=\mathrm{\infty}$, ${T}_{n2}(\mathrm{\infty})=\mathrm{\infty}$. Considering the real fixed points of ${T}_{n2}$, ${T}_{n2}(x)$ has no fixed point on $(-\mathrm{\infty},0)$. Note that ${T}_{n\lambda}(x)$ is monotone decreasing on $(0,1)$ from +∞ to 1, then ${T}_{n\lambda}(x)$ has no fixed point on $(0,1)$.

*i.e.*, $\mathrm{\#}(J({T}_{n2})\cap {\mathbb{R}}^{+})=2$.

- (b)
If $n>2$ and $\lambda \in (2,2+\frac{1}{\sqrt[n-1]{2}-1})$. By a similar discussion as (6) in Theorem 1, we can easily deduce that ${T}_{n\lambda}^{2}(0)>{T}_{n\lambda}(0)$. Then ${T}_{n\lambda}(0)>1$. We conclude ${T}_{n\lambda}(0)>q$ since ${T}_{n\lambda}(x)<x$ for $x\in (1,q)$. By (2), ${T}_{n\lambda}(x)$ is monotone decreasing on $(0,1)$. Then there exists a unique point ${q}_{1}\in (0,1)$ such that ${T}_{n1}({q}_{1})=q$. Note that ${q}_{1}\in (0,1)$, then $J({T}_{n\lambda})\cap {\mathbb{R}}^{+}=\{{q}_{1},q\}$,

*i.e.*, $\mathrm{\#}(J({T}_{n\lambda})\cap {\mathbb{R}}^{+})=2$. - (4)
We also distinguish two cases to discuss.

- (c)
If $n=2$ and $\lambda =3$. We know that ${T}_{23}$ has a unique repelling point ${x}_{0}=4$ on $(1,+\mathrm{\infty})$. It is easy to see that ${T}_{23}(x)<x$ for $x\in (1,4)$ and ${T}_{23}(x)>x$ for $x\in (4,+\mathrm{\infty})$. Note that ${T}_{23}(0)=4$ and ${T}_{23}(x)$ is monotone decreasing for $x\in (0,1)$, then ${T}_{23}(x)\in (1,4)$ for $x\in (0,1)$. Then $(0,4)\subset {A}_{\lambda}(1)$ and $(4,+\mathrm{\infty})\subset {A}_{\lambda}(\mathrm{\infty})$. We get $J({T}_{23})\cap {\mathbb{R}}^{+}=\{{x}_{0}\}$,

*i.e.*, $\mathrm{\#}(J({T}_{23})\cap {\mathbb{R}}^{+})=1$. - (d)
If $n>2$ and $\lambda \in (2+\frac{1}{\sqrt[n-1]{2}-1},+\mathrm{\infty})$. By a similar discussion as (6) in Theorem 1, we also get ${T}_{n\lambda}^{2}(0)<{T}_{n\lambda}(0)$. Hence ${T}_{n\lambda}(0)\in (1,q)$ since ${T}_{n\lambda}(0)>1$. Note that ${T}_{n\lambda}(x)$ is monotone decreasing for $x\in [0,1)$, then ${T}_{n\lambda}(x)\in (1,q)$ for $x\in [0,1)$. This means $[0,q)\subset {A}_{\lambda}(1)$. We obtain $J({T}_{n\lambda})\cap {\mathbb{R}}^{+}=\{q\}$,

*i.e.*, $\mathrm{\#}(J({T}_{n\lambda})\cap {\mathbb{R}}^{+})=1$. The proof of Theorem 3 is completed.

## Author’s contributions

The author is entirely responsible for this research. The author read and approved the final manuscript.

## Declarations

### Acknowledgements

The author would like to thank the referees for their valuable suggestions for improving this paper. This work was supported by the National Natural Science Foundation of China (No. 11371363, 11231009, 11261002, 11201474) and the Special Basic Scientific Research Funds of Central Universities in China.

## Authors’ Affiliations

## References

- McMullen C: Automorphisms of rational maps. In
*Holomorphic Functions and Moduli*. Edited by: Drasin D. Springer, New York; 1988:1-31.Google Scholar - Eremenko AE, Lyubich MY: The dynamics of analytic transformations.
*Leningr. Math. J.*1990, 1: 563-634.MathSciNetMATHGoogle Scholar - Beardon AF: The components of a Julia set.
*Ann. Acad. Sci. Fenn., Ser. A 1 Math.*1991, 16(1):173-180.MathSciNetView ArticleMATHGoogle Scholar - Morosawa S: On the residual buried Julia sets of rational functions.
*Ergod. Theory Dyn. Syst.*1997, 17(1):205-210. 10.1017/S0143385797069848MathSciNetView ArticleMATHGoogle Scholar - Qiao J: The buried points on the Julia sets of rational and entire functions.
*Sci. China Ser. A*1995, 38: 1410-1419.Google Scholar - Qiao J: Topological complexity of Julia sets.
*Sci. China Ser. A*1997, 40: 1158-1165. 10.1007/BF02931834MathSciNetView ArticleMATHGoogle Scholar - Curry CP, Mayer JC: Buried point in Julia set.
*J. Differ. Equ. Appl.*2010, 5: 435-441.MathSciNetView ArticleGoogle Scholar - Lee TD, Yang CN: Statistical theory of equations of state and phase transitions. II. Lattice gas and Ising model.
*Phys. Rev.*1952, 87: 410-419. 10.1103/PhysRev.87.410MathSciNetView ArticleMATHGoogle Scholar - Yang CN, Lee TD: Statistical theory of equations of state and phase transitions. I. Theory of condensation.
*Phys. Rev.*1952, 87: 404-409. 10.1103/PhysRev.87.404MathSciNetView ArticleMATHGoogle Scholar - Derrida B, DeSeze L, Itzykson C: Fractal structure of zeros in hierarchical models.
*J. Stat. Phys.*1983, 33: 559-569. 10.1007/BF01018834MathSciNetView ArticleGoogle Scholar - Bleher P, Lyubich M: Julia sets and complex singularities in hierarchical Ising models.
*Commun. Math. Phys.*1991, 141: 453-474. 10.1007/BF02102810MathSciNetView ArticleMATHGoogle Scholar - Derrida B, Itzykson C, Luck J: Oscillatory critical amplitudes in hierarchical models.
*Commun. Math. Phys.*1984, 94: 115-132. 10.1007/BF01212352MathSciNetView ArticleGoogle Scholar - Joao C, Barata A, Goldbaum P: On the distribution and gap structure of Lee-Yang zeros for the Ising model, periodic and aperiodic couplings.
*J. Stat. Phys.*2001, 103: 857-891. 10.1023/A:1010332500031View ArticleMATHGoogle Scholar - Qiao J: Julia sets and complex singularities in diamond-like hierarchical Potts models.
*Sci. China Ser. A*2005, 48(3):388-412. 10.1360/04ys0180MathSciNetView ArticleMATHGoogle Scholar - Qiao J, Li Y: On connectivity of Julia sets of Yang-Lee zeros.
*Commun. Math. Phys.*2001, 222: 319-326. 10.1007/s002200100507MathSciNetView ArticleMATHGoogle Scholar - Qiao J, Gao J: Jordan domain and Fatou set concerning diamond-like hierarchical models.
*Nonlinearity*2007, 40: 119-131.MathSciNetGoogle Scholar - Qiao J, Yin Y, Gao J: Feigenbaum Julia sets of singularities of free energy.
*Ergod. Theory Dyn. Syst.*2010, 30: 1573-1591. 10.1017/S0143385709000522MathSciNetView ArticleMATHGoogle Scholar - Beardon A:
*Iteration of Rational Functions*. Springer, Berlin; 1991.View ArticleMATHGoogle Scholar - Carleson L, Gamelin TW:
*Complex Dynamics*. Springer, Berlin; 1991.Google Scholar - Milnor J Annals of Math. Studies 160. In
*Dynamics in One Complex Variable*. 3rd edition. Princeton University Press, Princeton; 2006.Google Scholar

## Copyright

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.