1: \begin{definition}\label{sigmasharp}\text{ }

2: 

3: \medskip

4: \begin{enumerate}

5:     \item $\Sig^\#:=\{\ul{R}\in\Sig:\exists n_k,m_k\uparrow\infty\text{ s.t. }R_{n_k}=R_{n_0},R_{-m_k}=R_{-m_0},\forall k\geq0\}$.

6: %    \begin{align*}

7: %        &\Sigma^\#:=\{\ul{u}\in\Sigma:\exists n_k,m_k\uparrow\infty\text{ s.t. }u_{n_k}=u_{n_0},u_{-m_k}=u_{-m_0}\forall k\geq0\},\\

8: %        &

9: %        \end{align*}

10:     \item Every two partition members $R,S\in \mathcal{R}$ %(where $\mathcal{R}_\epsilon$ is the Markov partition  given by the Bowen-Sinai refinement of the cover $\mathcal{Z}_\epsilon$, see \cite[proof of Theorem 1.1]{SBO}), 

11: are said to be {\em affiliated} if $\exists u,v\in\mathcal{V}$ s.t. $R\subseteq Z(u), S\subseteq Z(v)$ and $Z(u)\cap Z(v)\neq\varnothing$ (this definition is due to O. Sarig, \cite[\textsection~12.3]{Sarig}).

12: %    \item Every sequence (finite or infinite) of symbols in $\mathcal{V}_\epsilon$ is called {\em admissible} if it can be extended to a chain in $\Sigma_\epsilon$. Similarly for sequences of symbols in $\mathcal{R}$, and $\Sig$. Chains in $\Sigma,\Sig$ are also sometimes referred to as admissible.

13: \end{enumerate}

14: \end{definition}

15: