1: \begin{definition}\label{def:Dgamma}

2: Fix a regularity structure $\TT$ and a model $(\Pi,\Gamma)$.

3: Then, for any $\gamma \in \R$,  the space $\CD^\gamma$ consists of all $T_\gamma^-$-valued functions 

4: $f$ such that, for every compact set $\K \subset \R^d$, one has

5: \begin{equ}[e:boundDgamma]

6: \$f\$_{\gamma;\K} = \sup_{x\in \K} \sup_{\beta < \gamma} {\|f(x)\|_\beta}  + \sup_{(x,y) \in \K \atop \|x-y\|_\s \le 1} \sup_{\beta < \gamma} {\|f(x) - \Gamma_{xy} f(y)\|_\beta \over \|x-y\|_\s^{\gamma - \beta} } < \infty\;.

7: \end{equ}

8: Here, the supremum runs only over elements $\beta \in A$. 

9: We call elements of $\CD^\gamma$ \textit{modelled distributions} for reasons that will

10: become clear in Theorem~\ref{theo:reconstruction} below.

11: \end{definition}

12: