1: \begin{definition} \label{varsigma-def}

2: Define the involution $\varsigma$ on the $X_{n-1}$ highest weight vectors of $C^{r,s}$, as

3: indexed by $\pm$ diagrams, as follows.

4: Let $P$ be a $\pm$ diagram with $\os(P)=\Lambda$ and $\is(P)=\lambda$. For each

5: $1 \leq i \leq r-1$:

6: \begin{enumerate}

7: \item If  $i\equiv r-1 \pmod{2}$ then above each column of $\la$ of height $i$, there must

8: be a $+$ or a $-$. Interchange the number of such $+$ and $-$.

9: \item If $i\equiv r \pmod{2}$ then above each column of $\la$ of height $i$,

10: either there are no signs or a $\mp$ pair. Interchange the number of columns of each type.

11: \end{enumerate}

12: By Theorem~\ref{branching}, this can be extended in a unique way to an involution

13: $\varsigma$ on $C^{r,s}$.

14: \end{definition}

15: