Add bridge-derived census (n=6..10) to the disjunction section

Cross-tabulate bridge-derived vs intertwining-tree coverage: the
bridge-derived share falls from 100% (n=6) to 62.7% (n=10), the
disjunction never relies on it alone, and the "neither" column is
identically zero throughout.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

papers/even_level_graph_generators/paper.tex

@@ -492,6 +492,45 @@ $n = 21$ and there are exactly $6$ of them. Below $n = 21$ every
 maximal planar graph is an intertwining tree, which is why the
 disjunction holds trivially in that range.
 
+The intertwining-tree disjunct therefore carries the conjecture by itself
+for all small $n$, but this leaves open how much of the load the
+bridge-derived disjunct is independently able to bear. To measure that we
+classified every triangulation at $6 \leq n \leq 10$ as bridge-derived or
+not, deciding bridge-derivedness exhaustively: a triangulation is
+bridge-derived iff some valid parity partition admits an Even Level Graph
+in its backward bridge-switch orbit (a search feasible only at these
+sizes). Table~\ref{tab:bridge-census} records the result. Three features
+stand out. First, the bridge-derived disjunct is substantive but far from
+universal on its own: its share of all triangulations falls steadily, from
+all of them at $n = 6$ to under two-thirds by $n = 10$. Second, the
+disjunction never relies on it in this range -- the \emph{intertwining
+only} column counts triangulations covered by the tree disjunct alone, and
+it grows, while no triangulation here is bridge-derived without also being
+an intertwining tree. Third, and consistent with the conjecture, the
+\emph{neither} column is identically zero throughout.
+
+\begin{table}[ht]
+\centering
+\begin{tabular}{cccccc}
+$n$ & triangulations & bridge-derived & \% & intertwining only & neither \\\hline
+$6$  & $2$    & $2$    & $100.0$ & $0$   & $0$ \\
+$7$  & $5$    & $4$    & $80.0$  & $1$   & $0$ \\
+$8$  & $14$   & $12$   & $85.7$  & $2$   & $0$ \\
+$9$  & $50$   & $36$   & $72.0$  & $14$  & $0$ \\
+$10$ & $233$  & $146$  & $62.7$  & $87$  & $0$ \\
+\end{tabular}
+\caption{Bridge-derived census for $6 \leq n \leq 10$. \emph{bridge-derived}
+counts plane-triangulation iso classes that are bridge-derived level graphs
+of some Even Level Graph, decided by exhaustive backward bridge-switch
+search over all valid parity partitions; \% is its fraction of all
+triangulations. \emph{intertwining only} counts those that are intertwining
+trees but not bridge-derived; \emph{neither} counts those covered by no
+disjunct. Every triangulation in this range is an intertwining tree, and
+every bridge-derived one is too, so bridge-derived $\subseteq$ intertwining
+tree here.}
+\label{tab:bridge-census}
+\end{table}
+
 \subsection*{The boundary case $n = 21$}
 
 The first triangulations that are \emph{not} intertwining trees are the