NAG-002232 · Accepted whole

Finite forests outside a semistable vertex set

4 Analytic curves and skeleta

Open in the English reader · PDF page 753

Statement · exact TeX
\begin{lemma}[Finite forests outside a semistable vertex set]
\label{v035:f5:finiteforest}
Let $\mathcal C$ be a split nodal algebraic model with vertex set $W$,
skeleton $\Sigma$, and retraction $r:X\to\Sigma$. Let
$V\subset W$ be a semistable vertex set, and let $U$ be a connected
component of $X\setminus V$ meeting $W$. Then
$S=\Sigma\cap U=r(U)$ is a nonempty finite tree with its boundary
endpoints removed. It has one boundary end if $U$ is an open ball
and two if $U$ is an open annulus. Each end has a unique limit in $V$;
the two annular ends are counted separately when their limits coincide.
Every special component indexed by $W\cap U$ is a smooth
$\mathbf P^1_k$.
\end{lemma}
Full selected proof · exact TeX
\begin{proof}
We first record the topology of the complement of $\Sigma$.
Its components are open balls $B$, and
\[
                      \overline B=B\cup\{r(B)\}.
\]
For a smooth residual tube this is
Proposition~\ref{cand:separablesmoothtubehomotopy}.
Inside a nodal tube the annular Gauss retraction leaves residue-disc
components. In an annular coordinate each such disc is
$|T-a|<|a|$, attached at $\zeta_{0,|a|}$. Its closure in a compact
closed subannulus is the disc together with this Gauss point, by the
Gauss seminorm formula and polynomial factorization. This subannulus
is compact and hence closed in $X$, so the same is its closure in $X$.

\emph{The retraction stays in $U$.}
The set $U$ is open in $X$ and closed in $X\setminus V$, so its
boundary lies in $V$. If $y\in U$ retracted to $v\in V$, then $y$
would lie in a smooth residual $W$-tube attached to $v$: residue discs
inside a nodal tube retract to interior edge points, not vertices.
That smooth tube is open and closed in $X\setminus V$ and therefore
would equal $U$. It contains no vertex of $W$, a contradiction.
Consequently $r(U)$ avoids $V$. It is connected and contains every
$w\in W\cap U$, so it lies in $U$. Since $r$ fixes $\Sigma$, this
proves $r(U)=\Sigma\cap U$. It is a connected part of a finite graph
obtained by deleting vertices in $V$.

\emph{There are no circuits.}
Use an actual ball or annulus coordinate on $U$.
Theorem~\ref{fnd:classification} describes every type II or III
point as a Gauss point $\zeta_{a,t}$. At a type II point, removal
partitions the analytic projective line into the inward residue
discs and the outward disc. To check exhaustiveness, if
$|T-a|(x)=t$ and $x\ne\zeta_{a,t}$, there is $b$ with
$|b-a|\leq t$ and $|T-b|(x)<t$. Otherwise every linear factor would
have its Gauss value at $x$, and factorization over $K$ would give
$x=\zeta_{a,t}$. At a type III point, the inward and outward pieces
give the analogous two-sided separation. These are disjoint open
sets and distinguish the two sides of a radial interval.

The Gauss interval between two points of $U$ lies in $U$.
In a ball its joining radius is still smaller than the outer radius.
In an annulus the value of $|T|$ on the interval is between the two
endpoint values, or is their common value. Every interior point of
this interval separates its endpoints. Therefore every path between
the endpoints contains the interval. An embedded compact arc must
equal the interval, since an interval contained in an arc with the
same endpoints is the whole arc. This also agrees with the explicit
interval calculation of Lemma~\ref{cand:gaussdiscmetriccalculus}.
A circuit in $S$ would give two different arcs between two of its
Gauss points. Thus $S$ is a tree.

\emph{Radial ends have limits in $V$.}
For a ball $U=\{|T|<R\}$, the connected outer tails
$E_t=\{t<|T|<R\}$ have nested nonempty compact connected closures
in $X$. Their intersection is nonempty and connected. It misses $U$
and is contained in its finite boundary $V$, hence is a singleton
$\{v_U\}$. Every neighborhood of $v_U$ contains a whole tail:
otherwise the nested compact sets outside that neighborhood would
have a common point. The closed disc $|T|\leq t$ is compact in $U$
and closed in $X$. It follows that every boundary point lies in every
$\overline E_t$, so $\partial_XU=\{v_U\}$.
For an annulus, the same argument applies to each of its two radial
ends. A compact middle subannulus shows that its boundary consists
of the two resulting limits $v_-,v_+$; they may be equal.

\emph{Each radial end occurs exactly once in the finite tree.}
Choose a compact disc or middle annulus containing the finite set
$W\cap U$. Every graph end of $S$ leaves all compact subsets of $U$,
because its endpoint in the compact graph $\Sigma$ lies in $V$.
Its final portion therefore approaches one coordinate end. The
cut-point separation just proved forces this portion to contain a
sufficiently short radial Gauss tail. Two different graph ends
cannot approach the same coordinate end, since they would share
a nonempty Gauss interval.

Suppose a coordinate end were unrepresented. Some sufficiently far
tail $E$ would then miss $S$. Indeed, if all tails met $S$, one of
its finitely many edges would have points escaping toward that end;
compactness of the closed edge would produce its endpoint in $V$
and a graph end approaching the chosen coordinate end.
The connected set $E$ lies in one ball component $B$ of
$X\setminus\Sigma$. Because $B$ meets $U$ and avoids $V$, we have
$B\subset U$. Also $r(B)\in r(U)\subset U$, so the compact closure
$B\cup\{r(B)\}$ misses $V$. All shorter tails lie in $B$, contrary
to the nonempty limiting set in $V$ proved above. Every coordinate
end is therefore represented. This gives exactly one or two ends.

\emph{The unwanted components are smooth rational curves.}
For $w\in W\cap U$, translation and scaling identify its type II
Gauss field with the unit Gauss field. Its residue field is $k(t)$
by Proposition~\ref{new:gauss_residue_formula}.
Proposition~\ref{cand:componentvaluedresiduefield} identifies this
with the function field of the special component $D_w$.
The normalization $D_w^\nu\to D_w$ is finite by
\cite[Tag 0BXR]{StacksAlg}. It is therefore proper, and $D_w^\nu$
is proper over $k$ by \cite[Tags 01WM, 01W3]{StacksAlg}.
The normal proper curve equivalence \cite[Tag 0BY1]{StacksAlg}
now identifies $D_w^\nu$ with $\mathbf P^1_k$.
A self-node of $D_w$ would give a loop based at $w$
in $\Sigma$. Its interior is connected, avoids $V$, and meets the
open set $U$ near $w$, so the entire loop lies in $U$. This contradicts
acyclicity. Thus $D_w$ has no self-node. At a node shared with another
component, its own branch is smooth, by the pointed branch calculation
in Lemma~\ref{cand:pointednodalbranches}. Hence $D_w$ itself is smooth
and equals its normalization, as asserted.
\end{proof}
Typeset reader excerpt
Typeset mathematical passage from PDF page 753
PDF page 753 · excerpt continues in the reader

Typed proof inputs

Named external registrations
Reader cross-references

Proof map

Verified map inputs · NAG-002232

AND · joint registered proof inputs

Unlinked source pointers

These pointers have no operative linked edge in this map.

  • APP-STACKS-01W3@NAG-002232:existing-13-1
  • APP-STACKS-01WM@NAG-002232:existing-12-1
  • APP-STACKS-0BXR@NAG-002232:existing-11-1
  • APP-STACKS-0BY1@NAG-002232:existing-14-1