2
$\begingroup$

Let $A$ be a simply connected domain, $\gamma : I \to A$ a closed contour. Let $P$ be the set of all points $z \in \overline{A} - \gamma(I)$ such that $\text{Ind}_\gamma(z) \ne 0$. I want to show that that $\overline{P} = P \cup \gamma(I)$.

By continuity of $\text{Ind}_\gamma(z)$, we have $P \cup \gamma(I) = \overline{A} \cap \text{Ind}_\gamma^{-1}(\mathbb{Z} - \{0\}) = \overline{P \cup \gamma(I)} = \overline{P}\cup \gamma(I)$, so it is sufficient to show that $\gamma(I) \subset \overline{P}$.

Information that may be useful: $\text{Ind}_\gamma(z)$ is constant in each connected component of $\mathbb{C} - \gamma(I)$; $\text{Ind}_\gamma(z) = 0$ if $z \in \mathbb{C} - A$; $\overline{P}$ is compact.

$\endgroup$
  • $\begingroup$ Maybe Jordan curve theorem is needed here? $\endgroup$ – Crostul Nov 10 '15 at 12:54
  • $\begingroup$ @Crostul $\gamma(I)$ is not necessarily a simple closed curve. $\endgroup$ – Randy Randerson Nov 10 '15 at 17:56
0
+100
$\begingroup$

This isn't true in general. For instance, suppose $\gamma$ goes around a circle once counterclockwise, then goes back around the same circle clockwise. Then $P$ is empty.

If you assume $\gamma$ only intersects itself finitely many times, then it is true. For in that case the non-self-intersection points of $\gamma$ are dense in $\gamma(I)$, so it suffices to show such points are all in $\bar{P}$. But at any such point $p$, $\gamma$ is locally injective; it then follows from the Jordan–Schoenflies theorem that there is a homeomorphism of a small ball around $p$ with a disk which turns $\gamma(I)$ into a diameter of the disk. The index of $\gamma$ on the two half-disks separated by the diameter will then differ by $1$, so at least one of them must have nonzero index.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.