When I try to find the covering space of a punctured unit circle, I met some obstacles. The only one remain is given as follows:

Suppose $H$ is the left half plane, $D^*$ is the punctured unit disk with $D^*=\{z\in\mathbb{C}:0<|z|<1\}$. $X$ is a Riemann surface, where $f:X\to D^*$ is an unbranched holomorphic covering mapping. Since $\exp:H\to D^*$ is the universal covering, it follows that there exists a holomorphic mapping $\psi$ such that the following diagram commutes. $$\require{AMScd} \begin{CD} X @<{\psi}<< H\\ @V{f}VV @VV{\exp}V \\ D^* @<{id}<< D^* \end{CD}$$ My question is, is $\psi$ a surjective mapping?

This question arouse when I try to classify the covering space with the deck transformation subgroup $G=Deck(H/X)$. When $G$ contains only the identity, $\psi:H\to X$ is injective, then it's biholomorphic and onto. But when $G$ contains other stuff, I can't see why it's onto. Or can I prove that there is a biholomorphic mapping without using the condition that $\psi$ is onto?

Can I show that $\psi$ is actually a covering map? Since covering map is sure to be surjective. Forster's defines the universal covering like this:

Suppose $X$ and $Y$ are connected topological spaces and $p:Y\to X$ is covering map, $p:Y\to X$ is the universal covering of $X$ if it satisfies the following universal property. For every covering map $q:Z\to X$, with $Z$ connected, and every choice of points $y_0\in Y,z_0\in Z$ with $p(y_0)=q(z_0)$, there exists exactly one continuous fiber-preserving mapping $\psi:Y\to Z$ s.t. $\psi(y_0)=z_0$.

The definition of universal covering in Wikipedia is different, and the only difference is that $\psi$ should be a covering map. So I wonder are these two definitions equivalent? Since continuous fiber-preserving mapping does not implies covering, I can't see why they are equivalent. Thanks for attention!

  • $\begingroup$ I don't understand what lifting property you use to define $\psi$. $\endgroup$ – Seub Dec 14 '13 at 15:12
  • $\begingroup$ To have the existence of $\psi$, you probably have to assume that $f$ is an unbranched unlimited covering. Then $\psi(H)$ as well as its complement are open. $\endgroup$ – Daniel Fischer Dec 14 '13 at 15:16
  • $\begingroup$ @Seub Sorry, I forgot to point out that "$f$ is a covering map". Now an edit has been made. $\endgroup$ – Golbez Dec 14 '13 at 17:30
  • $\begingroup$ @DanielFischer Yes, but how to show that "both $\psi(H)$ and its complement are open"? I stuck here for some hours. It seems really trivial, but I can't see what I have missed. Could you give more details? Thanks! $\endgroup$ – Golbez Dec 14 '13 at 17:32
  • $\begingroup$ Stop me if I'm saying something stupid, but it seems to me that the theorem (sort of characterization of the universal cover) giving the existence of $\psi$ says that it is a covering. So it's surjective by definition (?) $\endgroup$ – Seub Dec 14 '13 at 17:48

With $f$ an unbranched unlimited covering, we have $\psi(H) = X$, since

  1. $\psi$ is a non-constant holomorphic mapping (and $H$ is connected), so $\psi(H)$ is open, and

  2. $X\setminus \psi(H)$ is also open (see below), hence, since $X$ is connected too, we must have $\psi(H) = X$.

To see that $X \setminus \psi(H)$ is open, consider an $x \in X\setminus\psi(H)$.

$f(x)$ has an evenly covered (by $f$) connected neighbourhood $U$. We may assume that $U$ is also evenly covered by $\exp$ (otherwise choose a smaller disk around $f(x)$). Let $V$ be the connected component of $f^{-1}(U)$ that contains $x$. If there was a $z \in H$ with $\psi(z) \in V$, then let $W$ be the component of $\exp^{-1}(U)$ containing $z$. Then $(f\lvert_V)^{-1} \circ (\exp\lvert_W)$ and $\psi\lvert_W$ coincide in a neighbourhood of $z$, hence in all of $W$, and thus it follows that $x\in (f\lvert_V)^{-1}\circ (\exp\lvert_W)(W) = \psi(W) \subset \psi(H)$ contradicting the assumption $x \notin \psi(H)$. Hence $V \cap \psi(H) = \varnothing$, and $X\setminus \psi(H)$ is recognised as open.

  • $\begingroup$ Really helpful! Thanks! This also answer the question that these two definitions are equivalent. $\endgroup$ – Golbez Dec 14 '13 at 17:57

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.