MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

A finite branched holomorphic covering is a holomorphic map $f : V \to W$ between holomorphic varieties $V$ and $W$ such that

  • $f$ is a finite branched covering (in the topological sense)
  • There is a regular part $V_0 \subset V$ on which $f$ is a locally biholomorphic mapping

On the regular part, the behavior of $f$ is relatively simple since it is locally biholomorphic, and locally $V_0$ looks just like $W_0 = f(V_0)$. My question is: do we (can we) know anything about the nonregular part (maybe that's not the right term)? I.e., can we know the structure of the map $f:(V \setminus V_0) \to W$?

In particular I am interested in the case where $W$ is simply a open disc of $\mathbb{C}$ centered at the origin, and $W_0$ is the whole disk without the origin. How should I study the structure of $f^{-1}(0)$?

share|cite|improve this question

This is only a partial answer. Regarding the last part of your question:

Every (connected) finite branched cover $f: V\rightarrow \mathbb{D}$ of the disc is isomorphic to one of the form $p_n: \mathbb{D}\rightarrow \mathbb{D}$, where $p_n(z) = z^n$ and $n\in \mathbb{Z}^+$. In particular, $f^{-1}(0)$ consists of a single point since $z^n$ has a unique zero.

To get a better idea of why only the map $z\rightarrow z^n$ appears, note that every such branched covering arises from an unbranched covering $g:X\rightarrow \mathbb{D}^*$ which one can also show is of the form $p_k: \mathbb{D}^* \rightarrow \mathbb{D}^* $ up to isomorphism. The original branched covering $V\rightarrow \mathbb{D}$ is obtained by completing, i.e. "adding an extra point" to $\mathbb{D}^*$ and extending $g$ so that it takes the value $0$ on this new point. So $f^{-1}(0)$ consists of a single point (of "multiplicity" $k$) by construction. Now, if you don't require $V$ to be connected, you'll end up with possibly several disjoint copies of this situation, namely $N$ discs $\mathbb{D}_j$ equipped with maps $z\rightarrow z^{n_j}$. Then $f^{-1}(0)$ consists of $N$ points $v_j$ with multiplicities $n_j$. This is the situation you'll find when analyzing the preimage of a small embedded disc under a non-constant map of Riemann surfaces.

In higher dimensions, branched coverings are more difficult to understand. One useful fact is that such maps are proper (by definition). So if $V,W$ are smooth, then the Proper Mapping Theorem applies. In particular, $f(V\backslash V_0)$ is an analytic hypersurface in $W$ (since the ramification divisor $V\backslash V_0$ is an analytic hypersurface in $V$ - see Griffiths & Harris).

Edit: One nice reference for the basics of branched covers is the book "Holomorphic Functions to Complex Manifolds" by Grauert and Fritzsche.

share|cite|improve this answer
Thank you so much for the answer. For the first part, what's the name of the theorem? I have heard different names like "local uniformization theorem" and "local canonical form", and I had trouble finding the original statement and proof. Why $f^{-1}(0)$ is just a single point: why can't a few sheets coalesce into one point in $f^{-1}(0)$ while the remaining sheets coalesce into a different point? – ssquidd Jul 5 '12 at 15:49
Hi ssquidd. I was going to reply with a comment, but it ran too long. I've edited the my answer to reflect your question. – Kevin Jul 5 '12 at 17:48
Miranda calls the local expression $z\rightarrow z^n$ of a holomorphic map between Riemann surfaces the "local normal form". That such maps have a this local form follows easily from power series manipulations. Check out his book "Algebraic Curves and Riemann Surfaces" for more. – Kevin Jul 6 '12 at 5:16
So, did you want to know more about the higher dimensional situation or is my answer satisfactory? – Kevin Jul 20 '12 at 18:31

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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