Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

Let $X$ be a topological space and $U,V \subset X$ open subsets. Let $x:U \to \mathbb{C}$ and $y:V \to \mathbb{C}$ be homeomorphisms. If $x\circ y ^{-1}$ is holomorphic how do I show that $y\circ x^{-1}$ is holomorphic?

share|cite|improve this question
@potato: there is nothing to look up because the result is false over $\mathbb R$. Counterexample: $x\mapsto x^3$ – Georges Elencwajg Jun 24 '12 at 5:02
@GeorgesElencwajg You're right; my mistake. My advice to prove the more general statement that "if $f$ is bijective and holomorphic then so is $f^{-1}$" stands though. – Potato Jun 24 '12 at 5:37
It has been a while, but I believe you just invert the power series at an arbitrary point to show the inverse is holomorphic there? – Potato Jun 24 '12 at 5:52
Use the Open Mapping Theorem for non-constant analytic functions. – Robert Israel Jun 24 '12 at 7:09
Sorry, I wasn't thinking. The Open Mapping Theorem ensures that a nonconstant one-to-one analytic function is a homeomorphism onto its image. The fact that the inverse is analytic is from the formula for the derivative of an inverse function: $(f^{-1})'(z) = 1/f'(f^{-1}(z))$ (and that $f$ is not one-to-one in a neighbourhood of a point where $f'=0$). – Robert Israel Jun 24 '12 at 8:08

Holomorphic functions in one variable are so rigid that they admit of a complete local description, an exceptionally pleasant situation .

Namely, if $f:U\to V$ is a non-constant holomorphic map between connected open subsets of $\mathbb C$ (or between Riemann surfaces), then for $a\in U$ we can write $f(z)=\phi(z)^n$ on an open neigbourhood $W$ of $a$, with $\phi: W\stackrel {\cong}{\to} W' \subset \mathbb C$ a holomorphic isomorphism .

[The proof is not so difficult : suppose $a=f(a)=0$ and write $f(z)=cz^n(1+zg(z))$, with $c\neq 0\in \mathbb C$ and $g$ holomorphic.
Since $c(1+zg(z))$ is non zero near $z=0$, we can extract a holomorphic $n$-th root of it, i.e. find on some neighbourhood $W$ of $a$ a function $h\in \mathcal O(W)$ such that $h(z)^n=c(1+zg(z))$.
The required isomorphism is then $\phi(z)= z\cdot h(z)$ ]

Using this local description it is then clear that injectivity of $f$ forces $n=1$ (because $f(z)=\phi(z)^n$ and $\phi$ is bijective) and thus the bijective function $f$ is an isomorphism because locally it is one: $f \mid W=\phi$.
Notice that the structure theorem also implies that non constant holomorphic functions are open, a non trivial result evoked by Robert in his comment.

share|cite|improve this answer
Dear @GeorgesElencwajg I didn't understand how your answer prove the statement of my question – Jr. Jun 25 '12 at 3:12
Dear Jr.,you know that $f=x\circ y^{-1}$ is holomorphic and you know that it is a homeomorphism, so in particular that it is bijective.The result I proved in the answer then says that $f$ is a holomorphic isomorphism. Hence $f^{-1}=y\circ x^{-1}$ is also holomorphic. Isn't that what you wanted ? – Georges Elencwajg Jun 25 '12 at 5:30
wouldn't it be "$z \in U$"? – Jr. Jun 25 '12 at 16:55
Yes, there was a typo: I corrected it by replacing $z$ by $a$ in the fifth line.Thanks. – Georges Elencwajg Jun 25 '12 at 17:26
How does one know that $f$ can be written in the form $f(z)=cz^n(1+zg(z))$? – Jr. Jun 26 '12 at 17:50

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.