0
$\begingroup$

I am learning the Riemann Surfaces and I have a question when I read the appendix A.1.5. of textbook A course Complex analysis and Riemann surfaces by W.Schlag.

One way to define the degree of the non-constant and analytic map $f$ between Riemann surfaces is that $$\deg(f) := \sum_{p\in f^{-1}(q)}(b_{p}(f)+1), $$ where $p$ is a branch point and $(b_{p}(f)$ barnch numbers.

And the other notion of the degree from topology,

Let $f: M \to N$ be a smooth map and $f^{*}: H^{n}(N) \to H^{n}(M)$ the induced map defined via the pullback. There exists a real number denoted by $\deg(f)$ s.t. $$ \int_{M} f^{*}(\omega)=\deg(f) \int_{N} \omega,\ \forall \omega \in H^{n}(N).$$

For the Riemann surfaces $n=2$, hence the definition of the degree is that $$\int_{M} f^{*}(d\sigma)=\deg(f) \int_{N} d\sigma,\ \forall d\sigma \in H^{2}(N).$$

$\textbf{My question}$ is that why it is easy to verify that for any $\textbf{regular}$ value $q\in N$, which means that $Df(p): T_{p}M \to T_{p} N$ is invertible for every $p$ with $f(p)=q$, $$\deg(f)= \sum_{p\in f^{-1}(q)} Ind(f;p),$$ where $Ind(f;p)=\pm 1$ depending on whether $Df(p)$ preserves or reverses the orientation.

$\endgroup$
1
$\begingroup$

First, observe that the first definition is valid at any given point $q$ and implies that the degree depends only on a neighbourhood of the fiber (since you look at the ramification indices to determine the degree).

For each point $p_i$ in the fiber let $U_i$ be a neighbourhood such that $f_i:=f_{|U_i}$ is a diffeomorphisme onto its image (local inversion theorem).

You can assume that all the $f_i$ have the same image V.

Let $\sigma$ be a differential 2-form. $deg(f) \int_V \sigma =\int_{\cup \, U_i} f_i ^{*}\sigma =\Sigma (\int_{U_i} f_i ^{*}\sigma )=\Sigma (\pm\int_V \sigma )$ where $\pm$ depends on if $f_i$ is orientation-preserving or orientation-reversing.

Let us clarify what it means to be orientation-preserving in this particular case.

Since Riemann surfaces are orientable (see orientability of riemann surface), you have an oriented atlas over $M$ and $N$. You can always assume that each set is contained in such an atlas. If $dx^1\wedge dx^2$ in $V$ is sent to $\lambda \times dy^1\wedge d y^2$ in $U_i$ for some $\lambda>0$, $f_i$ is orientation-preserving. If not, it sends $dx^1\wedge dx^2$ to $-\lambda\times dy^1\wedge d y^2$ and is orientation reversing. If you are unfamiliar with differential geometry, think of it as being in $\mathbb{R}^2$ and applying the formula $\int_{U_i}f_i^*(g(x)dx^1\wedge dx^2)=\int_{U_i}\pm|det(J_{f_i}(y))|(g(f_i(y))dy^1\wedge\ dy^2)=$ $\int_{U_i}\pm|det(J_{f_i}(y))|(g(f_i(y))dy_1 dy_2=\pm\int_{V}g(x)dx_1 dx_2=\pm\int_{V}g(x)dx^1\wedge dx^2)$

$\endgroup$

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.