2
$\begingroup$

I was reading a book [ Teichmuller Theory and quadratic differential and Farb-Margalits' A Primer on MCG ] where they define the natural co-ordinate of holomorphic quadratic differential on a compact Riemann surface without boundary. They define the natural co-ordinate by taking the line integral of $ \sqrt \phi $ , where the quadratic differential is $ q = \phi(z) dz^2 $ locally on X. But if $\phi$ has a zero of odd order, how can I define the square root ? In the books they always avoided that case. Any help ?

$\endgroup$
0

1 Answer 1

3
$\begingroup$

The following result can be found in the book of Strebel "Quadratic differentials", p. 29.

Theorem

In the neighborhood of any zero $P_0$ we can introduce a local parameter $\xi$, such that $P_0$ corresponds to $\xi=0$, in terms of which the quadratic differential $q$ has the representation

$\phi(\xi) d \xi^2=\big(\frac{n+2}{2} \big)^2 \xi^n d \xi^2$.

The integral $\Phi(\xi)$ has then the simple form

$\Phi(\xi)=\xi^{\frac{n+2}{2}} + \textrm{const.}$

The proof is straightforward (manipulation of power series), and can be carried out for both $n$ even and $n$ odd.

$\endgroup$
1
  • $\begingroup$ Late question: The problem I have with his proof is, that he integrates over some root z^(n/2) to get an antiderivative (to get his series over c_n times z^(n+2/2). I fail to see why we can do this, since we would have to make a cut somewhere to make it analytic? $\endgroup$
    – ctst
    Commented Jan 8 at 22:10

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .