Linear fractional transformation of quadratic differentials on the Riemann sphere

165 Views Asked by At

Suppose I have the following quadratic differential on the Riemann sphere with four punctures:

\begin{equation} q = -\frac{9 t \left(216+t^3\right)}{\left(-27+t^3\right)^2} dt^2 \end{equation}

This has four zeroes, and four second-order poles (one for each of the punctures).

Now, since $q$ is on the sphere, we should be allowed to do an $SL\left(2, \mathbb{C}\right)$ reparameterisation, of the form $t \rightarrow \frac{az+b}{cz+d}$. So, to test, choose $a=c=b=1$, and $d=2$. Then, using the quadratic differential transformation rule that if $q = f\left(z\right)dz^2 = g\left(w\right)dw^2$, then

\begin{equation} f\left(z\right) = g\left(w\left(z\right)\right)\left(\frac{d w\left(z\right)}{dz}\right)^2 , \end{equation}

I find

\begin{equation} q = -\frac{9 \left(1729+4324 z+3894 z^2+1516 z^3+217 z^4\right)}{(2+z)^2 \left(215+321 z+159 z^2+26 z^3\right)^2} dz^2 . \end{equation}

This now has five second-order poles. But it should still be living on a sphere with four punctures, so it looks like it has picked up another pole. This confuses me; I had really hoped the number of poles would be invariant under such reparameterisations.

So, my question is:

  • Is such an $SL\left(2,\mathbb{C}\right)$ transformation on a quadratic differential on a Riemann sphere with punctures actually legitimate? If not, why not?

  • If the transformation is legitimate, what is the explanation for the extra pole in the quadratic differential after the transformation?

Thanks in advance for any help!

1

There are 1 best solutions below

0
On BEST ANSWER

The answer to the question, as Alan Muniz points out, is that such $SL\left( 2, \mathbb{C} \right)$ transformations of the quadratic differentials are legitimate, and they do preserve the number of poles of the quadratic differential. The problem in the above is that I incorrectly counted five poles in $q\left(z\right)$, when in fact there are only four - since $q\left(z\right)$ is holomorphic at $\infty$.