1

This is exercise 2.2 in Hartshorne's AG: enter image description here

What I have done:

a):Apply Hurwitz theorem and observe that ramification index is no greater than degree. b):There is $ k[x]\rightarrow k[x,z]/(z^2-\prod (x-\alpha_i ))$.So explicit calculation can be done. c):Follow the hints.d):Nothing to prove...

Here is my question:

How the one-to-one correspondence in e) actually comes out?One direction(namely given some points of $P^1$,we construct a curve)is b).The other(given a curve $f:X\rightarrow P^1$,we consider these points of $P^1$ over which $f$ has ramificaition)is a).But it's not obvious for me to see how these two procedures are actually inverse to each other...namely from $X$ by procedure in a) we get some points $p_i$,then from $p_i$ by the procedure in b) we get another curve $Y$,What's the relationship of $X$ and $Y$?

I have look up several solutions of Hartshorne about e),they all say "this follows directly from a,b,c,d" but I still don't see why.

Any reference or help is appreciated.

edit: Here is my understanding of "$f$ is uniquely determined by its branch locus up to automorphisms of $\mathbb{P}^1$".suppose $f:X\rightarrow \mathbb{P}^1$,take an affine cover $U=\text{Spec} k[x]$ of $\mathbb{P}^1$,assume ramification points of $f$ all belongs to $f^{-1}(U)=\text{Spec} A$.$A$ can be supposed to take this form $k[x,z]/(u(x,z))$,since $f$ is of degree 2,we have $u(x,z)=z^2+g(x)z+h(x)$.So the rest is to prove under some automorphism $x\rightarrow \frac{ax+b}{cx+d}$,$u$ will take the form $z^2+\prod_i (x-\alpha_i )$.Still don't know how.:(

XiaYu
  • 587
  • 2
    The fact that these are inverse of each other follows from the last parts of (1) and (2) - where it says that f is uniquely determined by its branch locus up to automorphisms of $\mathbb{P}^1$, and that the cover you construct from the 6 points is the one given the canonical linear system. – loch Oct 19 '18 at 19:16
  • @loch thanks for your comment.I can't figure out why "f is uniquely determined by its branch locus up to automorphisms of $\mathbb{P}^1$".Could you give me some reference? Anyway I want to skip this question for now and try it later after I learn more from other book...... – XiaYu Oct 20 '18 at 14:22
  • @loch by the way I update my understanding of "f is uniquely determined by its branch locus" in the post. – XiaYu Oct 20 '18 at 14:24
  • 1
    $f$ is the map determined by the canonical linear system $|K_X|$. It follows from Riemann-Roch that this is the only linear system with dimension $g-1$ and degree $2g-2$, hence $f$ is unique up to automorphism. – DKS Oct 23 '18 at 11:14
  • @DKS thank you!! – XiaYu Oct 23 '18 at 11:22

0 Answers0