@def maxtoclevel=3

Algebraic Curves Lecture 9:

The 9th lecture of Algebraic Curves by Karl Christ
  1. Coverings

Coverings

Let f:YXf:Y\to X be a map of curves, pYp\in Y, q=f(p)q = f(p).

The ramification index at pp is vp(t)v_p(t).

If vp(t)=1v_p(t) = 1 we say that ff is unramified at pp. Otherwise pp is a ramification point of ff, in which case we say that qq is a branch point.

The ramification divisor of ff is

R=pY(vp(t)1)p.\begin{aligned} R = \sum_{p\in Y} (v_p(t) - 1)\cdot p. \end{aligned}

Implicit in this definition is the fact that ff has finitely many ramification points. Our first theorem is the Riemann-Hurwitz formula.

Theorem: In this situation in characteristic 00,
KY=fKX+R\begin{aligned} K_Y = f^*K_X + R \end{aligned}
and in particular
2g(Y)2=d(2g(X)2)+pY(vp(t)1)\begin{aligned} 2g(Y) - 2 = d(2g(X) - 2) + \sum_{p\in Y}(v_p(t) - 1) \end{aligned}
where d=deg(t)d = \deg(t).

Remark: We say that ff is separable if the induced field extension K[Y]/K[X]K[Y]/K[X] is separable. In positive characteristic

  1. If ff is not separable then it is ramified along all of YY

  2. Even if it is separable the formula might fail if the ramification is not tame (meaning the ramification index is divisible by the characteristic of the base field).


Proof.   (Sketch of an analytic proof.)   Let ω\omega be a meromorphic 1-form on XX the target curve. Pulling it back by ff gives us a 1-form on YY. Away from branch points, ff is a local isomorphism (here we have a drawing of a stack of analytic disks in YY over an analytic disk around the point qq in XX) and thus any zero or pole of ω\omega is given dd zeros or dd poles on YY.

At a ramification point pp, ff is given analytic locally be zzvp(t)z\mapsto z^{v_p(t)}. On the other hand, ω\omega is locally given by dt=dzvp(t)=(vp(t))zvp(t)1dzdt = dz^{v_p(t)} = (v_p(t))z^{v_p(t)-1}dz on YY.

\square
 

We can see where this fails when the ramification is not tame; if the characteristic of the base field divides vp(t)v_p(t) then ω\omega vanishes on YY.

Cor: There is no nontrivial étale cover YP1Y\to \mathbb P^1.

Proof.   Recall that étale = flat + unramified. Any finite surjective map of nonsingular varieties is flat, but a cover f:YP1f:Y\to \mathbb P^1 will always fail to be unramified. Indeed, suppose ff is étale. Riemann-Hurwitz says that
2g(Y)2=2d+deg(R)=2d\begin{aligned} 2g(Y) - 2 = -2\cdot d + \deg(R) = -2d \end{aligned}
which implies g(Y)=0,d=1g(Y) = 0, d = 1. The important part is that deg(f)=d=1\deg(f) = d=1, hence ff is an isomorphism on function fields. hence ff is an isomorphism. The only étale map is therefore trivial.
\square
 

Cor: If f:YXf:Y\to X is a cover then g(Y)g(X)g(Y) \geq g(X).

Proof.  

If g(X)=0g(X) = 0 then there is nothing to show. Otherwise, we get by Riemann-Hurwitz that

2g(Y)2=d(2g(X)2)+deg(R)    g(Y)=d(g(X)1)+deg(R)2+1    g(Y)=g(X)+(d1)(g(X)1)+deg(R)2.\begin{aligned} &2g(Y) - 2 = d\cdot (2g(X) - 2) + \deg(R) \\ &\implies g(Y) = d(g(X) - 1) + \frac{\deg(R)}{2} + 1 \\ &\implies g(Y) = g(X) + (d-1)(g(X) - 1) + \frac{\deg(R)}{2}. \end{aligned}
If g(X)>0g(X) > 0 then (d1)(g(X)1)0(d-1)(g(X) - 1)\geq 0 and deg(R)20\frac{\deg(R)}{2}\geq 0 so the claim follows.
\square
 

Cor: deg(R)\deg(R) is even.
Cor: If g2g\geq 2 and g(X)=g(Y)=gg(X) = g(Y) = g then any cover f:YXf:Y\to X is an isomorphism.
Proof.   2g2=d(2g2)+deg(R)2g - 2 = d\cdot (2g - 2) + \deg(R) which implies d=1d = 1 and hence ff is an isomorphism.
\square
 
Example: Take f:P12:1P1f:\mathbb P^1\xrightarrow{2:1} \mathbb P^1. Then
2=2(2)+deg(R)    deg(R)=2\begin{aligned} -2 = 2\cdot (-2) + \deg(R) \implies \deg(R) = 2 \end{aligned}
and so ff is ramified at two points with ramification index 22: [x:y][x2:y2][x:y] \mapsto [x^2:y^2].
Cor: (Hurwitz bound).   Suppose XX has genus 2\geq 2 and finitely many automorphisms. Then Aut(X)84(g1)|\operatorname{Aut}(X)| \leq 84(g - 1).

Remark:

  1. This bound is sharp in the sense that it is achieved for infinitely many gg, but not all of them. For example, up to g=6g = 6 it is achieved only in g=3g = 3 by the Klein quartic:

{x3y+y3z+z3x=0}P2\begin{aligned} \{x^3y + y^3z + z^3 x = 0\} \subseteq \mathbb P^2 \end{aligned}
  1. The upshot is that if you fix gg, there is a bound on the automorphisms of ALL curves of that genus. A priori, it might be the case that for fixed genus gg you could find curves of that genus whose automorphism group has arbitrarily large size.

  2. The Hurwitz bound fails in positive characteristic.

Proof.   You can check that X/Aut(X)=YX/\operatorname{Aut}(X) = Y is again a curve. So you get a covering XYX\to Y of degree Aut(X)\operatorname{Aut}(X)|. Ramification points of this cover are the points with non-trivial stabilizers.

Denote by ri,jr_{i,j} the ramification points lying over a branch point pip_i for 1ib1\leq i\leq b. (A picture of a line representing the base curve YY with a tick for point pip_i lying in YY. Above pip_i is a loopy curved line representing XX so that each point ri,jr_{i,j} is a ramification point. Each ri,1,ri,2,ri,3r_{i,1}, r_{i,2}, r_{i,3} is a place where the fiber is "tangent to the curve XX") Over pip_i all points in the fiber f1(pi)f^{-1}(p_i) have the same ramification index fif_i. If ri=f1(pi)r_i = |f^{-1}(p_i)| then firi=Aut(X)f_i\cdot r_i = |\operatorname{Aut}(X)|. Then Riemann-Hurwitz says

2g(X)2=Aut(X)(2g(Y)2)+iri(fi1)=Aut(X)(2g(Y)2)+ri,j(fi1)\begin{aligned} 2g(X) - 2 &= |\operatorname{Aut}(X)| \cdot (2g(Y) - 2) + \sum_{i}r_i\cdot (f_i - 1) \\ &= |\operatorname{Aut}(X)| \cdot (2g(Y) - 2) + \sum_{r_{i,j}}(f_i - 1) \end{aligned}

\square
 

©Isaac Martin. Last modified: February 14, 2024.