4
$\begingroup$

This comes from Ex IV.4.2 in Hartshorne. Namely we want to show that if $D$ is a divisor of degree at least three on an elliptic curve $X$, and we embed $X$ into $\mathbb{P}^n$ via this very ample divisor, then the image is projectively normal.

I want to use the trace condition, namely that for every $d>0$, the natural map $$\Gamma(\mathbb{P}^n,\mathcal{O}(d)) \to \Gamma(X,i^*\mathcal{O}(d))$$ is surjective (this is essentially Ex II.5.14(d)). I can show this for $d=1$, but I'm not sure how to show this more generally. I figure something cohomological or the $d$-Uple embedding might help, but I'm just not sure.

Any ideas?

$\endgroup$
2
+100
$\begingroup$

A solution to this is given in Eisenbud's book "The Geometry of Syzygies", section 6.D.

The upshot is that for an elliptic curve $X$ and a point $p\in X$, the natural map $H^0(X, \mathcal{O}(dp)) \otimes H^0(X, \mathcal{O}(ep)) \rightarrow H^0(X, \mathcal{O}((d+e)p))$ is surjective, for $d\ge 2$ and $e\ge 3$. The proof of this is similar to how one puts an elliptic curve in Weierstrass form. Using Riemann-Roch, choose $x\in H^0(\mathcal{O}(2p))$ with a pole at $p$ of order $2$, and $y\in H^0(\mathcal{O}(3p))$ with a pole at $p$ of order $3$. Then you can construct a basis for $H^0(\mathcal{O}(np))$ for any $n$ by considering $x^i$ and $x^jy$. Then it is not too hard to show surjectivity by counting dimensions and using that rational functions with different order of vanishing at $p$ have to be linearly independent.

Let us suppose that this is true. We will prove the trace condition by induction. Consider the following commutative diagram:

$$\begin{array}{ccc} H^0(\mathbb{P}^n, \mathcal{O}(1)) \otimes H^0(\mathbb{P}^n, \mathcal{O}(d-1)) & \longrightarrow & H^0(X, \mathcal{O}_X(1)) \otimes H^0(X, \mathcal{O}_X(d-1))\\ \downarrow && \downarrow \\ H^0(\mathbb{P}^n, \mathcal{O}(d)) & \longrightarrow & H^0(X,\mathcal{O}_X(d)) \end{array}$$

The top arrow is surjective by inductive hypothesis, and the arrow on the right is surjective by the above fact (note that for any divisor $D$ of degree $n$ on an elliptic curve $X$, we have that $D = nP$ for some $P$) which implies that the bottom arrow is surjective.

$\endgroup$
  • $\begingroup$ This is encouraging, I had discovered that the map $H^0(X,\mathcal{O}_X(p))^{\otimes k\} \to H^0(X,\mathcal{O}_X(kp))$ was surjective, but wasn't able to put it all together. So close! Thank you for your help. $\endgroup$ – DKS Apr 9 '18 at 13:10

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.