MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I get stuck with what Example II.7.6.3 (Page 156 in Hartshorne's book Algebraic Geometry) claims. Let me recite the example here: Let $X$ be an elliptic curve $y^2z = x^3 - xz^2$ in $\mathbf P_k^2$ defined over an algebraically closed field $k$ with characteristic $\ne 2$. Let ${\scr L}={{\scr L}}(P_0)$ be the invertible sheaf corresponding to the divisor $(P_0)$ where $P_0=\infty$ (I think the claim is true for any closed point on the curve). Then this example claims $\scr L$ is not very ample because $\scr L$ is not generated by global sections and the reason why it is not generated by global sections is "if it were, then $P_0$ would be linearly equivalent to some other point $Q \in X$". I do not see why this is true. Could somebody give some hints, or answers?

Thank you CamSar to provide a detailed argument which gives further insights. Here I want to prove CamSar's argument "$h^0((X, \mathcal O_X(P_0))) \ge 3$ implies that there exists an effective divisor $Q$ linearly equivalent to $P_0$". Since the elliptic curve is defined over a field $k$, $H^0(X, \mathcal O_X(P_0))$ is a vector space over $k$ with finite dimension $\ge 3$ by assumption. Fix a set of generators of this vector space, then each element induces a positive divisor $D$ equivalent to $(P_0) = (\infty)$. Since for any ration function $f$, $\mathrm{deg} (\mathrm{div}(f)) = 0$, $D=(Q)$ for some closed point $Q \in X$. Since the dimension of the vector space over $k$ is $\ge 3$, at least one element of generators induces a divisor $(Q)$ with $Q \ne P_0$.

share|cite|improve this question
OK, here is a hint. You can find an effective divisor linearly equivalent to $P_0$ but disjoint from it. – Donu Arapura May 2 '14 at 13:04
up vote 4 down vote accepted

If $\mathcal{O}_X(P_0)$ is very ample then it induce an embedding $f:X\rightarrow\mathbb{P}(H^{0}(X,\mathcal{O}_X(P_0))\cong\mathbb{P}^n$. Therefore $n\geq 2$ and $h^{0}((X,\mathcal{O}_X(P_0)))\geq 3$. Now, $h^{0}((X,\mathcal{O}_X(P_0)))\geq 3$ implies that there exists an effective divisor $Q$ linearly equivalent to $P_0$. In particular $P_0-Q$ is the divisor of a rational function $g = \frac{p}{q}$ on $X$. The rational map $$\phi:X\dashrightarrow\mathbb{P}^1,\: x\mapsto [p(x):q(x)]$$ is birational. This implies that $X$ is rational. A contradiction because $g(X) = 1$. Roughly speaking we are saying here that if a divisor of degree one on a curve $X$ is very ample then the curve $X$ is mapped to a curve of degree one ($deg(P_0) = 1$) in a projective space, that is a line. Therefore $g(X) = 0$. In this way you get the following.

Let $X$ be a curve and $P_0\in X$ be a point. Then $\mathcal{O}_X(P_0)$ is very ample if and only if $g(X) = 0$.

From the first part of the argument you see that $\mathcal{O}_X(D)$ very ample imply $h^{0}((X,\mathcal{O}_X(D)))\geq 3$. You may use Riemann-Roch theorem (pag 294) with $D = P_0$. In this case $deg(K_X-P_0) = 2g-2-1 = -1$. Then $P_0$ is non-special and $h^{1}(X,\mathcal{O}_X(P_0)) = 0$. By Riemann-Roch yu get $$\chi(\mathcal{O}_X(P_0)) = h^0(\mathcal{O}_X(P_0)) = deg(P_0)-g+1 = 1.$$ Again you find a contradiction.

In general, a divisor $D$ on a curve $X$ is ample if and only if $deg(D)>0$. If $g(X) = 0$ then $D$ is ample if and only if it is very ample if and only if $deg(D)>0$. However, if $g(X)=1$ then $D$ is very ample if and only if $deg(D)\geq 3$. Indeed, by Riemann-Roch, $deg(D)\leq 2$ implies $h^{0}(X,\mathcal{O}_X(D)) = deg(D)\leq 2$. Therefore the morphism induced by $D$ can not be an embedding. On the other hand if $deg(D)\geq 3\geq 2g+1$ then $D$ is very ample.

share|cite|improve this answer
Thank you CamSar, this helps to understand this subject deeper. Also I give my thanks to Donu Arapura for the hints with reminds me the linear system of a divisor. – user41541 May 3 '14 at 17:11

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.