Let, $C$ be a projective curve (not necessarily reduced), $i:C \to \mathbb{P}^n$ be a closed immersion. Does there exist a bound on/geometric interpretation of the dimension of $H^0(i^*(\mathcal{O}_{\mathbb{P}^n}(1))$?
1 Answer
Thinking about $i^\ast \mathcal O_{\mathbb{P}^n}(1)$ as $\mathcal O_C(1)$, there is a natural exact sequence
$$0 \to I_C(1) \to \mathcal O_{\mathbb{P}^n}(1) \to \mathcal O_C(1) \to 0.$$
Assuming $C$ is non-degenerate, we obtain an exact sequence in cohomology
$$0\to H^0(\mathcal O_{\mathbb{P}^n}(1))\to H^0(\mathcal O_C(1))\to H^1( I_C(1))\to 0.$$
It folllows that $h^0(\mathcal O_C(1))\geq n+1$, with equality if and only if the $H^1$ group above vanishes. The vanishing of this $H^1$ is called linear normality of the embedded curve. There are several different properties which are equivalent to linear normality of an embedded curve, some of which are more or less geometric:
(1) Hyperplanes cut a complete linear series on $C$.
(2) $C$ is embedded by a complete linear series.
(3) $C$ is not a projection of a non-degenerate embedding from a higher-dimensional projective space.
(4) $H^1(I_C(1))=0$
In practice, the fourth property is usually the most useful for attempting to show a curve is linearly normal, as you can compute this cohomology group via other methods, such as resolving the ideal sheaf.
-
1$\begingroup$ Just to explicate, linearly normal means that the embedding is not the projection of an embedding in a bigger projective space. I think most people would view that as the geometric explanation. $\endgroup$– mehCommented Nov 7, 2013 at 19:12
-
$\begingroup$ @Huizenga: Could you give an example of linearly normal curve? $\endgroup$– ChenCommented Nov 7, 2013 at 19:19
-
$\begingroup$ @agniesky: Yes, I suppose that is a more satisfactory geometric condition. Edited. $\endgroup$ Commented Nov 7, 2013 at 19:24
-
$\begingroup$ @Chen: As the edit now says, a curve is linearly normal iff it is embedded by a complete linear series. $\endgroup$ Commented Nov 7, 2013 at 19:25
-
$\begingroup$ @Huizenga: Another small question. I am not able to find a good reference for non-degenrate curve. Does it mean that the curve is not contained in a linear subspace of $\mathbb{P}^n$? $\endgroup$– ChenCommented Nov 7, 2013 at 19:41