5
$\begingroup$

Let $C$ be a general genus $g$ curve, how can we describe the Neron-Severi group of its $n$-th self product $C^n=C\times \dots \times C$?

It is a lattice in $H^2(C^n,\mathbb{Z})\cong \mathbb{Z}^{n+2n(n-1)g^2}$, do we know its rank and basis if $C$ is general?

$\endgroup$

1 Answer 1

7
$\begingroup$

Let me give the answer for $n=2$.

Fix a point $p \in C$ and call $x_1, \, x_2, \, \Delta$ the divisor classes

$\begin{equation*} \begin{split} x_1 & := \{(p, \, x) \, | \, x \in C\} \\ x_2 & := \{(x, \, p) \, | \, x \in C \} \\ \Delta & := \{(x, \, x) \, | \, x \in C \}. \end{split} \end{equation*} $

Let $C^{(2)}$ be the second symmetric product of $C$, namely the quotient of $C^2$ by the action of $\mathbb{Z}/2 \mathbb{Z}$ exchanging the two factors, and let $x$, $\delta$ be the images of $x_1$ (or $x_2$) and $\Delta$ in $C^{(2)}$, respectively.

It is a classical result that for a general curve $C$ of positive genus the Neron-Severi group of $C^{(2)}$ is generated by $x$ and $\delta$, hence a basis for the Neron-Severi group of $C^2$ is given by $$x_1, \, x_2, \, \Delta.$$ The same result holds for the Neron-Severi group of the $n$-fold symmetric product $C^{(n)}$, hence one shows that the Neron Severi group of $C^n$ is generated, for general $C$, by the coordinate divisors and by the classes of the diagonals.

References. A. Kouvidakis: Divisors on symmetric product of curves, Trans. Amer. Math. Soc. 337 (1993), 117-128.

$\endgroup$
2
  • 4
    $\begingroup$ Do you mean general in the sense of algebraic geometry (omitting a proper Zariski closed subset), or do you mean "very general" as in omitting a countable union of proper algebraic subvarieties? (For $g = 1$, wouldn't all CM curves need to be excluded?) $\endgroup$ Commented Oct 28, 2015 at 17:01
  • 1
    $\begingroup$ Right, "general" here actually means "very general", i.e. lying in the complement of a countable union of closed subvarieties of the moduli space $\mathcal{M}_g$. $\endgroup$ Commented Oct 28, 2015 at 18:14

You must log in to answer this question.