2
$\begingroup$

Given two smooth elliptic curves $C_1$ and $C_2$ over $\mathbb{C}$. Assume they are not isogenous. I'm interested in the structure of $Pic(A)$ and $Pic^{0}(A)$ for $A:=C_1 \times C_2$.

Reading Birkenhake/Lange - Complex Abelian Varieties, i think this has to do with correspondences of curves. Since an elliptic curve is its own Jacobian and the two curves are not isogenous, we have $Hom(C_1,C_2)=0$. So the space of correspondences $Corr(C_1,C_2)$ is trivial, i.e. every line bundle $L$ on $A$ is of the form $L=q^{\*}M\otimes p^{\*}N$, where q and p are the projections on the factors and $M$ and $N$ are line bundles on the factors. This implies $Pic(A)=Pic(C_1)\times Pic(C_2)$.

Does this impliy $Pic^{0}(A)=Pic^{0}(C_1)\times Pic^{0}(C_2)$? That is, is the Picard variety of $A$ isomorphic to $A$ in this case?

$\endgroup$
1
  • 1
    $\begingroup$ Yes, any product $A$ of elliptic cures is principally polarizable(e.g. by the product polarization), hence isomorphic to its Picard variety. $\endgroup$ Oct 7, 2010 at 18:01

2 Answers 2

2
$\begingroup$

Yes: in fact $Pic^0(C_1\times C_2)=Pic^0(C_1)\times Pic^0(C_2)$ for any pair of curves. The fact that $C_1$ and $C_2$ are not isogenous in your case only affects the Neron-Severi group $Pic/Pic^0$ of $C_1\times C_2$, exactly for the reasons you describe.

$\endgroup$
3
  • $\begingroup$ Interesting, can you give an explanation or a reference why this is true for all curves? I see that we have $Pic^{0}(C_1)\times Pic^{0}(C_2) \subset Pic^{0}(C_1\times C_2)$, via the pullbacks coming from the projections. But i don't see the other inclusion. $\endgroup$
    – TonyS
    Oct 7, 2010 at 17:33
  • 1
    $\begingroup$ It comes out of the theory of correspondences. See Weil's classic "Courbes algebriques et les varietes qui s'en deduisent", or p.153 of Zariski's book on algebraic surfaces (2nd ed.) $\endgroup$ Oct 7, 2010 at 18:38
  • $\begingroup$ Thank you. I will try to get a look in that book and that article. $\endgroup$
    – TonyS
    Oct 7, 2010 at 20:24
1
$\begingroup$

Am I missing something? Doesn't this hold for arbitrary smooth pairs of varieties using the Kunneth decomposition?

$\endgroup$
1
  • $\begingroup$ Not for the Picard groups. See Ex. 4.10, chap. IV in Hartshorne. $\endgroup$ Nov 1, 2021 at 10:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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