11
$\begingroup$

In response to the affirmative answer to this question using symplectic methods, I am wondering if the equivalent statement holds in positive characteristic?

Explicitly, over an algebraically closed field $k$ of positive characteristic $p > 0$, does every smooth projective morphism $f : X \to \mathbb{P}^1$ admit a section?

If we do not require $k$ to be separably closed, the question becomes uninteresting because of examples such as $\mathbb{P}^1_{k'} \to \mathbb{P}^1_k$ for any finite separable $k'/k$.

My guess is there should exist an example with no section. My idea was to find a surface admitting a nontrivial elliptic fibration over $\mathbb{P}^1$ with smooth fibers. Do there exist such counterexamples?

$\endgroup$
1
  • 2
    $\begingroup$ I believe the symplectic proof of existence of holomorphic sections of a projective smooth morphism to $\mathbb{CP}^1$ also works if there is a single critical value. Glue the $2$-sphere to an "opposite" $2$-sphere along the boundaries of little disks centered at the critical values, and form the fibration over this new $2$-sphere by glueing the original fibration to its "opposite" fibration over the boundary of the disks. So one could ask about the positive characteristic case with a single critical value, but this fails since the fundamental group of $\mathbb{A}^1$ is nontrivial. $\endgroup$ Commented Nov 9, 2021 at 11:13

1 Answer 1

1
$\begingroup$

EDIT: It was wrong.

I think such examples exist, take as $f:Y\rightarrow \mathbb P^1$ any family of supersingular abelian varieties over $\overline{\mathbb F_p}$ which is non-constant but constant up-to isogeny (e.g. take the Moret-Bailly pencil of supersingular abelian surfaces). You cannot have a non-constant morphism $\mathbb P^1\rightarrow Y$, since the group $Y(\overline{\mathbb F_p}(t))$ is torsion (since the family is isotrivial up to isogeny) and any non-constant morphism $\mathbb P^1\rightarrow Y$ would correspond to a non-torsion element in $Y(\overline{\mathbb F_p}(t))$.

$\endgroup$
4
  • 1
    $\begingroup$ If $Y \to \mathbf P^1$ is a family of abelian varieties, doesn't that mean it has a zero section? $\endgroup$ Commented Nov 7, 2021 at 19:57
  • 1
    $\begingroup$ I think you want a nontrivial torsor for the family. $\endgroup$
    – A User
    Commented Nov 7, 2021 at 19:59
  • $\begingroup$ Ops...of course you are right... I don't know what I was thinking about... Sorry! $\endgroup$ Commented Nov 7, 2021 at 20:22
  • $\begingroup$ I doubt that there are nontrivial torsors for the family, since P^1 is simply connected, so one has to find a different route $\endgroup$ Commented Nov 7, 2021 at 20:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .