# Extending maps of curves

(I'm happy to work over an algebraically closed field....)

Let $\mathcal{C} \rightarrow Spec (R)$ be a (flat) family of proper, prestable curves where $R$ is a DVR. Suppose the generic fiber is smooth and the special fiber, $C_0$, is reduced but may be reducible.

Given a finite map of curves $f_0: D_0 \rightarrow C_0$ with $D_0$ also prestable, can this be extended to some map on some family?

That is, is there a flat family of proper curves $\mathcal{D} \rightarrow Spec(R)$ and an $R$-morphism $f: \mathcal{D} \rightarrow \mathcal{C}$ which reduces to $f_0$ on the special fiber?

Perhaps such an extension is possible only after a ramified cover of $Spec(R)$?

If so, can it be arranged that the generic fiber of $\mathcal{D}$ is smooth?

-

## 2 Answers

Stated as such, the question is really too vague. For example, if $C_0, D_0$ are smooth of genus bigger than 1, and $f_0$ is purely inseparable, then $f_0$ can not lift to $R$ (even after extention) if $R$ has characteristic zero. Actually, as the generic fibers of $\mathcal C$ and $\mathcal D$ are equal to $g(C_0)=g(D_0)$, this will contradict Riemann-Hurwitz formula.

In this survey paper you can find some sufficient conditions. For instance, the lifting property for finite generically étale morphisms of semi-stable curves are characterized.

-
My comment "too vague" was for the first version of the question... –  Qing Liu Jan 25 '10 at 22:04
I will take a look at this reference... what happens if I keep making the question more specific, like requiring $f_0$ to be separable, or else the genus is small? –  mdeland Jan 25 '10 at 22:07
The answer for f_0 separable (=generically étale) is in the paper. What do you mean by small genus ? –  Qing Liu Jan 25 '10 at 22:47
Say, g = 0 or g = 1. –  mdeland Jan 25 '10 at 23:20
Ah great! Section 5 of your survey is exactly the kind of thing I was looking for. –  mdeland Jan 25 '10 at 23:24

I don't think you can extend an arbitrary morphism, even if the curves are prestable.

For instance, if $C_0$ is a quartic with one node and the generic element is a smooth quartic then you do have a morphism from a smooth curve of genus $2$ to $C_0$ which does not extend, since any extension would map a curve of genus $2$ to a curve of genus $3$.

-
Yes great point - let me ask only about prestable curves - so that the singularities are at worst nodes. –  mdeland Jan 25 '10 at 21:51
I have slightly changed my answer to deal with prestable curves. –  jvp Jan 25 '10 at 22:31
Yes, your comment explains why the condition in Qing Liu's referenced paper above, 4.11a is a necessary one. Thanks! –  mdeland Jan 26 '10 at 13:48