8
$\begingroup$

I posted this question on MSE a few days ago, but I did not get much interest in it. So I thought I would try my luck here. If you are interested in answering the question, there is a bounty over on MSE.

https://math.stackexchange.com/questions/2911318/presentation-for-a-finite-etale-cover-of-an-affine-elliptic-curve


I will write the question first, then try and explain myself more clearly after:

Question: How can one find a presentation for a finite etale cover of an affine piece of an elliptic curve?

If $E/\mathbb{C}$ is an elliptic curve, then it is homeomorphic to a torus. For this reason we know $\pi_{1}(E) \cong \mathbb{Z}\times \mathbb{Z}$. If we puncture $E/\mathbb{C}$, then we obtain an affine curve, let us denote it $X$. Moreover, $X$ can be thought of as the elliptic curve minus the point at infinity. Since this is affine, we can give a presentation for it, say, $$X:= \text{Spec}(\mathbb{C}[x,y]/\langle y^{2} - f(x) \rangle)$$ for $f(x) = x(x-1)(x-2)$.

We know $\pi_{1}^{et}(X) \cong \hat{\pi}_{1}(X(\mathbb{C})^{an})$ and $X(\mathbb{C})^{an}$ is homeomorphic to a punctured torus. The fundamental group of a punctured torus is the free group on two generators $F_{2}$. The profinite completion of which is non-trivial. So, there are algebraic covers of such an affine curve. How can we get our hands on them? Is there a presentation which in some sense is in terms of $f(x)$? Does anyone have a reference for a discussion on such a construction?

Thanks in advance :)

$\endgroup$
6
  • $\begingroup$ If I'm not mistaken, in this situation any etale cover will be a smooth algebraic curve $C$ and this curve has to be affine. Now the cover extends to a covering $\overline{C} \to E$, it may be ramified but only above $0$. Now you can use the Riemann-Hurwitz formula to get some necessary conditions for such a cover to exist. $\endgroup$ Commented Sep 12, 2018 at 22:27
  • $\begingroup$ In fact, you're seeking for the analogue of the theory of "dessins d'enfants" for punctured elliptic curves. I think this is difficult and I'm not aware of any explicit general result. The étale covers of $E$ are easy to describe: they are given by isogenies and explicit equations can be computed using Vélu's formulas. On the other hand the universal cover of $E-\{0\}$ is isomorphic to the upper-half plane $\mathcal{H}$, so $E-\{0\}=\mathcal{H}/\Gamma$ for some discrete subgroup $\Gamma$ of $\mathrm{PSL}_2(\mathbb{R})$, but to my knowledge no one knows how to compute it. $\endgroup$ Commented Sep 13, 2018 at 14:40
  • 1
    $\begingroup$ I expect this is just as difficult as computing equations for covers of $\mathbb{P}^1 - \{0,1,\infty\}$. To get a sense of the difficulties involved in the latter problem, you might be interested in this article by Sijsling and Voight $\endgroup$
    – Will Chen
    Commented Sep 14, 2018 at 0:32
  • $\begingroup$ @FrançoisBrunault There is at least one punctured elliptic curve for which the discrete subgroup of $\text{PSL}_2(\mathbb{R})$ is "easy" to compute - it is the commutator subgroup of $\text{SL}_2(\mathbb{Z})$, which is a torsion-free congruence subgroup of index 12, level 6, a single cusp, and with $j$-invariant 0. $\endgroup$
    – Will Chen
    Commented Sep 14, 2018 at 0:35
  • 2
    $\begingroup$ @FrançoisBrunault Sorry for the late reply, but two references for your question are given on the second page (Rhodes + Petersson) in this interesting article: ams.org/journals/proc/2008-136-07/S0002-9939-08-09262-9/… $\endgroup$
    – Will Chen
    Commented Oct 8, 2018 at 19:44

1 Answer 1

4
$\begingroup$

If you are just looking for examples you can consider the following construction, which is geometric although I don't care to write down the equations explicitly.

Recall, that if $C$ is a compact Riemann surface and $F$ is a reduced divisor of degree $f$ then there is a $f$-fold cyclic cover of $C$, branched at $F$, which we denote by $D$. That is, there is a degree $f$ map: $$ \pi\colon D \rightarrow C $$ which has branching of order $f$ at each of the points of $F$. (Remark: this cover depends on a choice of $f$-th tensor root of the line bundle $\mathcal{O}_C(E)$. There are always $f$-th roots as the map: $$ (-)^{\otimes f}\colon\mathrm{Pic}_1(C) \rightarrow \mathrm{Pic}_f(C) $$ is surjective.)

Using this we can construct the following étale covers of the punctured torus. Start by considering an étale $d$-sheeted cover of your elliptic curve $E$: $$ \phi\colon C \rightarrow E. $$ (Thus, $C$ is a genus 1 curve as well.) Let $x\in E$ be a point and consider a reduced divisor of degree $f$, $F\subset\phi^{-1}(x)$, which is supported on the preimage of $x$ in $C$. Taking the cyclic cover branched at $F$ we get: $$ D\xrightarrow{\pi}C\xrightarrow{\phi}E. $$ Then the composition $\phi\circ \pi$ is a degree $d\cdot f$ map which is étale away from the point $x\in E$.

$\endgroup$

You must log in to answer this question.

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