5
$\begingroup$

Suppose I have an elliptic curve $E$ defined over a number field $K$. I know that if it has

  • a $2$ $K$-torsion, it has a model of the form:

$E: Y^2=X^3+aX^2+bX$

  • a $3$ $K$-torsion, it has a model of the form:

$E: Y^2 +cXY +dY=X^3$

My question is, do we have a nice description for elliptic curves with a $p$ $K$-torsion where $p$ is any rational prime?

Any help or reference would be very much appreciated!

$\endgroup$
1
  • 1
    $\begingroup$ Are you looking for an explicit equation for the universal elliptic curve over the modular curve $X_1(p)$? That is going to be difficult as even the equation for the modular curve is going to be very complicated. On the other hand you can always shift the curve in such a way that the $p$-torsion point is at $(0,0)$, but that only says $a_6=0$. $\endgroup$ Commented Apr 5, 2022 at 19:24

2 Answers 2

6
$\begingroup$

A standard method to find such an equation (in principle) is the following, shown to me by Tate, but maybe due to Mordell(?). If $P_0$ is a point of order at least $4$ on $E$ (infinite order is allowed), then after a change of coordinates, we can find an equation for $E$ of the form $$ E : y^2 + u x y + v y = x^3 + v x^2 \quad\text{with}\quad P_0=(0,0). $$ (It's a nice exercise to check this assertion.) So now we start with this equation and point, and we want $P_0$ to have order $N$. If $N$ is odd, say $N=2n+1$, then use the group law on $E$ to compute $x(nP_0)$ and $x\bigl((n+1)P_0\bigr)$. These will each be rational functions of $u$ and $v$; i.e., they're in $\mathbb Q(u,v)$. Setting $$ x(nP_0)=x\bigl((n+1)P_0\bigr)$$ and clearing the denominators will give you a polynomial in $(u,v)$ whose vanishing is an affine plane model for $X_1(N)$ if $N$ is prime. However it tends to be rather singular at points where the discriminant of $E$ vanishes. And if $N$ is composite, then the polynomial will factor, i.e., the curve will have components for each divisor of $N$ greater than $3$. Finally, if $N=2n$ is even, you can do the same thing with the decomposition $N=(n+1)+(n-1)$, or you could set the numerator of $(2y+ux+v)(nP_0)$ equal to $0$, which is another way of forcing $2nP_0$ to be $O$.

$\endgroup$
2
  • $\begingroup$ This method has been used before, not with the aim to find equations of $X_1(N)$, but to determine elliptic curves over $\mathbb{Q}$ with a given torsion subgroup. For example Billing and Mahler in "On exceptional points on cubic curves" (Journal of the LMS, vol. 15, 1940) showed that an elliptic curve over $\mathbb{Q}$ cannot have a rational point of order $11$. In the course of doing so, they prove in Lemma 2 (p. 41) that a certain cubic curve with explicit equation has exactly 5 rational points, the curve being of course $X_1(11)$. $\endgroup$ Commented Apr 6, 2022 at 9:01
  • $\begingroup$ @FrançoisBrunault Thanks, that's probably the paper that I was trying to recall when I mentioned Mordell's name. $\endgroup$ Commented Apr 6, 2022 at 10:46
3
$\begingroup$

For a given integer $N \geq 1$, the elliptic curves endowed with a point of order $N$ are parametrized by the modular curve $Y_1(N)$. For $N \geq 4$, there is a universal elliptic curve $E \to Y_1(N)$ and you are asking whether $E$ has a nice description (e.g. Weierstrass equation). To get started with the theory, you could look at these notes by Parson, which include the case $p=5$ (these notes are part of a 2003-04 working group organized by B. Conrad and Parson, for which you can find notes here). If you are after explicit equations for small values of $N$, there is an article by Baaziz, Equations for the modular curve $X_1(N)$ and models of elliptic curves with torsion points, Mathematics of Computation 79 (2010), no. 272, 2371–2386.

$\endgroup$

You must log in to answer this question.

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