Gluing genus 1 and genus 2 curves along ℓ \ell -torsion
Pitchayut Saengrungkongka, Noah Walsh · 2025
Read Gluing genus 1 and genus 2 curves along ℓ \ell -torsion by Pitchayut Saengrungkongka, Noah Walsh (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-09753.txt.
Introduction
Curves can be glued together along shared points to build new curves of higher genus, a bit like sewing shapes along matching marks. Pitchayut Saengrungkongka and Noah Walsh give a search method for such gluings in this 2025 paper. Start with a genus two curve over the rational numbers. The hunt seeks a genus one partner curve and a prime so that a genus three curve exists whose linked Jacobian splits, up to a mild twist, into the two partners' Jacobians.
Earlier work handled the prime two. This method extends the search to larger primes, and the team found several pairs for primes up to thirteen. They also sharpened the numerical gluing method from the earlier work. With that gain, they glued curves along their thirteen-torsion points. Such constructions feed the study of curves and their Jacobians.
Catalog record
| CREATOR | Pitchayut Saengrungkongka, Noah Walsh |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 86 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-09753
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Gluing genus 1 and genus 2 curves along ℓ \ell -torsion Pitchayut Saengrungkongka, Noah Walsh Abstract . Let Y Y be a genus 2 2 curve over ℚ \mathbb{Q} . We provide a method to systematically search for possible candidates of a prime ℓ ≥ 3 \ell\geq 3 and a genus 1 1 curve X X for which there exists a genus 3 3 curve Z Z over ℚ \mathbb{Q} whose Jacobian is, up to quadratic twist, ( ℓ , ℓ , ℓ ) (\ell,\ell,\ell) -isogenous to the product of Jacobians of X X and Y Y , building on the work by Hanselman, Schiavone, and Sijsling for ℓ = 2 \ell=2 . We find several such pairs ( X , Y ) (X,Y) for prime ℓ \ell up to 13 13 . We also improve their numerical gluing algorithm, allowing us to successfully glue genus 1 1 and genus 2 2 curves along their 13 13 -torsion. 2020 Mathematics Subject Classification 11G30, 14H25, 14H40 1. Introduction The study of curves is a central topic in arithmetic geometry. Exhaustive lists of curves have long been a useful tool in number theory, starting with the Antwerp tables that contain elliptic curves up to conductor 200 [ BK75 ] , continuing with Cremona’s tables of elliptic curves [ Cre06 ] , and persisting today with the collections of elliptic curves and genus 2 curves in the LMFDB [ LMFDB ] . A large number of higher genus curves can be constructed by gluing curves of smaller genera. For any two curves X and Y of genera g_{X} and g_{Y} , the process of gluing produces a curve Z of genus g_{X}+g_{Y} such that the Jacobian of Z , denoted \Jac(Z) , is isogenous to the product \Jac(X)\times\Jac(Y) . In other words, \Jac(Z) is a quotient of \Jac(X)\times\Jac(Y) by a finite subgroup G . If G is a subgroup of the product of the n -torsion subgroups \Jac(X)[n] and \Jac(Y)[n] for a positive integer n , we say that this gluing is along n -torsion . Given generic curves X and Y , there are many possible choices of subgroups G that give rise to gluings over \mathbb{C} , and one can construct such gluings analytically by considering \Jac(X) and \Jac(Y) as complex tori. However, for arithmetic applications, we are interested in curves defined over non-algebraically closed fields such as \mathbb{Q} . If we take curves X and Y randomly, then it is very likely there will be no gluing Z that is defined over \mathbb{Q} . We are interested in systematically producing pairs of curves that admit a gluing over \mathbb{Q} . These curves often have interesting properties that deviate from generic behavior, such as a special torsion structure, a nontrivial endomorphism ring, or an interesting Sato-Tate group. 1.1. Previous Work The simpler case of gluing two curves of genus 1 (i.e., 1+1=2 ) along n -torsion has been studied in many aspects. (i) Criteria for gluings to exist. Frey and Kani [ FK91 ] derive the necessary conditions for two curves to admit a gluing along n -torsion, namely, if E_{1} and E_{2} are gluable, then there exists an antisymplectic, \Gal(\overline{\mathbb{Q}}/\mathbb{Q}) -equivariant isomorphism E_{1}[n]\to E_{2}[n] . (ii) Find all gluable pairs. Given a list of elliptic curves (e.g., all elliptic curves in the L-functions and modular forms database, LMFDB [ LMFDB ] ), we are interested in systematically searching for all pairs that are gluable. When n=\ell is prime, Cremona and Freitas [ CF22 ] gave a complete algorithm for detecting all such antisymplectic isomorphisms and applied their methods to all pairs of elliptic curves in the LMFDB. (iii) Parametrize all gluable pairs. Building on (ii), we ask if one can parameterize all pairs of elliptic curves that are gluable. For a fixed elliptic curve E_{1} , the space of all elliptic curves E_{2} that are gluable is one-dimensional and is a twist of the modular curve X(n) . For n\in\{2,3,4,5\} , the space is isomorphic to \mathbb{P}^{1} . The details have been worked out by Rubin and Silverberg in [ RS01 , RS95 , Sil97 ] . Fisher [ Fis14 ] derives explicit formulas for n\in\{7,11\} , but in that case there are finitely many E_{2} ’s for a fixed E_{1} . There are also efforts to parametrize pairs of gluable elliptic curve along n -torsion for higher n : Fisher [ Fis20 , Cor. 1.3] shows that there are infinitely many pairs of gluable curves for all n\leq 10 . (iv) Computing gluing. One way method of computing gluings over \mathbb{Q} is via analytic techniques over \mathbb{C} . In simpler cases, we also have explicit formulas. Howe, Leprévost, and Poonen [ HLP00 , Prop. 4] gave an explicit formula for n=2 , and Bröker et. al. [ BHLS15 , §A.1] gave a formula for n=3 . As \ell grows larger, the explicit formulas quickly become unwieldy. Genus 1 plus genus 2 gluing (i.e., 1+2=3 ) has not been studied as widely. Ritzenthaler and Romagny [ RR18 ] gave a formula for recovering the equation of the genus 2 factor of a genus 1 plus genus 2 gluing along 2 -torsion, provided that the genus 1 factor is known. Then Hanselman, Schiavone, and Sijsling [ HSS21 ] gave a comprehensive algorithm and explicit formula for gluing along 2-torsion. They answer some of the questions above for gluing genus 1 and 2 curves as follows. • To answer (i), they [ HSS21 , §1] provide a concrete criterion for a gluing along \ell -torsion to exist, which we recall in Section 2 . • To answer (iv), they [ HSS21 , §2.1] outline the analytic algorithm to compute the gluing along \ell -torsion, which was implemented in [ HSS20 ] . In the case of \ell=2 , they also provide an explicit formula to glue genus 1 and 2 curves along 2 -torsion. • To answer (iii), if \ell=2 , for a genus two curve Y , Hanselman [ Han20 , §2.2] constructs an infinite family of genus 3 curves resulting from gluing Y with an elliptic curve along their 2 -torsion. More generally, for a fixed genus 2 curve Y such that there exists an elliptic curve X that admits gluing with Y along \ell -torsion, one can reduce the problem of finding all gluable elliptic curves X^{\prime} to finding elliptic curves X^{\prime} whose \ell -torsion is symplectically isomorphic (as a \Gal(\overline{\mathbb{Q}}/\mathbb{Q}) -module) to the \ell -torsion of X . See Section 6.2 for more details. Our work attempts to answer (ii) and improves existing methods in (iv). 1.2. Our Results This paper considers the problem of gluing two curves of genus 1 and genus 2 along their \ell -torsion to obtain a curve of genus 3 for a prime \ell . We focus on the case of \ell\geq 3 that has not been as widely studied. There are two key results. The first key result is that, given a curve Y of genus 2 , we demonstrate how to systematically and efficiently search for genus 1 curves X in the LMFDB and corresponding primes \ell such that X and Y are gluable along \ell -torsion. More specifically, given a