2025-11-12T08:04:09.503033

Counting rational points on elliptic and hyperelliptic curves over function fields

Gillibert, Hallouin, Levin
Combining $2$-descent techniques with Riemann-Roch and Bézout's theorems, we give an upper bound on the number of rational points of bounded height on elliptic and hyperelliptic curves over function fields of characteristic $\neq 2$. We deduce an upper bound on the number of $S$-integral points, where $S$ is a finite set of places. As a primary application, over small finite fields we bound the $3$-torsion of Jacobians of hyperelliptic curves and the $2$-torsion of Jacobians of trigonal curves. In this setting, these bounds improve on both the trivial geometric bound and the naive inequality coming from the Weil bound, as well as recent upper bounds on $2$-torsion in the work of Bhargava et al.
academic

Counting rational points on elliptic and hyperelliptic curves over function fields

Basic Information

  • Paper ID: 2510.13292
  • Title: Counting rational points on elliptic and hyperelliptic curves over function fields
  • Authors: Jean Gillibert, Emmanuel Hallouin, Aaron Levin
  • Classification: math.NT (Number Theory), math.AG (Algebraic Geometry)
  • Publication Date: October 2025
  • Paper Link: https://arxiv.org/abs/2510.13292

Abstract

This paper combines 2-descent techniques with the Riemann-Roch theorem and Bézout's theorem to establish upper bounds on the number of rational points of bounded height on elliptic and hyperelliptic curves over function fields of characteristic ≠ 2. The authors further derive upper bounds on the number of S-integral points, where S is a finite set of places. As main applications, the paper bounds the 3-torsion of Jacobians of hyperelliptic curves and the 2-torsion of Jacobians of cubic curves over small finite fields. These bounds improve upon trivial geometric bounds, naive applications of the Weil bound, and recent upper bounds of Bhargava et al. on 2-torsion.

Research Background and Motivation

  1. Core Problem: Investigating the counting problem for rational points on elliptic and hyperelliptic curves over function fields, particularly estimating the number of rational points of bounded height and S-integral points.
  2. Problem Significance:
    • Rational point counting is a fundamental problem in arithmetic geometry, closely related to the function field version of the Mordell conjecture
    • For elliptic curves, this directly connects to the study of the structure of the Mordell-Weil group
    • Applications over finite fields can bound torsion subgroups of Jacobians, which is of significant importance in algebraic and arithmetic geometry
  3. Limitations of Existing Methods:
    • Classical geometric bounds are often too crude
    • While the Weil bound is precise, it is not sharp in certain cases
    • Existing torsion bounds can be further improved in the small finite field case
  4. Research Motivation: To obtain more precise counting bounds through geometric methods (primarily analysis of 2-descent maps) and apply them to torsion estimation of Jacobians over finite fields.

Core Contributions

  1. Established new upper bounds for rational point counting: For elliptic and hyperelliptic curves over function fields, explicit upper bounds on the number of rational points of bounded height are provided (Theorem 1.1)
  2. Derived counting bounds for S-integral points: Using the function field version of the abc theorem, upper bounds on the height of S-integral points are given, yielding estimates on the number of S-integral points (Theorem 1.2)
  3. Improved torsion bounds for Jacobians over finite fields:
    • Bounds of qg/2+γg/log⁡gq^{g/2+\gamma g/\log g} for 3-torsion of hyperelliptic curve Jacobians (Theorem 1.3)
    • Bounds of (2q)g/3+γg/log⁡g(2q)^{g/3+\gamma g/\log g} for 2-torsion of cubic curve Jacobians (Theorem 1.6)
  4. Provided characteristic-independent geometric methods: The main techniques are based on the Riemann-Roch theorem, applicable to arbitrary fields of characteristic ≠ 2

Detailed Methodology

Problem Setup

Study rational points on hyperelliptic curves C:y2=f(x)C: y^2 = f(x) over function fields k(B)k(B), where ff is a monic separable polynomial of odd degree d≥3d \geq 3. The main objectives are to estimate:

  • The size of the set of rational points of bounded naive height C(k(B))≤c:={(x0,y0)∈C(k(B))∣deg⁡x0≤c}C(k(B))_{\leq c} := \{(x_0, y_0) \in C(k(B)) | \deg x_0 \leq c\}
  • The size of the S-integral point set C(RS)C(R_S)

Core Technical Framework

Analysis of 2-Descent Maps

Definition of 2-descent map (Lemma 2.3): δ:C(k(B))→H1(Cf∖π−1(Σ2),μ2)\delta: C(k(B)) \to H^1(C_f \setminus \pi^{-1}(\Sigma_2), \mu_2)

where CfC_f is the curve defined by f(x)=0f(x) = 0, and Σ2\Sigma_2 is the set of places where the connected component group of the Néron model has even order.

Key Observation: The number of rational points mapping to the same cohomology class is finite and controllable.

Main Counting Theorem

Theorem 2.4 provides an upper bound on the number of points mapping to the same class: 2Ω(c,f,g)+max⁡{c,12(c+g)}+12^{\Omega(c,f,g)+\max\{c, \frac{1}{2}(c+g)\}+1}

where Ω(c,f,g)\Omega(c,f,g) is defined as:

\max\{d(c+g) + h_B(f) - g_f, \frac{1}{2}(d(c+g) + h_B(f))\} & \text{if }C_f\text{ is irreducible} \\ d(c+g) + h_B(f) + \omega(f) - 1 & \text{otherwise} \end{cases}$$ #### Technical Innovations 1. **Geometric counting method**: Utilizing function counting in linear systems, controlling dimension via the Riemann-Roch theorem 2. **Application of Bézout's theorem**: Analyzing intersections of two algebraic varieties in projective space to obtain upper bounds on the number of irreducible components 3. **Optimization for special cases**: - Improvements when the base curve is $\mathbb{P}^1$ and $f$ has coefficients in $k[t]$ (Proposition 2.7) - Refinements for cubic curves using the Maroni invariant (Theorem 3.7) ### S-Integral Point Height Bounds **Theorem 3.1** uses the abc theorem to provide S-integral point height bounds: - Characteristic 0 case: $\deg x_0 \leq 4(2g-2+|S \cup \Sigma|) + \frac{3h_B(f)}{d}$ - Characteristic $p > d$ case: $\deg x_0 \leq 6\rho(2g-2+|S \cup \Sigma|) + \frac{3h_B(f)}{d}$ where $\rho$ is the inseparable degree, reflecting arithmetic properties of the curve. ## Experimental Setup ### Theoretical Verification Framework Since this is pure mathematical theory, there are no traditional "experiments," but the authors verify the theory through: 1. **Concrete application verification**: Applying abstract bounds to torsion estimation of Jacobians over finite fields 2. **Comparison with known results**: Comparing with Weil bounds, geometric bounds, and results of Bhargava et al. 3. **Special case analysis**: Providing more precise bounds for specific cases such as elliptic curves $y^2 = x^3 + B(t)$ ### Parameter Settings - **Finite field size**: $q = p^r$, where $p \geq 5$ - **Curve genus**: Primarily considering $g > 4$ to apply Maroni theory - **Height bounds**: $c$ typically chosen as an appropriate value related to curve invariants ## Main Results ### Core Theorem Results **Theorem 1.1** (Main counting bound): For hyperelliptic curves $C: y^2 = f(x)$, the number of rational points of bounded height satisfies: $$|C(k(B))_{\leq c}| \leq 2^{\Omega(c,f,g)+\max\{c, \frac{1}{2}(c+g)\}+1+\text{rk}_{\mathbb{Z}} LN(J)+\dim_{\mathbb{F}_2} LN(J)[2]}$$ **Theorem 1.3** (3-torsion bound): For hyperelliptic curves $X$ of genus $g$ over finite fields $\mathbb{F}_q$: $$|\text{Jac}(X)(\mathbb{F}_q)[3]| \leq q^{g/2+\gamma g/\log g}$$ **Theorem 1.6** (2-torsion bound): For cubic curves $X$: $$|\text{Jac}(X)(\mathbb{F}_q)[2]| \leq (2q)^{g/3+\gamma g/\log g}$$ ### Analysis of Improvements 1. **Compared to trivial bounds**: When $q < 81$, the 3-torsion bound asymptotically improves the trivial bound $3^{2g}$ 2. **Compared to Weil bound**: Improves upon the bound $(√q+1)^{2g}$ 3. **Compared to Bhargava et al.**: Asymptotically improves their 2-torsion bound for cubic curves ### Concrete Numerical Examples - When $q < 9$ and $q \not\equiv 1 \pmod{3}$, improves the bound $3^g$ given by Galois invariance - When $q < 32$, improves the trivial bound $2^{2g}$ for 2-torsion of cubic curves ## Technical Depth Analysis ### Application of Maroni Invariant For cubic curves $C$, the Maroni invariant $m$ satisfies: $$\frac{g-4}{3} \leq m \leq \frac{g-2}{2}$$ **Corollary 3.5** uses this invariant to provide precise bounds on linear system dimension: $$h^0(D) \leq \max\left\{\min\left\{m + \deg D + 2 - g, \frac{2(\deg D - 1) - g}{3} + 2\right\}, 1\right\}$$ ### Role of Inseparable Degree In positive characteristic, the inseparable degree $\rho$ is defined as: $$\rho := \inf\left\{\text{ideg}_K\left(\frac{e_k - e_1}{e_2 - e_1}\right) | k = 3,\ldots,d\right\}$$ This quantity controls the behavior of the curve under base field extensions and directly affects the bounds on S-integral point heights. ## Related Work ### Historical Background 1. **Function field version of Mordell's conjecture**: Resolved by Manin, Grauert, and others in the 1960s 2. **Height theory**: Developed by Silverman and others for elliptic curves 3. **Integral point counting**: Classical work by Hindry-Silverman, Pacheco, and others ### Advantages of This Work 1. **Generality of method**: Applicable to arbitrary fields of characteristic ≠ 2, without requiring algebraic closure 2. **Explicit bounds**: Provides concrete exponents suitable for applications 3. **Geometric approach**: Primarily based on Riemann-Roch, avoiding complex arithmetic analysis ## Conclusions and Discussion ### Main Conclusions 1. Established geometric methods for counting rational points on hyperelliptic curves over function fields 2. Significantly improved bounds on Jacobian torsion in the small finite field case 3. Provided a systematic approach from rational point counting to torsion estimation ### Limitations 1. **Characteristic restriction**: Methods do not apply to characteristic 2 2. **Asymptotic nature**: Bounds are primarily improvements in the asymptotic sense and may not be sharp for small genus 3. **Geometric assumptions**: Requires technical assumptions such as algebraic closure of the base field ### Future Directions 1. **Extension to characteristic 2**: Requires replacing étale cohomology with flat cohomology 2. **Higher torsion**: Extending to the general case of p-torsion 3. **Number field analogue**: Adapting the method to analogous problems over number fields ## In-Depth Evaluation ### Strengths 1. **Technical innovation**: Cleverly combines methods from algebraic geometry (Riemann-Roch) and number theory (2-descent) 2. **Precise results**: Provides explicit exponential bounds with practical application value 3. **Complete theory**: Forms a complete theoretical system from abstract counting to concrete applications 4. **Clear exposition**: Rigorous arguments with sufficient technical detail ### Weaknesses 1. **Limited scope**: Exclusion of characteristic 2 restricts the generality of the method 2. **Dependence**: Some results rely on deep theorems such as Brumer's rank bounds 3. **Room for optimization**: Certain bounds may admit further improvements ### Impact 1. **Theoretical contribution**: Provides new tools for arithmetic research on curves over function fields 2. **Application value**: Potential applications in cryptography, coding theory, and related fields 3. **Methodological significance**: Demonstrates the power of geometric methods in number-theoretic problems ### Applicable Scenarios 1. **Theoretical research**: Interdisciplinary research between algebraic geometry and number theory 2. **Computational applications**: Algorithms requiring estimates of properties of curves over finite fields 3. **Further generalizations**: Provides methodological guidance for related problems ## References The paper cites 32 important references, including: - Classical work of Brumer on elliptic curve ranks - Pioneering research of Hindry-Silverman on integral points - Recent results of Bhargava et al. on torsion - Classical theory of Maroni on cubic curves These references form a solid foundation for the research, reflecting the authors' deep understanding of the field's development.