The Circle Method for Quadrics over Function Fields

Technical Summary: The Circle Method for Quadrics over Function Fields

1. Problem Statement
The paper addresses the problem of counting rational points of bounded height on quadric hypersurfaces defined over function fields. Specifically, it considers a non-singular quadratic form $f \in \mathbb{F}_q[x_1, \dots, x_n]$ over the rational function field $\mathbb{F}_q(t)$, where $q$ is a power of an odd prime ($p > 2$).

The primary objective is to determine the exact number of $\mathbb{F}_q(t)$-rational points on the projective hypersurface $X = V(f) \subseteq \mathbb{P}^{n-1}$ with "naive height" bounded by $q^P$. Geometrically, this is equivalent to counting the number of morphisms $\mathbb{P}^1 \to X$ of a fixed degree $P$ defined over $\mathbb{F}_q$. Unlike classical number field settings where asymptotic formulas are the norm, the author aims to derive exact formulas for these counts, including secondary terms that depend on the parity of $n$ and the arithmetic properties of the quadratic form's determinant.

2. Methodology
The author employs the Hardy–Littlewood–Ramanujan Circle Method adapted to the function field setting. The methodology proceeds through the following steps:

• Counting Function Formulation: The problem is translated into counting primitive zeros of $f$ in the polynomial ring $\mathbb{F}_q[t]^n$ with degree constraints. A Möbius inversion formula relates the count of primitive solutions to the total count of solutions $N(P)$.
• Exponential Sums: The counting function $N(P)$ is expressed as an integral over the unit circle (the space of additive characters) of an exponential sum $S(\alpha) = \sum \psi(\alpha f(x))$.
• Diophantine Approximation: Using the function field analogue of Dirichlet's approximation theorem, the integration domain is decomposed into "major arcs" (rational approximations $a/r$) and "minor arcs." In the function field setting, the structure of the local field $\mathbb{F}_q((t^{-1}))$ allows for a precise decomposition where the integral over the minor arcs vanishes or is negligible due to the ultrametric inequality.
• Evaluation of Gauss Sums: The core technical challenge involves evaluating the exponential sums over the major arcs. These reduce to quadratic Gauss sums $S_{a,r}(f)$. The author explicitly computes these sums using the properties of the Jacobi symbol and the structure of the ring $\mathbb{F}_q[t]/(r)$.
• Case Analysis: The evaluation of the Gauss sums depends critically on:
  1. The parity of the number of variables $n$.
  2. Whether the quantity $(-1)^{n/2} \det(f)$ is a square in $\mathbb{F}_q^\times$ (for even $n$).
  3. The parity of the degree $P$ (in specific cases).

3. Key Contributions

• Exact Formulas: The paper provides closed-form exact formulas for the number of rational points/morphisms, rather than just asymptotic expansions. This is a significant refinement over previous results (e.g., by Lee) which focused on leading terms.
• Secondary Terms: The author explicitly calculates secondary terms in the counting formulas. In the case where $n=4$ and the determinant is a square, a term of order $P q^{2P}$ appears, where the factor $P$ plays the role of a logarithmic factor found in analogous number field problems.
• Geometric Interpretation of Parity: The results reveal a deep connection between the arithmetic of the quadratic form and the geometry of the variety:
  • If $n$ is even and the determinant is a non-square, the variety is a non-split quadric. The paper proves that no morphisms of odd degree exist in this case (the count is 0), reflecting the fact that the Picard group has rank 1.
  • If $n$ is even and the determinant is a square, the variety is split (isomorphic to $\mathbb{P}^1 \times \mathbb{P}^1$), and morphisms exist for all degrees.
• Cohomological Insight: The author suggests that the coefficients in these exact formulas can be interpreted via the Grothendieck–Lefschetz trace formula as traces of the Frobenius morphism acting on the $\ell$-adic étale cohomology of the moduli space of morphisms.

4. Main Results
The paper derives distinct formulas based on three cases:

• Case 1: $n$ is even, and $(-1)^{n/2}\det(f)$ is a square in $\mathbb{F}_q^\times$.

  • For $n=4$, the count includes a main term proportional to $P q^{2P}$ and a secondary term proportional to $q^{2P}$.
  • For $n \ge 6$, the count is a sum of two terms: a dominant term of order $q^{P(n-2)}$ and a secondary term of order $q^{P(n/2)}$.

• Case 2: $n$ is even, and $(-1)^{n/2}\det(f)$ is NOT a square in $\mathbb{F}_q^\times$.

  • For $n=4$, the count is non-zero only if $P$ is even. If $P$ is odd, the count is exactly 0.
  • For $n \ge 6$, the formula depends on the parity of $P$. The secondary term involves a factor of $(-1)^P$, oscillating based on the degree.

• Case 3: $n$ is odd.

  • Similar to Case 2, the existence of morphisms depends on the parity of $P$. For $n=3$, no morphisms exist for odd $P$.
  • For $n \ge 5$, the formulas involve terms of order $q^{P(n-2)}$ and $q^{P(n-1)/2}$, with the secondary term vanishing or appearing based on whether $P$ is even or odd.

5. Significance

• Manin's Conjecture: The results provide concrete evidence for Manin's conjecture (regarding the distribution of rational points on Fano varieties) in the function field setting. The main terms in the derived formulas align with the predicted asymptotic behavior.
• Function Field vs. Number Field: The paper highlights the unique advantages of the function field setting, where the non-archimedean nature of the absolute value and the discrete valuation ring structure allow for the derivation of exact formulas, a feat often difficult to achieve over number fields.
• Moduli Spaces: By counting morphisms $\mathbb{P}^1 \to X$, the work contributes to the understanding of the geometry and arithmetic of moduli spaces of stable maps (Kontsevich compactifications) over finite fields.

In summary, Johanna Mettasch's work successfully applies the circle method to quadrics over function fields to produce precise, exact counting formulas that reveal intricate dependencies on the arithmetic of the underlying field and the geometry of the quadric hypersurface.