Local Gekeler Ratios in Drinfeld Modules

• Local Gekeler ratios are arithmetic invariants that measure asymptotically normalized matrix counts with a fixed characteristic polynomial over finite local rings.
• They are computed by reducing the matrix enumeration problem to the structure of the ideal class monoid through explicit overorder classification and weak equivalence classes.
• The global ratio, derived as an infinite product of local factors, parallels classical mass formulas and provides key insights into the weighted sizes of Drinfeld module isogeny classes.

Local Gekeler ratios are arithmetic invariants associated with the distribution of matrices over finite local rings with prescribed characteristic polynomials, closely linked to isogeny classes of Drinfeld modules. For a given polynomial ring $A = \mathbb{F}_q[T]$, a nonzero prime ideal $\mathfrak{p} \subset A$, and a monic polynomial $f \in A[x]$ of degree $r$, the local Gekeler ratio $v_\mathfrak{p}(f)$ measures asymptotically (as the power $n \to \infty$) the normalized count of $r \times r$ matrices over $A/\mathfrak{p}^n$ with characteristic polynomial $f$ relative to the size of the special linear group $SL_r(A/\mathfrak{p}^n)$. Formally, the ratio is realized as the limit

$$v_\mathfrak{p}(f) := \lim_{n \to \infty} \frac{|\{M \in \mathrm{Mat}_r(A/\mathfrak{p}^n)\mid \mathrm{charpoly}(M)=f\}|}{|SL_r(A/\mathfrak{p}^n)| / |\mathfrak{p}|^{n(r-1)}}.$$

These local invariants yield, upon global multiplication, the weighted size of an isogeny class of Drinfeld modules, paralleling the classical mass formulas for abelian varieties over finite fields (Eggink, 19 Jan 2026).

1. Formal Definition and Existence of Local Gekeler Ratios

Let $A = \mathbb{F}_q[T]$, $\mathfrak{p} \subset A$ a nonzero prime, and $f(x) \in A[x]$ monic of degree $r$. For each integer $n \geq 1$,

• $$S_n(f;\mathfrak{p}) = \{ M \in \mathrm{Mat}_r(A/\mathfrak{p}^n) \mid \mathrm{charpoly}(M) = f \bmod \mathfrak{p}^n \}$$,
• $G_n = SL_r(A/\mathfrak{p}^n)$,
• $L_n = |G_n| / |\mathfrak{p}|^{n(r-1)}$.

The local Gekeler ratio at $\mathfrak{p}$ is defined as

$$v_\mathfrak{p}(f) = \lim_{n\to\infty} \frac{|S_n(f; \mathfrak{p})|}{L_n},$$

where $|\mathfrak{p}| = q^{\deg \mathfrak{p}}$. The existence and finiteness of this limit are established by analyzing the growth of $|S_n(f; \mathfrak{p})|$ as $n$ increases and normalizing appropriately by $L_n$ to account for volume in the matrix space. Proposition 2.1 in (Eggink, 19 Jan 2026) shows the ratio stabilizes for sufficiently large $n$, and hence the limit is both finite and positive.

2. Ideal Class Monoids and Weak Equivalence Classification

The calculation of $v_\mathfrak{p}(f)$ is reduced to a structure-theoretic problem in the arithmetic of the local order $R_\mathfrak{p} = A_\mathfrak{p}[x]/(f)$. The ideal class monoid $ICM(R_\mathfrak{p})$ comprises fractional $R_\mathfrak{p}$-ideals modulo principal $R_\mathfrak{p}$-ideals. Each similarity class of matrices with characteristic polynomial $f$ corresponds, via an explicit bijection, to the set of such weak equivalence classes.

For each overorder $S_\mathfrak{p} \supset R_\mathfrak{p}$, one computes the set of weak equivalence classes $W_{S_\mathfrak{p}}(R_\mathfrak{p})$. Lemma 1.2, Propositions 1.12, 1.13, and Corollary 1.7 in (Eggink, 19 Jan 2026) together establish that:

• The set of $\mathfrak{p}$-overorders $S_\mathfrak{p}$ is finite and effectively computable.
• Each weak class corresponds to a unique ideal under isomorphism.
• The monoid decomposes as $$ICM(R_\mathfrak{p}) = \bigsqcup_{S_\mathfrak{p}} W_{S_\mathfrak{p}}(R_\mathfrak{p})$$.

The reduction of the matrix-counting problem to enumeration in $ICM(R_\mathfrak{p})$ follows Proposition 2.3.

3. Computation Algorithm for $v_\mathfrak{p}(f)$

The main computational algorithm proceeds as follows:

1. Factorization: Factor $f$ in $A_\mathfrak{p}[x]$ to describe the complete local order $R_\mathfrak{p}$ as a product of local factors.
2. Overorder Computation: Determine all overorders $S_\mathfrak{p} \supset R_\mathfrak{p}$ using the index module $[S_\mathfrak{p}:R_\mathfrak{p}]$.
3. Weak Class Enumeration: For each $S_\mathfrak{p}$, classify its weak equivalence classes $W_{S_\mathfrak{p}}(R_\mathfrak{p})$, utilizing the criteria in Lemma 1.16 to distinguish between classes.
4. Ideal Representatives and Unit Groups: For each class, select a representative ideal $I$ and compute its automorphism group $$Aut_{R_\mathfrak{p}}(I) = \{ x \in Frac(R_\mathfrak{p})^\times \mid xI = I \}$$.
5. Summation: Sum contributions from all classes to yield

$$v_\mathfrak{p}(f) = \sum_{[I] \in ICM(R_\mathfrak{p})} \frac{1}{|Aut_{R_\mathfrak{p}}(I)|}.$$

The limit's convergence is justified via volume computations and lattice-counting in $R_\mathfrak{p}^r$, as detailed in Theorem 2.7 of (Eggink, 19 Jan 2026).

4. Global Ratio, Isogeny Classes, and Mass Formula

The global Gekeler ratio is the infinite product

$$v(f) = \prod_{\mathfrak{p} \subset A} v_\mathfrak{p}(f),$$

where the product converges and almost all local factors are $1$ (specifically, for primes $\mathfrak{p}$ not dividing the discriminant of $f$). Gekeler's theorem, as presented in (Eggink, 19 Jan 2026), establishes that $v(f)$ equals the mass—i.e., the weighted cardinality—of the isogeny class of rank-$r$ Drinfeld $A$-modules with characteristic polynomial $f$. Precisely, if $|IsogClass(f)|_{weighted}$ denotes the weighted count in which each module $\varphi$ is counted with weight $1/|\text{Aut}(\varphi)|$, then

$|IsogClass(f)|_{weighted} = v(f).$

This structure is analogous to the mass formulas for abelian varieties over finite fields.

5. Summary of Notation and Key Formulas

Symbol	Meaning	Reference/Formulas
$A$	Ring $\mathbb{F}_q[T]$	—
$|\mathfrak{p}|$	$q^{\deg \mathfrak{p}}$	—
$R$	$A[x]/(f(x))$	—
$R_\mathfrak{p}$	Completion at prime $\mathfrak{p}$	$R \otimes_A A_\mathfrak{p}$
$S_\mathfrak{p}$	Overorders $R_\mathfrak{p} \subset S_\mathfrak{p} \subset O_K$	Indexed by $[S_\mathfrak{p}:R_\mathfrak{p}]=\mathfrak{p}^k$
$W_{S_\mathfrak{p}}(R_\mathfrak{p})$	Weak equivalence classes for given overorder	Definition 1.8
$ICM(R_\mathfrak{p})$	Ideal class monoid	$$\bigcup_{S_\mathfrak{p}} W_{S_\mathfrak{p}}(R_\mathfrak{p})$$
$v_\mathfrak{p}(f)$	Local factor	$$\sum_{[I] \in ICM(R_\mathfrak{p})} \frac{1}{|Aut_{R_\mathfrak{p}}(I)|}$$
$v(f)$	Global mass	$\prod_{\mathfrak{p}} v_\mathfrak{p}(f)$

6. Illustrative Example

For $q=3$, $A=\mathbb{F}_3[T]$, $r=2$, and $f(x)=x^2+1$, consider $\mathfrak{p}=(T)$. As $x^2+1$ is irreducible over $\mathbb{F}_3$, $R_\mathfrak{p} = A_\mathfrak{p}[x]/(x^2+1)$ forms a ramified quadratic extension and is already the maximal order, with trivial class group (Pic$(R_\mathfrak{p})=1$). Therefore, $ICM(R_\mathfrak{p})$ consists of a single principal class, represented by $R_\mathfrak{p}$ itself, whose automorphism group is $R_\mathfrak{p}^\times$. It follows that

$$v_\mathfrak{p}(f) = \frac{1}{|R_\mathfrak{p}^\times|} \cdot |R_\mathfrak{p}^\times| = 1.$$

In cases where $f$ is not irreducible, such as $f(x) = (x-T)^2 \bmod (T)$, $R_\mathfrak{p} \cong A_\mathfrak{p}[x]/(x^2)$, yielding two overorders and weak-equivalence classes, leading to $v_\mathfrak{p}(f) = 1 + 1/|A_\mathfrak{p}^\times|$ upon computation.

7. Connections and Significance

The framework of local Gekeler ratios provides a bridge between matrix enumeration over finite local rings and the ideal-theoretic structure of nonmaximal and maximal orders in local fields. The explicit algorithmic reduction to $ICM(R_\mathfrak{p})$ and overorder classification allows for efficient computation of isogeny class sizes for Drinfeld modules and mirrors classical strategies in the theory of abelian varieties. This suggests that the local-to-global principle via Gekeler ratios not only quantifies isogeny classes but also encodes rich structural information about orders and modules over function fields (Eggink, 19 Jan 2026).