Generalization to quotients of arbitrary dimension

If $ R$ is a subring of $ {\bf {C}}$ , let $ S_2(R)=S_2(\Gamma;R)$ denote the $ {\bf {T}}$ -submodule of  $ S_2(\Gamma;{\bf {C}})$ of modular forms whose Fourier expansions have all coefficients in $ R$ .

Lemma 3.1   The Hecke operators leave $ S_2(R)$ stable.

Proof. If $ \Gamma = \Gamma_0(N)$ , then by the explicit description of the Hecke operators on Fourier expansions (e.g., see [DI95, Prop. 3.4.3]), it is clear that the Hecke operators leave $ S_2(R)$ stable. For a general $ \Gamma$ , by [DI95, (12.4.1)], one just has to check in addition that the diamond operators also leave $ S_2(R)$ stable, which in turn follows from [DI95, Prop. 12.3.11]. $ \qedsymbol$

Lemma 3.2   We have $ S_2(R)\cong S_2({\bf {Z}})\otimes R$ .

Proof. This is [DI95, Thm. 12.3.2] when our spaces $ S_2(R)$ and $ S_2({\bf {Z}})$ are replaced by their algebraic analogues (see [DI95, pg. 111]). Our spaces and their algebraic analogues are identified by the natural $ q$ -expansion maps according to [DI95, Thm. 12.3.7]. $ \qedsymbol$

If $ B$ is an abelian variety over $ {\bf {Q}}$ and $ S$ is a Dedekind domain with field of fractions $ {\bf {Q}}$ , then we denote by $ B_S$ the Néron model of $ B$ over $ S$ ; also, for ease of notation, we will abbreviate % latex2html id marker 7810
$ H^0(B_S, \Omega^1_{B_S/S})$ by  % latex2html id marker 7812
$ H^0(B_S, \Omega^1_{B/S})$ .

The inclusion $ X \hookrightarrow J$ that sends the cusp $ \infty$ to 0 induces an isomorphism

% latex2html id marker 7819
$\displaystyle H^0(X,\Omega^1_{X/{{\bf {Q}}}}) \cong H^0(J,\Omega^1_{J/{{\bf {Q}}}}).$

Let $ \phi_2$ be the optimal quotient map $ J \rightarrow A$ . Then $ \phi_2^*$ induces an inclusion % latex2html id marker 7827
$ \psi: H^0(A_{{\bf {Z}}},\Omega^1_{A/{{\bf {Z}}}}) \hookrightarrow H^0(J,\Omega^1_{J/{{\bf {Q}}}})[I] \cong
S_2({\bf {Q}})[I]$ , and we have the following commutative diagram:

% latex2html id marker 7829
$\displaystyle \xymatrix{
{H^0(A,\Omega^1_{A/{{\bf{...
...}}})}\ar@{^(->}[u]\ar@{^(->}[urr]_{\psi} & & {S_2({\bf{Z}})[I]}\ar@{^(->}[u]
}
$

Definition 3.3   The Manin constant of $ A$ is the (lattice) index

% latex2html id marker 7838
$\displaystyle {c_{\scriptscriptstyle{A}}}= [S_2({\bf {Z}})[I]: \psi(H^0(A_{{\bf {Z}}},\Omega^1_{A/{\bf {Z}}}))].
$

Theorem 3.4 below asserts that $ {c_{\scriptscriptstyle{A}}}\in{\bf {Z}}$ , so we may also consider the Manin module of $ A$ , which is the quotient % latex2html id marker 7844
$ M = S_2({\bf {Z}})[I] / \psi(H^0(A_{{\bf {Z}}},\Omega^1_{A/{\bf {Z}}}))$ , and the Manin ideal of $ A$ , which is the annihilator of $ M$ in $ {\bf {T}}$ .

If $ A$ is an elliptic curve, then $ {c_{\scriptscriptstyle{A}}}$ is the usual Manin constant. The constant $ c$ as defined above was also considered by Gross [Gro82, 2.5, p.222] and Lang [Lan91, III.5, p.95]. The constant  $ {c_{\scriptscriptstyle{A}}}$ was defined for the winding quotient in [Aga99], where it was called the generalized Manin constant. A Manin constant is defined in the context of $ {\bf {Q}}$ -curves in [GL01].

William Stein 2006-06-25