Modular forms, cusps, and transformation laws

A weight-k modular form is holomorphic on the upper half-plane and satisfies f(γτ)=(cτ+d)^k f(τ). The factor is generally nonconstant, so positive-weight forms are not ordinary invariant functions on the quotient. The slash notation f|ₖγ=(cτ+d)⁻ᵏf(γτ) packages the requirement as f|ₖγ=f. The automorphy factor satisfies a cocycle identity, making slash composition consistent with matrix multiplication. This is why checking compatible laws for generators can establish the law for the whole group.

Applying the law to −I gives f=(−1)^k f. For full level and trivial character, every odd-weight form is therefore zero. Checking only the maps induced on the upper half-plane could miss this obstruction because −I fixes every point. For even weight, the conditions f(τ+1)=f(τ) and f(−1/τ)=τ^k f(τ) express the two generator laws. A multiplier system or character changes these equations and must be included in the definition before discussing odd or fractional weights.

Unit periodicity lets f descend to a holomorphic function F on the punctured unit disk through q=exp(2πiτ). Holomorphy at infinity means that F extends holomorphically to q=0. Equivalently, its Laurent expansion has no negative powers, and it has a convergent power series with coefficients aₙ for n≥0. Boundedness near the puncture is another equivalent condition by the removable-singularity theorem. Periodicity alone supplies a Laurent expansion, not the nonnegative-index conclusion.

A cusp form has constant term zero in addition. At full level there is one cusp class, so checking infinity suffices. For a finite-index subgroup one must transport every cusp to infinity and inspect the corresponding slash-transformed function. If the effective cusp width is h and no character or odd-weight sign twist is present, the local variable is exp(2πiτ/h). The width is a property of that subgroup at that cusp; using the full-level variable indiscriminately can disguise fractional powers and incorrect vanishing orders.

Multiplication adds weights and preserves holomorphy, so forms of all nonnegative weights make a graded ring. Dividing two forms of equal weight cancels their automorphy factors and produces a modular function, but zeros of the denominator can introduce poles. A quotient of differing weights has the weight difference wherever it is defined and must still be checked for poles. The phrase weight zero therefore does not make a meromorphic quotient into a holomorphic modular form.

A holomorphic weight-zero form for full level extends to a holomorphic function on the compactified quotient, hence is constant. By contrast, the invariant j-function has a pole at the cusp and is a modular function, not an element of the holomorphic space M₀. Conversely, a q-series with no negative powers is only a candidate: most such series fail inversion symmetry. Definition checking must keep these two directions separate. Fourier data express cusp regularity, while the modular transformation law supplies the global relation between different parts of the half-plane.