A cusp form is bounded on the vertical segment above a rational cusp
ProvedMTT.cuspForm_bounded_near_rationalA cusp form of level is bounded on the vertical segment running from a rational point up to height one.
Precisely: for and there is a constant with
This is the statement that does not blow up as the vertical ray descends to the rational point , which is a cusp of . It is the delicate half of the assertion that the Mazur–Tate–Teitelbaum modular integral
converges: convergence at the top of the ray is the exponential decay of at the cusp , which is immediate from the -expansion, while convergence at the bottom is exactly this statement.
Why boundedness is not automatic. The Petersson bound , valid for any cusp form, gives only , which is unbounded for and not integrable near for . The bound stated here therefore genuinely uses the vanishing of at the cusp , not merely its modularity. Writing with , one has , so that
and the imaginary part on the right tends to as . Since vanishes at — and, being invariant under , which is normal in , is periodic of period and so decays like — the exponential factor overwhelms the pole . In fact as ; only boundedness is asserted here, since that is what the integrability argument consumes.
Formalization note. UpperHalfPlane.ofComplex is the section of the inclusion that takes an arbitrary junk value off the upper half-plane; for the point lies in , so ofComplex returns it and the statement says what it appears to say.
import Definitions.Def_MTT_Cohomology import Mathlib.RingTheory.Flat.Basic set_option autoImplicit false noncomputable section open scoped BigOperators TensorProduct open MTT.Cohomology
theorem MTT.cuspForm_bounded_near_rational
{N k : ℕ} (hN : 0 < N) (f : CuspForm (MTT.GammaOne N) (k : ℤ)) (r : ℚ) :
∃ C : ℝ, ∀ t : ℝ, 0 < t → t ≤ 1 →
‖f (UpperHalfPlane.ofComplex ((r : ℂ) + Complex.I * t))‖ ≤ C := by sorry