No label defined (Q4795402)

The Riemann-Siegel formula [see e.g., \textit{E. C. Titchmarsh}, The theory of the Riemann zeta-function, Clarendon Press, Oxford (1986; Zbl 0601.10026)] plays a major rôle in the theory of the Riemann zeta-function \(\zeta(s)\). It gives an asymptotic expansion of the error term in the approximate functional equation for \(\zeta(s)\) (\(s=\sigma+it\)), namely \[ \zeta(s)=\sum_{n\leq\sqrt{t/2\pi}}n^{-s} -\chi(s)\sum_{n\leq\sqrt{t/2\pi}}n^{s-1},\quad \chi(s)=2^s\pi^{s-1}\sin(\pi s/2)\Gamma(1-s), \] but it is not obvious how to generalize the method of proof either to \(\zeta^k(s)\) or to other Dirichlet series that share some of the basic properties of \(\zeta(s)\). In 1983 \textit{Y. Motohashi} obtained an expansion of the analogue of the formula above for \(\zeta^2(s)\) (the complete and detailed proof was published in [Sieve methods, exponential sums, and their applications in number theory (ed. by G. R. H. Greaves et al.), Cambridge University Press (1997; Zbl 0910.00038), 293-308].) Another type of formula for \(\zeta^2(s)\), based on a different method, was found by the author [Acta Arith. 82, 309-330 (1997; Zbl 0887.11035)]. His method was used by the reviewer to derive a Riemann-Siegel formula for \(\zeta^k(s)\) (\(k>2\) an integer) [Publ. Math. 52, 469-495 (1998; Zbl 0906.11045)]. The author generalized the Riemann-Siegel formula to the Dirichlet series associated with cusp forms in [Analytic and elementary number theory (eds. W. G. Nowak and J. Schoissengeier), Universität Wien, 53-69 (1996; Zbl 0881.11047)] and [Arch. Math. 69, 391-402 (1997; Zbl 0899.11045)]. The present paper successfully continues his earlier work on the subject, and provides a precise Riemann-Siegel formula for \(\varphi(s)=\sum_{n=1}^\infty a(n)n^{-s}\). Here \(a(n)\) is the \(n\)-th Fourier coefficient of a cusp form of weight \(k\) for the full modular group, which is an eigenfunction of the Hecke algebra. The author first states two lemmas, containing mostly asymptotic expansions of the functions which occur later. Then he studies the so-called ``incomplete cusp form'', namely the function \[ \psi_1(\eta,x):=\sum_{n>\eta}a(n)e^{-2\pi nx} \qquad(\eta \geq 1,\;\Re x > 0), \] that plays an important rôle in the derivation of the Riemann--Siegel formula for \(\varphi(s)\). This is given as Theorem 4 in Section 3, and the precise formulation of this important result is too lengthy to be stated here. Some consequences are given in Section 4, and we mention the following one: if the \(a(n)\)'s are real, \(\eta=t/(2\pi)\), \(t\geq 2\pi\), then \[ \begin{aligned} \varphi(k/2+it)&=\sum_{n\leq\eta}a(n)n^{-k/2-it} +(-1)^{k/2}X(k/2+it)\sum_{n\leq\eta}a(n)n^{-k/2+it}\\ &-2A(\eta)\eta^{-k/2}\cos(t-t\log\eta)+O(t^{-1/4}),\\ &X(s)=(2\pi)^{2s-k}\frac{\Gamma(k-s)}{\Gamma(s)},\;A(x)=\sum_{n\leq x}a(n). \end{aligned} \] This formula shows that the error term in the approximate functional equation for \(\varphi(k/2+it)\) is essentially determined by the behaviour of \(A(x)\).