The Dirichlet series of Koecher and Maaß and modular forms of weight 3/2 (Q1188022)

The basis problem for modular forms is the question which modular forms can be expressed as linear combinations of theta series. A special form of this problem is which cuspidal modular forms \(F\) of degree \(n\) and weight \(k\) for the group \(\Gamma_0^{(n)}\) can be expressed as linear combinations of the theta series of degree \(n\) of integral quadratic forms of rank \(m=2k\) and the same level \(N\). In this paper the authors give a proof of the result for \(n=1\), weight \(3/2\) and square free level. However, the result itself has already been proved by \textit{B. Gross} [Heights and the special values of \(L\)-series, Number Theory, Montreal 1985, CMS Conf. Proc. 7, 115--187 (1987; Zbl 0623.10019)]. The method of proof uses results from their previous work on Waldspurger's formula that the Koecher-Maaß Dirichlet series \(D_{K,M}(F,s)\) of certain linear combinations \(F\) of theta series of degree 2 and weight 2 is connected with the Rankin convolution of Dirichlet series attached to two modular forms \(g_1,g_2\) of weight 3/2 [the authors, Math. Ann. 288, 361--388 (1990; Zbl 0688.10025)]. By specializing to \(g_1=g_2\), they establish their solution to the basis problem with the help of a formula for the residue of \(D_{K,M}(F,s)\) at \(s=3/2\) (due to \textit{H. Maaß} [Math. Ann. 122, 90--108 (1950; Zbl 0038.24001)] and \textit{T. Arakawa} [Tôhoku Math. J., II. Ser. 42, 261--286 (1990; Zbl 0714.11031)]). They stress that the present method does not use Waldspurger's nonvanishing result for \(L(f,s)\).