Signature defects of Hilbert-Picard modular cusps (Q1319322)

A Hilbert-Picard modular cusp is a boundary point to the quotient space of the product of the complex unit balls, while a Tsuchihashi cups is a boundary point to the quotient of a tube domain [cf. \textit{H. Tsuchihashi}, Tôhoku Math. J., II. Ser. 35, 607-639 (1983; Zbl 0585.14004)]. Let \(F\) be a totally real algebraic number field of degree \(n(\geq 2)\) and \(K\) an imaginary quadratic extension of \(F\) with the ring \({\mathfrak O}_ K\). For a positive integer \(m\) the unitary group \(U(m+1,1;{\mathfrak O}_ K)\) with its entries in \({\mathfrak O}_ K\) acts on the product \((\mathbb{B}_{m+1})^ n\) of \(n\) copies of the complex unit ball in \(\mathbb{C}^{m + 1}\) through the \(n\) different embeddings of \(K\) into \(\mathbb{C}\). The quotient space is compactified by adding a finite number of points, which are called Hilbert-Picard modular cusps. In this paper we prove an equality between two invariants of a Hilbert-Picard modular cusp singularity \((V,P)\) of even dimension, namely, the signature defect \(\sigma(p)\) and the contribution of the cusp \(\chi_ \infty(p)\) to the arithmetic genus of the cusp singularity. Theorem. Let \((V,p)\) be a Hilbert-Picard modular cusp singularity of dimension \(n(m+1)\). When \(n(m+1)\) is even we have \(2^{n(m+1)}\chi_ \infty(p) = \sigma(p)\). The signature defect was defined by \textit{F. E. P. Hirzebruch} [Enseign. Math., II. Sér. 19, 183-281 (1973; Zbl 0285.14007)] in the case of Hilbert modular cusps and generalized by Satake. The invariant \(\chi_ \infty(p)\) of a Hilbert modular cusp is called the \(\psi\)-invariant by \textit{F. Ehlers} [Math. Ann. 218, 127-156 (1975; Zbl 0301.14003)]. Satake [loc. cit.] defined \(\chi_ \infty(p)\) in more general context, showing that it coincides with the contribution of the cusp to the dimension formula of the space of cusp forms as calculated by means of the Riemann- Roch theorem.