Invariants of an augmentation of \(\Gamma_{0}(n)\) (Q1037570)

The genesis of this article can be found in Fricke's early-twentieth-century study of modular curves defined by the Hecke groups \[ \Gamma^0(n):= \left\{\begin{pmatrix} a & b\\ c & d\end{pmatrix}\in \mathrm{SL}(2,\mathbb Z): b\equiv 0\pmod n,\, n\in\mathbb Z^+\right\}. \] Seeking simplification through reduction of the genus of these curves, \textit{R. Fricke} achieved this by augmenting \(\Gamma^0(n)\) through extension by involution arising from from the 2-group of primary factors of \(n\): \[ G^\times_n= \left\{t: t\mid n,\,(\tfrac nt, t)= 1\right\}, \] with the binary operator \(t\times u= tu/(t,u)^2\). (Here, \((a,b)\) denotes the greatest common divisor of \(n\).) [Lehrbuch der Algebra. III. Vieweg, Braunschweig (1928; JFM 54.0187.20).] As the authors' introduction points out, in later work ``Fricke extended the technique of augmenting \(\Gamma^0(n)\) by\dots involutions using a method he attributed to Bessel-Hagen, constructing `partial invariants' of the augmented group, based on \(\Delta\)-functions,'' or, equivalently, the Dedekind \(\eta\)-function. In several papers of the period 1989--1991 Harvey Cohn studied these partial invariants (generalizations of modular forms of weight \(0\) with multiplier systems) on the augmentation of \(\Gamma^0(n)\), in particular, their associated multipliers. In [Contemp. Math. 166, 9--34 (1994; Zbl 0813.11025)] \textit{H. Cohn} and the reviewer determined these multipliers almost fully, omitting a few irksome cases. This work was completed by the authors of the paper under review, who calculated the multipliers in the cases omitted [Ramanujan J. 8, No. 1, 47--56 (2004; Zbl 1065.11022)]. Because the partial invariants are formed from products of functions closely related to \(\eta(z)\), the multipliers are a priori 24th roots of unity. Surprisingly, they actually turn out to have the values \(\pm 1\). In the present article the authors derive results analogous to those found in the two papers cited in the previous paragraph, for the case in which the group augmented is \(\Gamma_0(n)\) rather than \(\Gamma^0(n)\). One again the partial invariants are derived from \(\eta\)-products and again the multipliers, which the paper determines completely, have the values \(\pm 1/\). This work, as well as that done earlier on \(\Gamma^0(n)\), can be reinterpreted from the viewpoint of vector-valued modular forms and the recently-emerging theory of generalized modular forms [J. Number Theory 99, No. 1, 1--28 (2003; Zbl 1074.11025)]. Whether this altered perspective on partial invariants can yield significant new information remains to be seen.