Pencils of symmetric surfaces in \(\mathbb{P}_3\). (Q5956273)

In the first paper under review, the author considers the case of the groups \(G=TT,OO,II\) which are the images in \(\text{SO}(4)\) of the direct products \(\widetilde T\times\widetilde T\), \(\widetilde O\times\widetilde O\), \(\widetilde I\times \widetilde I\), where \(\widetilde T\) denotes the binary tetrahedral group, \(\widetilde O\) the binary octahedral group, \(\widetilde I\) the binary icosahedral group in \(\text{SU}(2)\). The author denotes the groups by \(G_6\), \(G_8\) and \(G_{12}\) and describes the corresponding \(G_n\)-invariant pencils of surfaces in \(\mathbb{P}_3(\mathbb{C})\). In the second paper under review [Commun. Algebra 32, 3745--3770 (2004; Zbl 1064.14038)], the author describes the other \(G\)-invariant pencils of surfaces and shows that there are five more pencils and two of these contain surfaces with nodes. To be more precise, let's consider the Klein four-group \(V\subseteq \text{SO}(3)\). Let \(\widetilde V\) denote its inverse image in \(\text{SU}(2)\) under the universal covering \(\text{SU}(2) \to \text{SO}(3)\). The image of the direct product \(\widetilde V\times\widetilde V\) in \(\text{SO}(4)\) under the double covering \(\text{SU}(2)\times \text{SU}(2)\to \text{SO}(4)\) is the Heisenberg group \(H\). The author classifies all the subgroups \(G\) of \(\text{SO}(4)\) containing \(H\). Moreover, she gives generators for the spaces \(\mathbb{C}[x_0,x_1,x_2,x_3]^G_j\) of homogeneous \(G\)-invariant polynomials of degree \(j\) whenever the dimension of the spaces is two. In this case, the generators are the multiple quadric \(q^{j/2}=(x_0^2+ x^2_1 +x^2_2+x^2_3)^{j/2}\) (trivial invariant) and another polynomial of degree \(j\) that we denote by \(f\). The author describes the pencils \[ f+\lambda q^{j/2}=0, \quad\lambda\in\mathbb{P}_1, \] of surfaces in the three-dimensional complex projective space \(\mathbb{P}_3\) and, in particular, she finds the singular surfaces they contain. The study and classification that the author completes in the two papers are important for the following reasons. 1. Mukai identified the quotient space \(\mathbb{P}_3/ G_n\) with the Satake compactification \((n=6,8)\) [resp. with a certain modification of it \((n=12)]\) of the moduli spaces of abelian surfaces admitting a certain polarization. The author's surfaces descent to these quotients and should therefore be related to modular forms. 2. The reflection group of the regular four-dimensional 24-cell [resp. 600-cell] contains the group \(G_6\) [resp. \(G_{12}]\). The invariant of \(G_6\) [resp. \(G_{12}]\) are actually invariants under this larger group. The existence of these invariants for the reflection group is known, but no explicit equations seem to have been computed before. 3. At Europroj '96, V. Goryunov observed that \(G_{12}\) admits an invariant surface of degree 12 with 600 nodes. The author's explicit computations confirm his announcement. This gives the lower bound \(\mu(12)\geq 600\), where \(\mu(d)\) denotes the maximum number of nodes on a projective surface of degree \(d\). The best upper bound for \(\mu(12)\) is Miyaoka's \(\mu (12)\leq 645\). From a result obtained by Kreiss and later by Chmutov, it was known that \(\mu(12)\geq 576\). The above result explicitly improves this bound to show \(600\leq\mu(12)\leq 645\).