Real elements and real-valued characters of covering groups of elementary Abelian 2-groups. (Q1881116)

Let \(C_2^{(n)}\) denote an elementary Abelian 2-group of order \(2^n\), where \(n\geq 2\) is a positive integer. The Schur multiplier of \(C_2^{(n)}\) is an elementary Abelian group of order \(2^{n(n-1)/2}\) and thus a (Schur) covering group \(G\) of \(C_2^{(n)}\) has order \(2^{n(n+1)/2}\). We recall that in \(G\), \(G'=Z(G)\), where \(G'\) denotes the commutator subgroup of \(G\) and \(Z(G)\) the centre of \(G\). Moreover, \(Z(G)\) is elementary Abelian of order \(2^{n(n-1)/2}\) and \(G/Z(G)\) is isomorphic to \(C_2^{(n)}\). As \(n\) increases, the number of non-isomorphic covering groups increases rapidly. For example, the number of non-isomorphic covering groups is two when \(n=2\) and ten when \(n=3\). All covering groups share certain properties with respect to their complex characters, such as the number of irreducible characters of a given degree. However, they do not all have the same number of real-valued irreducible characters and the Frobenius-Schur indicators of the characters depend on aspects of the group structure which vary from group to group. The author asks which covering groups \(G\) have the largest number of real conjugacy classes. We first note that if a coset of \(Z(G)\) contains a real element, then all the elements in this coset are real. Moreover, each coset of \(Z(G)\) different from \(Z(G)\) is a disjoint union of conjugacy classes each of size \(2^{n-1}\). Thus since \(Z(G)\) consists of real central elements, the covering groups \(G\) with the largest number of real conjugacy classes (and hence the largest number of real-valued irreducible characters) are those for which the number of cosets of \(Z(G)\) consisting of real elements is as large as possible. The author proves that for all \(G\), the number of cosets of \(Z(G)\) consisting of real elements is at most \({n+1\choose 2}+1\) and she calls \(G\) `maximally real' if \(Z(G)\) achieves this number of cosets consisting of real elements. She then proves the following generating theorem. Let \(G\) be a maximally real covering group of \(C_2^{(n)}\). Then \(G\) is generated by \(n\) elements \(x_1,\dots,x_n\) satisfying \(x_1^2=\cdots=x_n^2=r\), and either \(r=1\) or \[ r=\prod_{1\leq i<j\leq k}[x_i,x_j], \] where \(k\) is even and \(2\leq k\leq n\). Here, \([a,b]\) denotes the commutator of elements \(a\) and \(b\). As a consequence, she deduces that the number of non-isomorphic maximally real covering groups of \(C_2^{(n)}\) equals \(1+\lfloor n/2\rfloor\). When \(n=2\), the two non-isomorphic covering groups (the dihedral and quaternion groups of order 8) have all their characters real and the author proves that, for all \(n\geq 2\), these are the only covering groups with this property. The author finishes the paper by considering how to calculate the Frobenius-Schur indicator of each irreducible character of a given covering group \(G\). She considers in detail the maximally real covering group \(G\) defined by setting \(r=1\) in the description above. This covering group, which is generated by \(n\) involutions, is unique in having the largest number of involutions of any covering group. She obtains three recurrence relations which relate the number of irreducible characters of \(G\) with Frobenius-Schur indicator equal to 1, \(-1\) or \(0\), respectively, in terms of corresponding numbers defined for the covering group defined by \(n-1\) involutory generators. The enumeration utilizes some ingenious graph-theoretic ideas.