On the skew-morphisms of dihedral groups (Q338552)

A \textit{skew morphism} of a finite group \(A\) is a bijection \(\varphi\!:A\to A\) such that \(\varphi\) fixes the identity element of \(A\), and there exists an auxiliary function \(\pi\!:A\to\mathbb Z_s\) such that \(\varphi(xy)=\varphi(x)\varphi^{\pi(x)}(y)\) for all \(x,y\in A\), where \(s\) is the order of \(\varphi\) (as a permutation). Here, \(\pi\) is called the `power function' of \(\varphi\), and the subgroup \(K=\{x \in A\mid\pi(x)=1\}\) is called the `kernel' of \(\varphi\). This notion arose from the study of regular Cayley maps, but also has connections to group factorisations of the form \(G=AB\) where \(A\) and \(B\) are subgroups of \(G\) with \(B\) cyclic. Further background and properties of skew morphisms may be found in [\textit{M. D. E. Conder} et al., J. Algebra 453, 68--100 (2016; Zbl 1338.20019)]. This paper gives a somewhat surprising answer to a question/conjecture raised at a conference by the reviewer about skew morphisms of dihedral groups. If \(a\) and \(b\) are generators for the dihedral group \(D_n\) satisfying \(a^n=b^2=(ab)^2=1\), then for \(1<n\leq 45\) the kernel of every skew morphism of \(D_n\) contains at least one element lying outside the cyclic subgroup generated by \(a\). Here, the authors prove that this does not happen for all \(n\). In fact, they prove that the kernel \(K=\ker\varphi\) lies in \(\langle a\rangle\) if and only if \(K\) has index \(4\) in \(D_n\) and \(\varphi\) preserves \(\langle a \rangle\), and then they give an infinite family of examples where these things hold: one with \(n = 8m\) for every odd integer \(m \geq 3\).