Periodic knots, Smith theory, and Murasugi's congruence (Q1179586)

A knot \(K\) in a homology 3-sphere \(\Sigma\) is said to have period \(n\) if \(K\) is invariant under an autohomeomorphism \(h\) of \(\Sigma\) of order exactly \(n\) with fixed set \(B\), a circle disjoint from \(K\). The reviewer proved a congruence relation between the Alexander polynomial \(\Delta_ K(t)\) of \(K\) and that of the quotient knot \(\tilde K\) in the quotient space \(\tilde\Sigma=\Sigma/h\) that is also a homology 3-sphere when \(n\) is \(p^ r\) for a prime \(p\). The authors prove an analogous congruence for a \(G\)-periodic \((q-2)\)-knot \(K\) and its quotient knot \(\tilde K\) in a homology \(q\)-sphere \(\Sigma\). (\(K\) is \(G\)-periodic if \(K\) is invariant under a semi-free \(G\)-action on \(\Sigma\) with fixed set \(B\) \((\cong S')\) disjoint from \(K\). An interesting but simple consequence of the theorem is Proposition C: If the \(\mathbb{Z}/2 \times \mathbb{Z}/2\) branched cover of a 2 component link \(L\) in a homology 3-sphere is again a homology 3-sphere, then the linking number of two component is congruent to \(\pm 1\pmod 8\).