Cyclic branched covers of alternating knots (Q2077175)

For an integer \(n \ge 2\), two knots in the 3-sphere are called \textit{\(n\)-twins} if their \(n\)-fold cyclic branched coverings are homeomorphic. The main result of the present paper then states that an \textit{alternating prime knot} has no \(n\)-twins, for \(n > 2\); in other words, an alternating prime knot is determined by its \(n\)-fold cyclic branched covering. The result is not true for composite knots, since ``mutation'' along a sphere of a decomposition into prime knots (intersecting the knot in two points) does not change the homeomorphism type of a cyclic branched covering; analogously, it is not true for \(n = 2\) since mutation along a Conway sphere (intersecting the knot in four points) does not change the 2-fold branched covering. Concerning the case \(n = 2\), it was shown by \textit{J. E. Greene} [Invent. Math. 192, No. 3, 717--750 (2013; Zbl 1278.57021)] that two prime alternating 2-twin knots are Conway mutants, i.e. obtained by mutation along Conway spheres; Greene also conjectured that if a knot \(K'\) is a 2-twin of a prime alternating knot \(K\), then \(K'\) is a Conway mutant of \(K\) and in particular also alternating. The proof of the main result of the present paper is based on two key ingredients. In a paper by \textit{B. Zimmermann} [Math. Ann. 311, No. 4, 665--673 (1998; Zbl 0913.57008)] it is shown how two hyperbolic knots (the basic case to be considered) with the same \(n\)-fold cyclic branched covering are related (see also a paper by \textit{M. Mecchia} [Topology Appl. 121, No. 3, 521--533 (2002; Zbl 0997.57004)]); in particular, each of the knots has a certain type of symmetry of period \(n\) (the covering transformation of the other knot). Now, for the case of an \textit{alternating knot}, such a symmetry is \textit{visible} on a minimal alternating diagram (projection) of the knot, and this finally allows the author to show that the two knots are equivalent.