Bitwist manifolds and two-bridge knots (Q305129)

The authors provide a description as identification spaces of the \(3\)-ball for the cyclic branched covers of \(2\)-bridge knots, where the identification is compatible with the action of the group of deck transformations of the cover. Recall that an identification space is a quotient space obtained by subdividing the boundary of the ball into \(2\)-cells and gluing these together in pairs. The main idea behind the construction is as follows: Describe the \(3\)-sphere as an identification space of the \(3\)-ball of a special type, called \textsl{bitwisted face paring}. Bitwisted face parings have the nice property that the associated identification spaces are manifolds and not just pseudomanifolds as is generically the case with arbitrary face parings. Assume that one can choose the face paring so that a North-South axis of the ball (joining two vertices of the cellular decomposition) is mapped to a \(2\)-bridge knot in the identification space. It is now easy to describe the \(n\)-fold cyclic branched cover of the given \(2\)-bridge knot as an identification space, just by taking the \(n\)-fold cyclic cover of the \(3\)-ball branched on the North-South axis and lifting the face paring. The main point is thus to provide, for each \(2\)-bridge knot \(K\), a bitwisted face paring such that the quotient of the pair (\(3\)-ball, North-South axis) is \((\mathbf S^3,K)\). It turns out that the structure of the face paring reflects the continued fraction expansion of the rational invariant of the \(2\)-bridge knot. As a by-product of the main result, it is shown that the fundamental groups of the cyclic branched covers of \(2\)-bridge knots admit cyclic presentations. Some considerations on these groups and their abelianizations are also provided.