Remarks on the zig-zag theorem (Q5932656)

\textit{O. Bottema} [Elem. Math. 20, 1-7 (1965)] has shown that if, for two given circles \(C_1\) and \(C_2\) in the Euclidean plane over the reals ((*) for which any circle of radius \(\rho\) centred on either circle intersects the other in exactly two points), a zig-zag closes in \(m\) steps for a certain side-length \(\rho\), i.e. if an equilateral closed \(2m\)-gon which shuttles between \(C_1\) and \(C_2\) exists with side length \(\rho\), then any two points on \(C_1\) and \(C_2\) at distance \(\rho\) are part of a zig-zag which closes in \(m\) steps. \textit{W. L. Black, H. C. Howland} and \textit{B. Howland} [Am. Math. Mon. 81, 754-757 (1974; Zbl 0291.50008)] proved a 3-dimensional generalization of the zig-zag theorem. This paper contains a generalization (in which only the local deformability of a zig-zag is proved, in the sense that there is a neighborhood around a vertex of an \(m\)-step zig-zag which contains vertices of \(m\)-step zig-zags) to \(n\)-dimensional (for all \(n\geq 2\)) Euclidean, hyperbolic, and spherical geometry over the reals. \(C_1\) and \(C_2\) are no longer required to be circles, but uniform curves, and need not satisfy (*). Some duality theorems are also proved, e.g. it is shown that the distance between the centres of the circles and the length \(\rho\) can be interchanged to yield closed zig-zags with the same number of steps.