Knot complements, hidden symmetries and reflection orbifolds (Q318713)

A conjecture of Reid and Walsh states that any commensurability class of cusped hyperbolic \(3\)-manifolds contains at most three knot complements. The authors of the paper under review showed that this conjecture holds true for knots without \textit{hidden symmetries}, that is, knots whose exteriors are regular covers of the minimal orbifold in their commensurability class [the authors, Geom. Topol. 16, No. 2, 625--664 (2012; Zbl 1258.57001)]. The above conjecture is then a consequence of a conjecture by Neumann and Reid claiming that only three knots, the figure-eight knot and the two dodecahedral knots, have complements admitting hidden symmetries. Here the authors provide evidence supporting Neumann and Reid's conjecture. They consider a natural class of potential examples, that is knots that are commensurable with reflection orbifolds. A key observation they make is the following: A reflection cannot be a symmetry of a hyperbolic knot. As a consequence, a hyperbolic knot whose complement contains a reflection in its commensurator must have hidden symmetries. The main result of the paper states that Neumann and Reid's conjecture holds true for the class of knots whose complements cover a reflection orbifold that is \textit{small}, i.e. does not contain an essential, closed \(2\)-suborbifold. As a consequence, the authors prove that the class of AP knots verifies Reid and Walsh's conjecture. Recall that a knot is an \textit{AP knot} if any closed essential surface in its complement contains an accidental parabolic. Note that this class of knots comprises small knots, alternating knots, and Montesinos knots amongst others. Finally, the authors describe the possible singular locus of an orientable minimal orbifold in the commensurability class of a hyperbolic non arithmetic knot complement, under the assumption that it admits a reflection (in a disc).