Not all boundary slopes are strongly detected by the character variety (Q937823)

Let \(M\) be a compact, orientable, irreducible \(3\)-manifold whose boundary is a single torus. A slope \(\sigma\) is called a boundary slope of \(M\) if there is an essential surface in \(M\) whose boundary consists of a non-empty parallel family of simple closed curves in \(\partial M\) of slope \(\sigma\). Moreover, a boundary slope is strict if it is the boundary slope of an essential surface which is not a fiber or a semi-fiber. \textit{M. Culler} and \textit{P. B. Shalen} [Ann. Math. (2) 117, 109--146 (1983; Zbl 0529.57005)] established a method of constructing essential surfaces in a \(3\)-manifold \(M\) using representations of the fundamental group \(\pi_1(M)\) into \(\text{SL}_2(\mathbb{C})\). More precisely, let \(X\) be an irreducible affine curve in the character variety of \(M\), and let \(\widetilde{X}\) be its smooth projection model. All but finitely many points of \(\widetilde{X}\) correspond to characters in \(X\). The points which do not correspond to characters are called ideal points. To any ideal point \(\tilde{x}\in \widetilde{X}\), an essential surface in \(M\) is associated. For a slope \(\sigma\), a rational function \(I_\sigma:\widetilde{X}\to \mathbb{C}\cup\{\infty\}\) is induced by a regular function \(I_\sigma:X\to \mathbb{C}\) with \(I_\sigma(\chi)=\chi(s)\), where \(s\) represents the slope \(\sigma\). Then either there is a unique slope \(\sigma\) such that \(I_\sigma(\tilde{x})\in \mathbb{C}\) or \(I_\sigma(\tilde{x})\in \mathbb{C}\) for every slope \(\sigma\). If an ideal point \(\tilde{x}\) is of the former type and \(I_\sigma(\tilde{x})\in \mathbb{C}\), then every essential surface associated to \(\tilde{x}\) has boundary slope \(\sigma\), and \(\sigma\) is said to be strongly detected by \(\tilde{x}\). For the latter case, a closed essential surface is associated to the ideal point by the construction. It is known that if \(M\) is small, then every ideal point is of the first type. Generally, strongly detected boundary slopes are common. The first examples of strict boundary slopes which are not strongly detected were found by \textit{S. Schanuel} and \textit{X. Zhang} [Math. Ann. 320, No.1, 149-165 (2001; Zbl 0982.57008)]. However, the manifolds involved are graph manifolds. Thus, Schanuel and Zhang asked whether every strict boundary slope of an orientable \(1\)-cusped hyperbolic \(3\)-manifold is strongly detected. The main result of the paper under review gives a negative answer to this question as follows. Let \(M\) be the complement of the pretzel knot \(K_n\) of type \((3,5,2n+1,2)\), \(n>1\). Then \(M\) is well known to be hyperbolic, and the slope \(4(n+4)\) is a strict boundary slope which is not strongly detected. Consider a mutant \(K_n^{\tau}\), which is the pretzel knot of type \((5,3,2n+1,2)\). By \textit{A. E. Hatcher} and \textit{U. Oertel} [Topology 28, No. 4, 453--480 (1989; Zbl 0686.57006)], \(K_n\) has boundary slope \(4(n+4)\), but \(K_n^{\tau}\) does not. This slope is shown not to be strongly detected by using a relationship between strongly detected boundary slopes of mutants established by \textit{D. Cooper} and \textit{D. D. Long} [J. Knot Theory Ramifications 5, No. 5, 609--628 (1996; Zbl 0890.57012)].