On the Neuwirth conjecture for knots (Q357199)

\textit{L. P. Neuwirth} conjectured in 1964, that every knot group \(G\) is a non-trivial free product with amalgamation, the amalgamation subgroup being free [Topology 2, 359--365 (1964; Zbl 0135.41804)]. This conjecture was proved by \textit{M. Culler} and \textit{P. B. Shalen} [Invent. Math. 75, 537--545 (1984; Zbl 0542.57011)], by proving that for any-nontrivial knot \(K\), there exists a properly embedded separating, orientable, incompressible and boundary incompressible surface in the exterior of \(K\), \(E(K)\). This conjecture was inspired by another conjecture also proposed by Neuwirth in the same paper, now called the Neuwirth conjecture, and which says that for any non-trivial knot \(K\), there exists a closed surface \(F\) containing \(K\) such that \(F\cap E(K)\) is connected, incompressible and boundary incompressible. This conjecture remains open, but it is satisfied by several classes of knots. There is a strong version of the conjecture [\textit{K. Ichihara, M. Ohtouge} and \textit{M. Teragaito}, Topology Appl. 122, No. 3, 467--478 (2002; Zbl 0997.57009)], which says that for any prime non-torus knot \(K\), there exists a non-orientable spanning surface \(S\) for \(K\) such that \(S\cap E(K)\) is algebraically incompressible and boundary incompressible. The boundary of a regular neighborhood of such a surface \(S\) then would satisfy the Neuwirth conjecture. The paper under review approach these conjectures by three methods. First by using Murasugi sums and previous results by the first author [J. Aust. Math. Soc. 91, No. 3, 391--404 (2011; Zbl 1241.57012)], it is proved that Montesinos knots satisfy the strong Neuwirth conjecture, and that all knots with 11 crossings or fewer, except perhaps for \(K11_{n}118\) and \(K11_{n}126\), satisfy the Neuwirth conjecture. Second, by using pre-essential surfaces, the following result is proved: Let \(K\) be a knot and suppose there is a closed incompressible surface \(\Sigma\) in \(E(K)\). Let \(i:\Sigma \rightarrow E(K)\) be the inclusion map, suppose that the induced map \(i_{*}:H_1(\Sigma, \mathbb{Z}) \rightarrow H_1(E(K),\mathbb{Z})\) has non-zero image but the induced map \(i_{*}:H_1(\Sigma, \mathbb{Z}_2) \rightarrow H_1(E(K),\mathbb{Z}_2)\) has zero image. Then \(K\) satisfies the Neuwirth conjecture. Third, by using degree one maps it is shown that if there is a degree one map \((S^3,K)\rightarrow (S^3,K')\) and \(K'\) satisfies the Neuwirth conjecture then \(K\) also satisfies it. In particular, if \(K'\) satisfies the Neuwirth conjecture and \(K\) is a satellite obtained from \(K'\), i.e. \(K\) is a satellite of some other knot with pattern given by \(K'\), then \(K\) also satisfies it. So to prove the Neuwirth conjecture, it suffices to prove it for simple knots and for satellite knots obtained from the trivial knot. At the end of the paper it is shown that there exist algorithms to decide if a given knot \(K\) satisfies the Neuwirth conjecture or the strong Neuwirth conjecture; these follow from normal surfaces theory.