Symmetry of links and classification of lens spaces (Q1397915)

In 1935, Reidemeister [\textit{K. Reidemeister}, Abh. Math. Semin. Hamb. Univ. 11, 102-109 (1935; Zbl 0011.32404)] classified lens spaces up to orientation-preserving PL homeomorphisms. This proof became valid in the topological category with the proof of the Hauptvermutung by Moise in 1952 [\textit{E. E. Moise}, Ann. Math. (2) 56, 96-114 (1952; Zbl 0048.17102)]. In 1960, Brody [\textit{E. J. Brody}, Ann. Math. (2) 71, 163-184 (1960; Zbl 0119.18901)] established the topological classification without using the Haupvermutung, following an approach first given by Fox ([\textit{S. Eilenberg}, Ann. Math. (2) 50, 247-260 (1949; Zbl 0034.25304)], see Problem 2, [\textit{R. H. Fox}, Proc. Internat. Congr. Math. (Cambridge, Mass., Aug. 30-Sept. 6, 1950) 2, 453-457 (1952; Zbl 0049.13002)]). In the paper under review, the authors provide another proof of the classification of lens spaces up to orientation-preserving homeomorphisms. The method is motivated by that of Fox and Brody. Fox and Brody studied the Alexander polynomial of knots in lens spaces; the authors of the present paper study the Alexander polynomial of certain ``symmetric'' links in \(S^3\). Let \(l_i, m_i\) be the meridian and longitude of the solid torus \(T_i, (i=1,2)\), conveniently oriented. For coprime numbers \(p\) and \(q\) the lens space \(L(p,q)\) is obtained by identifying the boundaries of the solid tori in such a way that \(-m_2=pl_1+qm_1\) and \(-l_2=\overline{q}l_1+rm_1\) where \(q\overline{q}-pr=1\). The authors prove that \(L(p,q)\) and \(L(p,q')\) are equivalent up to orientation-preserving homeomorphism if and only if \(q-q'\) or \(qq'-1\) are multiples of \(p\). For \([x], [y]\) elements of the first homology group of the lens space \(L(p,q)\), write \(lk([x],[y])=\frac{c\cdot y}{n} \in \mathbb Q \;/\mathbb Z\) if \(c\) is a \(2\)-chain whose boundary is \(nx\) and \(c \cdot y\) is the intersection number. Let \(\rho :S^3 \rightarrow L(p,q)\) be the universal cover of \(L(p,q)\). Since \(\rho ^{-1}(l_1)\) is the trivial knot, the classification theorem (Corollary 2 in the paper) follows (quite immediately) from the following result (Theorem 1 in the paper): Assume that \(K\) is a knot in \(L(p,q)\) that represents a generator of its first homology group. If the Conway-normalized Alexander polynomial of \(\rho ^{-1}(K)\) is~\(1\), then \(lk([K],[K])=-q/p\) or \(-\overline{q}/{p}\) up to an integer.