Branched spines and Heegaard genus of 3-manifolds (Q2385056)

The paper under review compares two ways of constructing a closed, orientable 3-manifold \(M\). The first is Heegaard splitting, namely, the construction of \(M\) as the union of two handlebodies. The complexity associated with a Heegaard splitting is the genus of the splitting surface, and the associated complexity of \(M\) is the genus of the least genus a Heegaard surface, denoted \(HG(M)\). The second construction is defined via a flow spine, which we now explain. A flow spine is a branched surface \(B\subset M\) whose branches are all of valence 2 (that is, at each branch two sheets are attached smoothly from one side to a third sheet), so that cl\((M \setminus N (B))\) is a ball. Moreover, we require that the horizontal boundary of \(N (B)\) is exactly an annulus. See the paper under review and references therein for further details. The combinatorial information that determines \(B\) is called the \(\mathcal E\)-datum; see, for example [\textit{R. Benedetti} and \textit{C. Petronio}, Branched standard spines of 3-manifolds. Lecture Notes in Mathematics 1653. Berlin: Springer (1997; Zbl 0873.57002)]. \textit{M. Endoh} and \textit{I. Ishii} [Jap. J. Math., New Ser. 31, No.~1, 131--156 (2005; Zbl 1081.57017)] defined a complexity for \(\mathcal E\)-data, called the block number, and extended it to a complexity of \(M\), called the block number of \(M\) and denoted \(Bl(M)\); \(Bl(M)\) is the minimum of all block numbers of \(\mathcal E\)-data of flow spines of \(M\). Endoh and Ishii showed that \(HG(M)\leq Bl(M)\) for every closed orientable 3-manifold \(M\) except for \(S^2 \times S^1\). The main construction of this paper uses a Heegaard surface and produces a flow spine. As a result, the author proves: Theorem 1. Every closed orientable 3-manifold \(M\) except for \(S^ 2 \times S^ 1\) and \(S^ 3\) satisfies \(Bl(M) =HG(M)\). The author also proves that the \(\mathcal E\)-datum determines the manifold: Theorem 2. Let \(P_ 1\) and \(P_ 2\) be flow-spines of closed oriented 3-manifolds \(M_ 1\) and \(M_2\), respectively, such that they both have the same \(\mathcal E\)-datum \(E\). Then \(M_ 1\) is homeomorphic to \(M_ 2\) via an orientation-preserving homeomorphism. This result is surprising in light of a construction of Benedetti and Petronio [loc. cit., Theorem 5.1.3] that demonstrated two non-homeomorphic spines with the same \(\mathcal E\)-datum.