Indecomposable PD\(_{3}\)-complexes (Q422091)

By Turaev's splitting theorem a \(\text{PD}_3\)-complex decomposes as a connected sum if and only if its fundamental group is a free product. Crisp showed that every indecomposable \(\text{PD}_3\)-complex is either aspherical or its fundamental group is the fundamental group of a finite graph of finite groups. The main result of the paper under review is that ``if \(X\) is an indecomposable \(\text{PD}_3\)-complex and \(\pi_1(X)\) is the fundamental group of a reduced finite graph of finite groups but is neither \(\mathbb Z\) nor \(\mathbb Z \oplus \mathbb Z / 2 \mathbb Z\) then \(X\) is orientable, the underlying graph is a tree, the vertex groups have cohomological period dividing 4 and all but at most one of the edge groups is \(\mathbb Z / 2 \mathbb Z\). If there are no exceptions then all but at most one of the vertex groups are dihedral of order \(2m\) with \(m\) odd. Every such group is realized by some \(\text{PD}_3\)-complex. Otherwise, one edge group may be \(\mathbb Z / 6 \mathbb Z\).'' The author uses a reformulation of Turaev's criterion for a group to be the fundamental group of a \(\text{PD}_3\)-complex and one of Crisp's results. A section on indecomposable nonorientable \(\text{PD}_3\)-complexes shows that the only such complexes with virtually free fundamental group are \(S^1 \tilde \times S^2\) and \(S^1 \times \mathbb RP^2\). The final section is concerned with the question of whether every aspherical \(\text{PD}_3\)-complex is homotopy equivalent to a closed 3-manifold which the author proposes to reduce to a question about Dehn surgery on links.