Geosphere laminations in free groups (Q431037)

Geodesic laminations were constructed in surfaces. In the paper under review, laminations are defined for free groups. More specifically, laminations are defined for 3-dimensional manifolds whose fundamental group is a free group. The main difference to other types of laminations defined so far is that the new definition defines codimension one laminations. The manifold \(M\) that is used is the connected sum of \(k\)-copies of \(S^2{\times}S^1\) with fundamental group the free group on \(k\)-generators. The manifold \(M\) can be constructed by identifying certain boundary components from the sphere after removing the interiors of embedded 2-spheres. Starting from a system of 2-spheres in \(M\), there is a standard construction, \textit{A. E. Hatcher} [Comment. Math. Helv. 70, No. 1, 39--62 (1995; Zbl 0836.57003)], that associates a tree \(T\) to the universal cover \(\tilde{M}\). The analogue of geodesics are geospheres that are defined as triples \(({\tau}, {\epsilon}, e_+)\) where \(\tau\) is a subtree, \(\epsilon\) is a sign assignment on the set of non-standard bivalent vertices of \(\tau\) and \(e_+\) associates, to each terminal vertex of \(\tau\), a vertex not in \(\tau\). This definition generalizes the triple defined by a sphere in normal form in \(\tilde{M}\). The only difference is that the normal sphere determines a finite subtree, something that is not assumed in the geospheres. The authors define a topology on the set of all geospheres in \(\tilde{M}\) and they show that the resulting space is totally disconnected. Similarly, they define geospheres on subtrees of the tree. Then, the space of all geospheres in a subtree is a compact space. Let \(E(\tilde{M})\) be the space of ends of \(\tilde{M}\). Then, homology classes of embedded spheres are determined completely by a partition of \(E(\tilde{M})\) into two open subsets. For the tree \(T\), a geosphere provides a splitting of \(E(T)\) into three subspaces, one closed and two open. That allows to define when two geospheres do not cross. Now the authors have all the pieces for defining a geosphere lamination in \(M\). That is a closed subset of geospheres in \(\tilde{M}\) that is \({\pi}_1(M)\)-invariant and is embedded in \(M\) (i.e. any two geospheres do not cross). They define a topology on the set of all geosphere laminations and show that it s compact. Also, they describe the space of geospheres as the space of partitions of \(E(T)\). At the end of the paper, there is a list of open questions related to geosphere laminations.