Reduced Bers boundaries of Teichmüller spaces (Q486744)

Let \(S\) be an orientable surface of genus \(g\) with \(p\) punctures with \(\xi(S) =3g+p\geq 4\). A quasi-Fuchsian group \(G\) is a quasiconformal conjugate of a Fuchsian group which uniformizes a Riemann surface homeomorphic to \(S\). The limit set of \(G\) is a Jordan curve in the Riemann sphere, the ideal boundary of hyperbolic \(3\)-space \(\mathbb{H}^3\), and its region of discontinuity \(\Omega\) is the union of two topological disks. The quotient manifold \(\mathbb{H}^3\cup\Omega/G\) is homeomorphic to \(S\times[0, 1]\). Let \(QF(S)\) denote the subspace of marked quasi-Fuchsian groups in the space of faithful discrete representations of \(\pi_1(S)\) to \(\text{PSL}_2(\mathbb{C})\). The Teichmüller space of \(S\) is denoted by \(\mathcal{T}(S)\). By the Ahlfors-Bers theory, there is a parametrization \(qf :\mathcal{T}(S)\times \mathcal{T}(\bar{S}) \to QF(S)\), where \(\mathcal{T}(\bar{S})\) is the Teichmüller space where markings are orientation reversing. For each \(m\in \mathcal{T}(S)\), the image of \(qf_{m}=qf(m,\cdot) : \mathcal{T}(\bar{S})\to QF(S)\) is called a \textit{Bers slice} and using the Bers embedding it is realized as a bounded domain in \(\mathbb{C}^{3g-3+p}\) so that to each point of its boundary \(\partial^{B}_{m}\), called the \textit{Bers boundary} with base point \(m\), corresponds a Kleinian group with unique invariant domain. This Kleinian group is called a \(b\)-group. It was proved by Kerckhoff and Thurston that if \(\xi(S)>4\), then there exist Bers slices for which the canonical identification between them does not extend to a homeomorphism between their Bers boundaries and that there is a Bers slice for which the action of the mapping class group has no continuous extension to its Bers boundary. The \textit{reduced Bers boundary} \(\partial^{RB}_{m}\) is the quotient space of \(\partial^{B}_{m}\) obtained by identifying two points if their related \(b\)-groups are quasiconformally equivalent. Since the phenomenon observed by Kerckhoff-Thurston is caused by the existence of quasiconformal deformation spaces contained in the Bers boundary, Thurston conjectured that (1) the reduced Bers boundary \(\partial_{m}^{RB}\) is independent of the base point. He also conjectured that (2) \(\partial_{m}^{RB}\) is homeomorphic to the space of unmeasured laminations on \(S\). The purpose of the paper under review is to settle these conjectures. The author solves Conjecture (1) affirmatively in Theorem 3.7. The natural identification between \(qf_{m_1}(\mathcal{T}(\bar{S}))\) and \(qf_{m_2}(\mathcal{T}(\bar{S}))\) extends to a homeomorphism from \(\partial^{RM}_{m_1}\) to \(\partial^{RM}_{m_2}\) for two points \(m_1\), \(m_2\) of \(\mathcal{T}(S)\) (so we will omit the reference to the base point \(m\) from the notation \(\partial^{RB}_m\)). The author also proves that the action of the extended mapping class group on \(\mathcal{T}(S)\) extends continuously to \(\mathcal{T}(S)\cup \partial^{RB}\). For Conjecture (2), it is important to know that the union of ending laminations and parabolic curves is an invariant of the quasiconformal equivalence class of boundary \(b\)-groups. So there is a bijection \(e\) from \(\partial^{RB}\) to a subspace of unmeasured lamination space of \(S\). The author shows in Propositions 3.4 and 3.5 that neither \(e\) nor \(e^{-1}\) is continuous if \(\xi(S)>4\), contrary to Thurston's expectation. Besides the solution of Thurston's conjectures, the author considers other aspects of the reduced Bers boundary. Section 4 treats the rigidity of auto-homeomorphisms of the reduced Bers boundary. The main theorem here, Theorem 4.1, states that if \(\xi(S)>4\), then any auto-homeomorphism \(f\) of \(\partial^{RB}\) is induced from an extended mapping class, and, unless \(S\) is a closed surface of genus \(2\), \(f\) is induced from a unique extended mapping class. This theorem should be compared with the analogous result by Papadopoulos and the author about the auto-homeomorphisms of the space of unmeasured laminations. In the last section, Section 5, the author considers the problem of characterizing the limit point in the reduced Bers boundary of a sequence in \(\mathcal{T}(S)\). Given an unbounded sequence \(\{m_i\}\) in \(\mathcal{T}(S)\), we can find a subsequence which determines a \textit{multi-layered Thurston limit} \((\lambda_1,..., \lambda_{n})\). Here \(\lambda_j\) is a measured lamination on a subsurface \(S_{j}\) of \(S\) such that \(S=S_1\supset \cdots \supset S_n\). Then one can detect the limit of the subsequence by \((\lambda_1,..., \lambda_{n})\) up to some ambiguity caused by arc components which \(\lambda_j\) may contain. The proofs of the results of this paper rely heavily on a paper by the author and \textit{T. Soma} [``Geometry and topology of geometric limits. I.'', Preprint (2010; \url{arXiv:1002.4266})]. This paper founded the theory of brick manifolds which gives models of hyperbolic \(3\)-manifolds arising from the geometric limits of sequences of Kleinian groups. For example the author uses the result that every quasiconformal deformation of a geometric limit \(H\) of a sequence of quasi-Fuchsian groups cannot be supported on its limit set even if \(H\) is infinitely generated.