Is there always an extremal Teichmüller mapping? (Q2565889)

Given a quasisymmetric self-homeomorphism \(h\) of the unit circle \(S^1\), let \(Q(h)\) be the set of all quasiconformal mappings with the correspondence \(h\). In ''Unique extremality'', [J. Anal. Math. 75, 299--338 (1998; Zbl 0929.30017)], by \textit{V. Božin, N. Lakic, V. Marković} and \textit{M. Mateljević}, it was shown that there exists \(h\) for which no extremal extension as a Teichmüller mapping is possible. This disproved some long standing conjectures. In the example constructed there, the boundary correspondence had a single extremal quasiconformal extension. In this paper we show that even when there are infinitely many quasiconformal extensions of the boundary values, it may still happen that none of them is a Teichmüller mapping. An infinitesimal version of this result is also established.