The end of the curve complex (Q646381)

If \(S\) is a genus \(g\) surface with \(b\) boundary components, so that \(3g-3+b \geq 2\), then the curve complex \(\mathcal C(S)\) has a vertex for each isotopy class of essential non-peripheral simple closed curves in \(S\) and a \(k\)-simplex for each collection of \(k+1\) disjoint vertices having disjoint representatives. By regarding each simplex as a Euclidean simplex of side-length one, \(C(S)\) turns out to be Gromov hyperbolic [\textit{H. A. Masur} and \textit{Y. N. Minsky}, Invent. Math. 138, No. 1, 103--149 (1999; Zbl 0941.32012)]. The present paper proves that, if the surface \(S\) has exactly one boundary component and genus two or more, then for each vertex \(\omega\in \mathcal C(S)\) and for any \(r \in \mathbb N\), the subcomplex spanned by \(\mathcal C^0(S)- B(\omega, r)\) is connected (where \(B(\omega, r)\) denotes the ball of radius \(r\) about the vertex \(\omega\)). In order to prove the above result, the author makes use of the fact that the complex of curves has no dead ends (Prop. 3.1 of the paper) and of the so-called Birman short exact sequence, see [\textit{J. S. Birman}, Braids, links, and mapping class groups. Based on lecture notes by James Cannon. Annals of Mathematics Studies. 82. Princeton, N.J.: Princeton University Press and University of Tokyo Press (1975; Zbl 0305.57013)] and [\textit{I. Kra}, Acta Math. 146, 231--270 (1981; Zbl 0477.32024)]. Note that, for the considered surfaces, the above result directly answers a question of Masur, and it also answers a question of \textit{G. C. Bell} and \textit{K. Fujiwara} [J. Lond. Math. Soc., II. Ser. 77, No. 1, 33--50 (2008; Zbl 1135.57010)] in the negative. It is also evidence for a positive answer to a question of P. Storm (already verified in an independent way by \textit{D. Gabai} in [Geom. Topol. 13, No. 2, 1017--1041 (2009; Zbl 1165.57015)].