Knots with \(g(E(K))=2\) and \(g(E(K\#K\#K))=6\) and Morimoto's conjecture (Q1009758)

For a knot \(K\) in the \(3\)-sphere, \(t(K)\) denotes its tunnel number. An easy observation implies that \(t(K_1\sharp K_2)\leq t(K_1)+t(K_2)+1\). Equivalently, \(g(E(K_1\sharp K_2))\leq g(E(K_1))+g(E(K_2))\), where \(g\) denotes the Heegaard genus. If \(K_i\) admits a \((t(K_i),1)\) position, then the equality does not hold by \textit{K. Morimoto} [Math. Ann. 317, No. 3, 489--508 (2000; Zbl 0981.57005)]. He also showed that the converse holds if both knots are \(m\)-small, and conjectured that the converse holds in general. The main result of the paper under review is the existence of infinitely many knots \(K\) with \(g(E(K))=2\) and \(g(E(K\sharp K\sharp K))=6\). (This is just a special case of a result by \textit{T. Kobayashi} and \textit{Y. Rieck} [Algebr. Geom. Topol. 8, No. 2, 953--969 (2008; Zbl 1149.57009)].) This implies counterexamples, consisting of composite knots, to Morimoto's conjecture. The argument is as follows. Given an integer \(d>0\), \textit{J. Johnson} and \textit{A. Thompson} [On tunnel number one knots which are not \((1,n)\), preprint], \textit{Y. N. Minsky, Y. Moriah} and \textit{S. Schleimer} [Algebr. Geom. Topol. 7, 1471--1483 (2007; Zbl 1167.57002)] constructed infinitely many knots \(K\) so that \(E(K)\) admits a genus two Heegaard splitting of distance more than \(d\) in the sense of the curve complex. Fix such a knot \(K\) for \(d=10\). Then the Heegaard genus \(g(E(K\sharp K\sharp K))\) is shown to be six by using mainly technique developed by \textit{T. Kobayashi} and \textit{Y. Rieck} [Commun. Anal. Geom. 14, No. 5, 1037--1077 (2006; Zbl 1113.57006)].