Consider a complete bipartite graph \({{\bf{K}}_{{\bf{m,n}}}}\) with vertex set \({\bf{(}}{{\bf{u}}_{\bf{1}}}{\bf{,}}{{\bf{u}}_{\bf{2}}}{\bf{,}}....{{\bf{u}}_{\bf{m}}}{\bf{)}}\) and \({\bf{(}}{{\bf{v}}_{\bf{1}}}{\bf{,}}{{\bf{v}}_{\bf{2}}}{\bf{,}}....{{\bf{v}}_{\bf{n}}}{\bf{)}}\).Using breadth first search algorithm and choosing vertices alphabetically ,if we start at \({{\bf{v}}_{\bf{1}}}\) as the root of the spanning tree then at level 1 there will be all the vertices of the half of the bipartition \({\bf{(}}{{\bf{u}}_{\bf{1}}}{\bf{,}}{{\bf{u}}_{\bf{2}}}{\bf{,}}....{{\bf{u}}_{\bf{m}}}{\bf{)}}\) ,because they are the only neighbors \({{\bf{v}}_{\bf{1}}}\) of who are no so pairwise among themselves. Choose and all the remaining vertices of the other bipartition are neighbors of \({{\bf{u}}_{\bf{1}}}\) any edges between \({{\bf{v}}_{\bf{i}}}{\bf{,}}{{\bf{v}}_{\bf{j}}}\) .
Hence, they are the children of at level 2 as well as the leaves of the tree.