“…The main tool is a powerful embedding technique, sometimes called extendability methods or tree embeddings with rollbacks, which was first introduced by Friedman and Pippenger [7] in 1987 and subsequently improved by Haxell [11] in 2001. Here we will use a modern reformulation of this technique which is attributed to Glebov, Johannsen, and Krivelevich [8], and which has played a major role in the solution of several problems in the last few years (see [3,4,6,9,10,15,13,18,19] for instance). Roughly speaking, the extendability method (Lemma 3.13) says that if we are given a subgraph S i ⊂ G which is 'extendable' and G has good expansion properties, then we can extend S i by adding a leaf e i with one of its endpoints in V (S i ) and other in V (G) \ V (S i ) so that S i + e i remains extendable.…”