In this note we point out an error in [2]. We show how to repair the proof in dimension 5. The results are true in general as can easily be seen from recent work of Borman, Eliashberg and Murphy [1]. The proof of Lemma 3.4 in [2] is incorrect. Below we will describe the problem with the proof and then show how it can easily be repaired in dimension 5. We then observe that Lemma 3.4, and thus the main results of the paper, is true in all dimensions based on recent work of Borman, Eliashberg and Murphy [1]. However this approach does not give an explicit construction and hence goes against the sprit of the original paper and in addition all the results of [2] follow directly from [1]. Acknowledgement: We thank Yasha Eliashberg for pointing out the error in the proof of Lemma 3.4 in [2]. The first author was partially supported by a grant from the Simons Foundation (#342144) and NSF grant DMS-1309073. 1. Exact Lagrangians, Liouville flows, and the error in the proof of Lemma 3.4 We begin by recalling the statement of Lemma 3.4 from [2]. To state the lemma we first establish some notation (that is slightly different that what was used in [2]). Consider T 2 × [0, 1] with coordinates (θ, φ, r) and the contact structure ξi = kerαi, i = 1, 2, where αi = ki(r) dθ + li(r) dφ. Here we have k1(r) = cos π 2 r and l(r) = sin π 2 r, and for i = 2 we have k2 and l2 agreeing with k1 and l1 near r = 0 and 1, and the curve (k2(r), l2(r)) in R has 5π/2 winding about the origin. In particular notice that ξ2 is obtained from ξ1 by adding Giroux torsion. Lemma 3.4 from [2] now reads as follows. Lemma 1. Let W be a manifold with contact form λ, there is a contact structure on W × [0, 1] × ([0, 1]× T ) such that the following properties are satisfied: (1) near W×{0}× [0, 1]×T 2 and W× [0, 1]×{0, 1}×T 2 the contact structure is contactomorphic to λ+ eα1, and (2) near W × {1} × [0, 1]× T 2 the contact structure is contactomorphic to λ+ eα2. Here t is the coordinate on the first [0, 1] factor. See [2] for details on how the main constructions and theorems of the paper follow from this lemma. The strategy of the proof in [2] was: (1) To construct a contact structure on W×[0, 1]×T 3 that near W×{0}×T 3 is given by λ+eβ0 and near W ×{1}× T 3 is given by λ+ e × β1, where βi is the contact structure on T 3 with Giroux torsion i and we are thinking of T 3 as S×T 2 with the S-factors Legendrian curves. (2) Then cut W × [0, 1]×T 3 along W × [0, 1]× ({θ0, θ1}×T ) so that one of the resulting pieces is as described in the lemma. To try to arrange this let β = p1 dθ1+p1 dθ2 be the Liouville form on T ∗T 2 = R×T 2 with coordinates (p1, p2, θ1, θ2). Notice that α = λ + β is a contact form on W × T ∗T . We will see below that we can arrange the two items above that are needed for our proof if there is a radial vector field v in R centered at a point p whose flow expands dβ (that is, Lvdβ = dβ) and a Lagrangian torus T 2 in a small neighborhood of {q}×T 2 ⊂ T ∗T 2 that is exact with respect to ιvdβ that is isotopic to {q}×T 2 by an isotopy disjoint from {p} × T . One may easily arrange all of this except for either the last 2010 Mathematics Subject Classification. 57R17 (primary), and 53D35(secondary).