We show that if a graph G has average degree d ≥ 4, then the Ihara zeta function of G is edgereconstructible. We prove some general spectral properties of the edge adjacency operator T : it is symmetric for an indefinite form and has a "large" semi-simple part (but it can fail to be semi-simple in general). We prove that this implies that if d > 4, one can reconstruct the number of non-backtracking (closed or not) walks through a given edge, the Perron-Frobenius eigenvector of T (modulo a natural symmetry), as well as the closed walks that pass through a given edge in both directions at least once.The appendix by Daniel MacDonald established the analogue for multigraphs of some basic results in reconstruction theory of simple graphs that are used in the main text.Date: August 22, 2018 (version 2.0). 2010 Mathematics Subject Classification. 05C50, 05C38, 11M36, 37F35, 53C24.A degree-one vertex is called an end-vertex. All results in this paper hold for connected finite undirected multigraphs without end-vertices, and from now on we will use the word "graph" for such multigraphs.If e = {v 1 , v 2 } ∈ E, we denote by e = (v 1 , v 2 ) the edge e with a chosen orientation, and by e = (v 2 , v 1 ) the same edge with the inverse orientation to that of e. Let o( e) = v 1 denote the origin of e and t( e) = v 2 its end point. If there are multiple edges between v 1 , v 2 then we will label them e i = (v 1 , v 2 ) i . A nonbacktracking edge walk of length n is a sequence e 1 e 2 ....e n of edges such that t(e i ) = o(e i+1 ), but e i+1 = e i . We call it tailless if e n = e 1 . Just like walks in the graph can be studied using the adjacency matrix, nonbacktracking walks are captured by the edge adjacency matrix T = T G studied by Sunada [24], Hashimoto [14] and Bass [2]. Letting E denote the set of oriented edges of G for any possible choice of orientation, so | E | = 2|E|, T is defined to be the 2|E| × 2|E| matrix, in which the rows and columns are indexed by E, and T e 1 , e 2 = 1 if t( e 1 ) = o( e 2 ) but e 2 = e 1 ; 0 otherwise.