The N-heterocyclic carbene-coordinated ruthenium benzylidene complex [(H 2 IMes)(PCy 3 )(Cl) 2 Ru¼CHPh] (1) is a highly active catalyst for a wide variety of olefin-metathesis reactions, [1] including those with sterically demanding [2] and electron-deficient olefins (Scheme 1). [3] In spite of recent advances, there are several processes that remain challenging, such as olefin cross metathesis (CM) with directly functionalized olefins. [4] For example, acrylonitrile CM has only been successful with Schrock©s arylimido molybdenum alkylidene catalyst [5] and the ether-tethered ruthenium alkylidene derivative [(H 2 IMes)(Cl) 2 Ru¼CH(o-iPrOC 6 H 4 )] (4). [6,7] Phosphane-ligated ruthenium catalysts have given poor results for this transformation, [3a, 5c, 6c, 8, 9] except for one report of efficient CM between purified acrylonitrile and 1-decene mediated by 1. [10] We have determined that dissociation rates of ligands are related to catalyst efficiency during CM with acrylonitrile. On this basis, we have developed a new, highly efficient ruthenium complex to perform acrylonitrile CM with unpurified acrylonitrile; this catalyst is the fastest initiator of any ruthenium-based catalyst reported to date. [11] Previous studies have shown that precatalysts of the type [L 2 X 2 Ru ¼ CHR] initiate by dissociating one L-type ligand before entering the catalytic cycle (Scheme 2); [12] in complexes 1±3, L is a phosphane (PR 3 ), and in complex 4, L is a tethered ether ligand (iPrO). Importantly, complexes 1±4 all provide the same propagating species (A and B) after a single turnover. [13] If either A or B is trapped by L, dissociation of L must occur before catalysis can continue. The relative affinity of A and/or B for the olefin in preference to L (i.e., favoring propagation) controls how long these species remain in the catalytic cycle. Consequently, the differences in activity between catalysts 1±4 depend on their rates of initiation and rebinding of L, both of which can be tuned by the nature of L. [12b, 14±16] Complexes 1±4 are all active for a variety of metathesis processes, such as CM, ring-closing metathesis (RCM), and ring-opening-metathesis polymerization (ROMP). However, the situation involving acrylonitrile CM is more complex; CM between acrylonitrile and allylbenzene proceeds efficiently with 4 (68 % yield) but not with 1±3 (21 %, 35 %, and 36 % yields, respectively; Table 1). [17] Clearly, this difference cannot simply be an issue of precatalyst initiation; 2 and 4 have roughly the same initiation rates and initiation is faster with 3 than 1, 2, or 4 Table 2. The difference is also not a result of COMMUNICATIONS