“…A first attempt to predict the observations of Feir (1967) theoretically was made by Roskes (1977), using the NLS, modified with some additional terms representing higher-order modulation effects. It is noteworthy that, although, as it turns out, Roskes (1977) neglected the interaction of the wave envelope with the induced mean flow, his numerical calculations were capable of reproducing, a t least qualitatively, the group splitting observed by Feir (1967). Later, Dysthe (1979) carried out a formal derivation of an envelope equation which extends the range of validity of the NLS to longer time, comparable with O(E-~) wave periods; in addition to the higher-order terms proposed by Roskes (1977), Dysthe's equation also includes the effect of the wave-induced mean flow.…”