Note also that our Eqs. ~10! are more general than
those in Ref. 13, since we start from a nonlinear
equation as in Eq. ~1! rather than the linearized approach
taken in Ref. 13. In this sense our analysis
takes into account diffraction from 2K gratings, such
as the term E2Y0Y2 in Eq. ~10a!. Furthermore,
terms such as Y1
2 in Eq. ~10a! are not present