To go back to the origin of the thread, I believe taking the idea of "crystal planes" too literally is one source of the confusion with reflection. Assuming an incoming plane wave with k0(vector), (magnitude: 2 pi/lambda) each scattering center at position r(vector), is at a different relative phase, given by k0 (dot) r . If we look at outgoing waves in a particular direction given by k1, and detect them at a distant position R, then the phase of each scattering contribution is given by: k0(dot)r + k1(dot)(R-r)or (k0-k1)(dot)r + k1(dot)R If we ask for the locus of scattering centers for which the first term is constant, ie r for which V(dot)r = const., then that is a geometric plane. We then ask for the phase difference between "planes" and find that for some k1 values, that difference can be a multiple of (2 pi). Note that these "planes" are not necessarily face planes of the macroscopic crystal.