In their paper, Kent et al. (1996) present an excellent case history of the use of dip moveout (DMO) and velocity-filtering in the common midpoint (CMP) domain for the suppression of out-of-plane arrivals scattered from a deep sea-floor. However, they imply that as a result of a “small offset approximation” the use of DMO in this way is limited to surveys recorded in water depths of at least a few kilometers with conventional streamer offsets. This is incorrect. I argue here that the application of DMO will reduce to water velocity the stacking velocity of arrivals scattered from the upper surface of the seafloor without any restriction on water depth. Furthermore, I argue that this use of DMO is simply an example of the equivalence between 2-D and 3-D DMO for marine surveys where all source-receiver azimuths are equal, and that no “small offset approximation” is required. I first present a counter-example to the claim of Kent et al. (1996) that seafloor scattering cannot be suppressed using DMO in shallow water, and then consider in more detail their argument for the application of DMO to out-of-plane scattering. In the discussion that follows, I only consider DMO in the context of a constant velocity medium.