Acoustic backscatter data were collected from the ocean bottom at four sites on the Sohm Abyssal Plain. The bottom sediment at the four sites varied from mud/clay to silt/gravel. Using a simple array consisting of a free-flooding-ring projector which was omnidirectional in azimuth, and an omnidirectional hydrophone, data were collected over the frequency range of 800–2400 Hz [Hines and Barry, J. Acoust. Soc. Am. 92, 315–323 (1992)]. To obtain sufficient signal-to-noise ratio without impairing spatial resolution, 200-Hz-wide linear FM pulses were employed. By deploying the array at a height of 500 m above the seabed, bottom backscatter data were obtained at grazing angles down to 4 degrees before the onset of the first surface return. For data arriving after the first-bottom–first-surface interaction, the geometry approximates an in-plane bistatic experiment. Estimates of the in-plane bistatic scattering strength were obtained for pairs of incident and scattered grazing angles (φi,φs)<(50 degrees,88 degrees), at frequencies of 900, 1200, and 1600 Hz. At one site the grazing angle dependence matched Lambert’s rule; at all other sites the data exhibited a steeper slope than that of Lambert’s rule. There was insufficient data to extract the frequency dependence. The fathometer returns were used to estimate the normal-incidence bottom loss at the four sites. The measured bottom loss ranged from a low of 2 dB to a high of 16 dB. The in-plane bistatic data and the monostatic data were compared to examine the validity of the separable and half-angle approximations [Ellis and Crowe, J. Acoust. Soc. Am. 89, 2207–2214 (1991)]. The separable approximation provided a reasonable fit to the data at three of the four sites; the half-angle approximation did not match the data well at any of the sites.