Seismic compressional-wave data from short refraction shooting carried out during the 1974–75 and 1976–77 seasons at J-9, the site of the Ross Ice Shelf Drilling Project, have been compared. Significant dissimilarities were found to exist between the two sets of data. The measurements were made at locations about 2 km apart, with three unreversed profiles 60° apart recorded during the 1976–77 season and one unreversed profile during the 1974–75 season. The resulting velocity–depth profiles, and hence the derived density–depth profiles, differ by as much as 8%, with the 1976–77 results indicating a maximum velocity, corresponding to solid ice, at a shallower depth than the 1974–75 data. Both profiles were subjected to the same analysis, and a comparison of travel-time curves shows the differences to be real. Densities measured on cores from a 100 m bore hole drilled in 1974–75 about 50 m from the center of the 1974–75 profile agree well with densities computed from that profile. The density difference is believed to be due to the passage of the ice through the high-stress system associated with the interaction between Ice Stream B, flowing in from the West Antarctic ice sheet, and the Ross Ice Shelf. A reversed refraction profile carried out at station B.C. about 30 km up-stream, shows evidence of dipping layers that may be similarly caused.