A “tethered stake” apparatus is used to measure basal sliding in a borehole on Ice Stream B, West Antaretica, about 300 km upstream (east) from its grounding line near the head of the Ross Ice Shelf. A metal stake, emplaced at the top of a laver of unfrozen till underlying the ice, is connected by a tether line to a metering unit that measures the tether line as it is pulled out from the borehole by the stake as a result of basal sliding. The measured sliding motion includes any actual slip across the ice–till interface and may include in addition a possible contribution from shear deformation of till within about 3 cm of the interface. This 3 cm figure follows from a qualitative model of the movements of the stake in the course of the experiment, based on features of the record of apparent sliding. Alternative but less likely models would increase the figure from 3 cm to 10 cm or 25 cm. In any case it is small compared to the seismically inferred till thickness of 9 m. Measured apparent sliding averages 69% of the total motion of 1.2 m d−1 over 26 days of observation if a 3.5 day period of slow apparent sliding (8% of the total motion) is included in the average. The occurrence of the slow period raises the possibility that the sliding motion switches back and forth between c.80% and c. 8% of the total motion, on a time-scale of a few days. However, it is likely that the period of slow apparent sliding represents instead a period when the stake got caught on the ice sole. If the slow period is therefore omitted, the indicated average basal sliding rate is 83% of the total motion. In either case, basal sliding predominates as the cause of the rapid ice-stream motion. In the last 2 days of observation the average apparent sliding rate reached 1.17 m d−1, essentially 100% of the motion of the ice stream. If till deformation contributes significantly to the ice-stream motion, the contribution is concentrated in a shear zone 3 cm to possibly 25 cm thick at the top of the 9 m thick till layer. These observations, if applicable to the West Antaretic ice sheet in general, pose complications in modeling the rapid ice-streaming motion.