flow in a pipe. The injection point was situated where the flow was welldeveloped. Thus, subsequent streamwise variations in pressure drop were due to the injected polymer spreading out across the pipe. The axial development of local drag reduction was monitored by a series of closely-spaced pressure tappings. The corresponding radial dispersion of the injected polymer, as it travelled downstream, was assessed by sampling the flow at various points.Local drag reduction, due to either point injection at the centerline or injection through a slot in the wall, was found to increase with distance downstream. This increase was related to the streamline increase of polymer concentration in a narrow annulus near the pipe wall. It was tentatively concluded that the effective annulus was bounded by 15 5 y + I 100, in agreement with previous deductions from less direct evidence.