Flow-deflecting vanes were installed in the streambed along two meander bends with eroding bluffs in 2000 and 2001 in the upper main stem of North Fish Creek, a tributary to Lake Superior in Wisconsin. About 45 vanes were arranged in 15 arrays at each site to deflect the flow away from the eroding toe or base of the bluff (outside of a bend) and toward the point bar (inside of a bend). Channel cross-section and bluff-erosion surveys were done and streamflow and stage were measured before, during, and after vane installation to monitor changes in channel morphology and bluff erosion in the context of hydrologic conditions. There were two large floods in the study area in spring 2001 (recurrence interval of approximately 100 years) and in spring 2002 (recurrence intervals of approximately 50 years). Some maintenance and replacement of vanes were needed after the floods. Most of the channel-morphology changes resulted from the large floods, and fewer changes resulted from near-bankfull or at-bankfull flows (one in October 2002 and four in April and May 2003). At the bluff located 16.4 river miles upstream of the creek mouth (site 16.4), the vanes deflected flow and caused the channel to migrate away from the base of the bluff and toward the point bar, allowing sediment to deposit along the bluff base. The 361-foot reach at site 16.4 had a net gain of 6,740 cubic feet of sediment over the entire monitoring period (2000?03). Deposition (10,660 cubic feet) occurred mainly along the base of the bluff in the downstream part of the bend. Erosion occurred at site 16.4 along the streambed, the point bar side of the channel, and along a midchannel bar (1,220, 1,610, and 1,090 cubic feet, respectively). Less channel migration was observed during 2001-03 at another bluff located 12.2 river miles upstream of the creek mouth (site 12.2), which had a net loss of sediment through the 439-foot reach of 2,800 cubic feet over the monitored time period. The main volume of sediment was lost from the bluff toe in the downstream part of the bend (7,100 cubic feet). Monitored channel-morphology changes at site 12.2 were less than at site 16.4, most likely because installation was done after the April 2001 flood, which caused major changes in channel morphology at site 16.4, and because the monitoring period was shorter than at site 16.4. Bluff-erosion data from both sites indicate that mass wasting and block failures from the bluff top occur episodically and will continue to occur for decades or more.