Stream restoration attempts to reverse the global degradation of rivers and streams, but rigorous evaluations are needed to advance the science. We evaluated a 3^rd-order channelized Indiana (USA) stream that was restored in 1997 by constructing two meanders, each ∼400 m long. Pool and riffle sequences were constructed, coarse substrate and wood were added to the channel, banks were stabilized and revegetated, and sedimentation was reduced by creating a sediment retention basin upstream. Habitat, periphyton, macroinvertebrates, and fishes were measured before restoration and for 5 y after restoration in the restored reaches and in an upstream, unrestored reach. Restoration improved habitat conditions (e.g., more pools, fewer fine sediments) in both restored reaches compared to the unrestored reach. Within 1 y after restoration, major trophic groups (i.e., periphyton, macroinvertebrates, and fishes) recovered to or exceeded levels in the degraded, unrestored reach. However, biotic responses varied with time, trophic level, and community parameter measured. Five years after the restoration, habitat quality, algal abundance, and macroinvertebrate density remained higher in the restored reaches, whereas macroinvertebrate diversity and fish abundance in the restored reaches were similar to or below levels in the unrestored, channelized reach. Although biotic recovery was relatively rapid, long-term persistence is uncertain because of continued sedimentation at a watershed scale. In many instances, reach-scale restorations may be ineffective in the face of basin-wide degradation. This study illustrates the importance of conducting long-term assessments of stream restorations, which can improve both knowledge and management of stream ecosystems.