We proposed and implemented procedures for partitioning future fish yields from the Great Lakes into taxonomic components. These projections are intended as guidelines for Great Lakes resource managers and scientists. Attainment of projected yields depends on restoration of stable fish communities containing some large piscivores that will use prey efficiently, continuation of control of the sea lamprey (Petromyzon marinus), and restoration of high-quality fish habitat. Because Great Lakes fish communities were harmonic before their collapse, we used their historic yield properties as part of the basis for projecting potential yields of rehabilitated communities. This use is qualified, however, because of possible inaccuracies in the wholly commercial yield data, the presence now of greatly expanded sport fisheries that affect yield composition and magnitude, and some possibly irreversible changes since the 1950s in the various fish communities themselves. We predict that total yields from Lakes Superior, Huron, and Ontario will be increased through rehabilitation, while those from Lakes Michigan and Erie will decline. Salmonines and coregonines will dominate future yields from the upper lakes. The Lake Erie fishery will continue to yield mostly rainbow smelt (Osmerus mordax), but the relative importance of percids, especially of walleye (Stizostedion vitreum vitreum) will increase. In Lake Ontario, yields of salmonines will be increased. Managers will have to apply the most rigorous management strictures to major predator species.