The Northern Refuge in Lake Michigan was established in 1985 as part of a rehabilitation program to stock yearling lake trout Salvelinus namaycush in areas with the best potential for success. Stocking of hatchery-reared lake trout within the refuge began in 1986 at three reefs: Boulder Reef, Gull Island Reef, and Richards Reef. On each reef from 1991 to 1997 we conducted gill-net surveys during the fall spawning season to evaluate performance of adult lake trout, and we conducted beam trawl surveys for naturally reproduced age-0 lake trout in the spring. Criteria to evaluate performance included spawner density, growth, maturity, and mortality. We found no evidence of natural reproduction by lake trout from our surveys. Nevertheless, density of spawning lake trout on Boulder Reef (69 fish/305 m of gill net/night) and Gull Island Reef (34 fish/305 m of gill net/night) appeared to be sufficiently high to initiate a self-sustaining population. Growth and maturity rates of lake trout in the Northern Refuge were similar to those for lake trout stocked in the nearshore region of Lake Michigan. In the Northern Refuge, growth rate for the Marquette strain of lake trout was slightly higher than for the Lewis Lake strain. Annual mortality estimates from catch curve analyses ranged from 0.46 to 0.41, and therefore, these estimates approached a level that was considered to be sufficiently low to allow for a self-sustaining population. Thus, it appeared that the lack of evidence for natural reproduction by lake trout in the Northern Refuge should not be attributed to inability of the population to attain a sufficiently large stock of spawners.