Although few biological data exist on the now extinct native lake trout, Salvelinus namaycush, of Lake Michigan, historical records and interviews with former commercial fishermen strongly suggest that this once widespread resource was composed of a number of discrete spawning populations or stocks. A natural division of the resource into distinct stocks is consistent with the size of Lake Michigan and its varied physiography. The native trout may have undergone subtle genetic changes over the millennia, although we cannot determine whether the physical and behavioral differences represented different genotypes or only temporary effects of the local environment. Because of physiographic similarities among the upper Great Lakes and probable interchanges of lake trout during the last glacial period, we recommend that progeny of extant wild stocks, particularly from Lake Superior, are genetically most suitable for recolonizing Lake Michigan. Moreover, the hatchery-held parents of such fish should be infused periodically with genes of the wild or feral donor populations. Despite the sound historical basis for these recommendations, we also recognize that sufficiently high stocking rates coupled with a reduction of heavy exploitation may be even more important than heritability in obtaining self-sustaining populations.Key words: lake trout, discrete stocks, Lake Michigan, historical changes, phenotype, homing, extinction, planted strains, rehabilitation