Walleyes (Stizostedion vitreum vitreum) from Lake Erie differed in allele frequencies from walleyes in Lake St. Clair (N = 1,680; 25 loci); however, only slight differences were found among walleyes from different spawning sites in each lake. Analyses of allele frequency data from samples of nonspawning walleyes taken in Lake St. Clair provided conditional maximum likelihood estimates that 86% of these walleyes were from Lake St. Clair and 14% from Lake Erie (SD = 19.7%) in 1983 and that 63% were from Lake St. Clair and 37% were from Lake Erie (SD = 20.0%) in 1984. About 30% of 1,159 recoveries of walleyes tagged in Lake Erie were taken from the Detroit River northward to southern Lake Huron, thus demonstrating extensive mixing of stocks from Lake Erie and Lake St. Clair. Additionally, tags recovered from the upper half of the St. Clair River provided an estimate of mixed stock composition of 76% Lake St. Clair fish and 24% Lake Erie fish. Analyses of tags returned during successive spawning seasons showed that walleyes strongly tended to return to suspected natal spawning areas. The tagging data thus corroborate the genetic evidence that walleye stocks from Lake St. Clair and Lake Erie are different, and that large numbers of Lake Erie walleyes enter Lake St. Clair during nonspawning seasons and return to their Lake Erie spawning sites each year. Western Lake Erie and Lake St. Clair walleyes should be considered as separate stocks for management purposes.