We incubated lake trout (Salvelinus namaycush) eggs over winter at shallow (10 m) and deep locations (20 m) on Gull Island Shoal, Lake Superior; at a shallow-water (10 m) site off the mainland (Bark Point); and in flowing Great Lakes water at two laboratories. Survival to hatch was significantly higher in the laboratories and averaged 80.9%. In Lake Superior, egg survival among incubators at all sites was significantly higher (P < 0.0001) for incubators that remained buried in spawning substrates (15.1–21.0%) than for incubators that were partially or completely exposed to water currents (1.0–12.6%). Egg survival for incubators that remained buried at the shallow-water sites was significantly higher at Bark Point (44.6%) than at Gull Island Shoal (21.0%). Egg survival among incubators that remained buried at the deep (14.4%) and shallow-water sites (21.0%) on Gull Island Shoal was not significantly different. Because incubators that were completely buried or partially exposed only appeared to differ in their degree of exposure, we concluded that survival of eggs in the lake was reduced by mechanical stress associated with water turbulence. Lower egg survival at Gull Island Shoal, a known lake trout-spawning site, was not expected and appeared to have been caused by a strong gale that occurred when these eggs were in late epiboly, a sensitive embryological stage. We present a hypothesis suggesting that lake trout recruitment in the Great Lakes is limited by availability of spawning habitat.