Experimental results suggest that anadromous salmon and trout learn (imprin
t) the odors of their natal site just prior to or during seaward migration.
In contrast, information on the life histories of several species and the
genetic structure of populations indicate that they must imprint earlier in
life, probably during incubation in the gravel or when they emerge as free
-swimming fry. To test the hypothesis that sockeye salmon (Oncorhynchus ner
ka) home to their incubation-emergence sites (rather than just to the lake
where they reared before migrating to sea), we took advantage of the natura
l variation in otolith microstructure caused by differences in thermal regi
mes during incubation. We collected otoliths from adult sockeye salmon that
returned to discrete spawning areas in Iliamna Lake, Alaska, and Lake Wash
ington, Washington, and, in blind trials, these were classified based on co
mparison with otoliths from juveniles from the same sites and using informa
tion on site-specific thermal regimes. Our analysis showed that the salmon
were much more likely to return to their natal incubation site than would h
ave occurred by chance. Estimated straying rates were about 0.1% from the W
oody Island population to the Pedro Pond population in Iliamna Lake and abo
ut 1% from the Cedar River population to the Pleasure Point population in L
ake Washington. The results were consistent with genetic evidence for fine-
scale structure of salmon populations and with conservation based on spatia
l scales appropriate for the early life history of the fish.