GROWTH AND MATURITY OF SALMON IN THE OCEAN 
65 
There is a remarkably close agreement between the figures for these two collec- 
tions, when it is considered that they represent different years and the work of two 
entirely independent observers. The proportion of fish with ocean and with stream 
nuclei is almost identical — 6.3 per cent of the 1918 collection and 6.7 of the 1921 col- 
lection having scales with the stream type of nuclei. 
It was shown above that the percentage of fish with stream nuclei among the 
schools off the mouth of the Columbia River was close to 22 per cent, and the fact 
was mentioned that Fraser has found that 34.6 per cent of the chinook salmon taken 
during 1915 and 1916 in the straits of Georgia showed the stream type of nucleus. 
Although in 1917 Fraser found that the percentage of fish with the ocean type of 
nuclear growth was 78 per cent, identical with our findings in the fish taken near the 
mouth of the Columbia, it would seem probable that the salmon in the Straits of 
Georgia tend to contain a larger percentage of individuals that have remained an 
entire year in fresh water before migrating seaward. Gilbert (1922a) has found 
that the chinook or king salmon of the Yukon River invariably have scales with the 
stream type of nucleus. 
These data indicate with considerable certainty that latitude, or more properly 
the differences in climatic conditions obtaining in different latitudes, has a definite 
effect upon the early history of the chinook salmon. In the more southerly latitudes 
the tendency is for the young fish to migrate soon after they become free-s\vimming, 
early in their first year. Under the influence of the more severe climatic conditions 
of the north, which cause later hatching and slower growth, the tendency is for 
greater and greater percentages of the fish to remain in the home stream during the 
first year. 
A similar condition has been found to exist among the salmon of Norway. 
Dahl (1910) gives tables showing for thi'ee localities and for two years the per- 
centages of these salmon that migrate seaward after spending 2, 3, 4, and 5 winters in 
fresh water. He concludes : 
A close examination of these tables shows that the age of the smolts at migration varies 
between 2 and 5 winters. In the south the smolts are generally young, but the farther north we 
go the more pronounced is the tendency for the fish to remain longer in the river before migration. 
In this connection it is interesting to note that in a large river basin such as 
that of the Columbia River a distribution of races is found which is suggestive of 
this distribution in latitude. It is certain, from unpublished data in the hands of 
the writer, that as a rule the races in the Columbia which spawn in the higher 
tributaries are predominantly those which have the habit of remaining in fresh 
water during the first year and whose scales, therefore, show the stream type of 
nucleus. The races in the lower tributaries, on the other hand, migrate, as a 
rule, during the first year, and as a result the scales of the adults show the ocean 
type of nuclear growth. It is evident, therefore, that the effect of altitude on the 
early history of the chinook salmon is very similar to the effect of latitude — an 
increase in either altitude or latitude resulting in an increase in the number of 
young that remain in the stream during the entire first year. This similar influence 
of altitude and latitude is well known from the numerous studies of the geographical 
distribution of land animals and plants, and it is interesting to note the parallel 
effect on racial habit in such an anadromous form as the chinook salmon. 
