1881.] Observations on the Salmon of the Pactfie: 179 
each of the subordinate runs, varying with each different river. 
In general, the runs are slack in the summer and increase with 
the first high water of autumn. By the last of August only 
straggling blue-backs can be found in the lower course of any 
stream, but both in the Columbia and the Sacramento the quinnat 
runs in considerable numbers till October at least. In the Sacra- 
mento the run is greatest in the fall, and more run in the summer 
than in spring. In the Sacramento and the smaller rivers south- 
ward, there is a winter run, beginning in December. 
The spring salmon ascend only those rivers which are fed by 
the melting snows from the mountains, and which have sufficient 
volume to send their waters well out to sea. Such rivers are the 
Sacramento, Rogue, Klamath, Columbia and Frazer's rivers. 
Those salmon which run in the spring are chiefly adults (sup- 
posed to be at least three years old). Their milt and spawn are no 
more developed than at the same time in others of the same spe- 
cies which will not enter the rivers until fall. It would appear that 
the contact with cold fresh water, when in the ocean, in some way 
caused them to turn toward it and to “run,” before there is any 
special influence to that end exerted by the development of the 
organs of generation. ‘ . 
High water on any of these rivers in the spring is always fol- 
lowed by an increased run of salmon. The canners think, and 
this is probably true, that salmon which would not have run till 
later, are brought up by the contact with the cold. water. The: 
cause of this effect of cold fresh water is not understood. We 
may call it an instinct of the salmon, which is another way of 
expressing our ignorance. In general, it seems to be true that 
in those rivers and during those years when the spring run is 
greatest, the fall run is least to be depended on. ' 
As the season advances, smaller and younger salmon of these 
two species (quinnat and blue-back) enter the rivers to spawn, 
and in the fall these young specimens are very numerous. We 
have thus far failed to notice any gradations in size or appearance 
of these young fish by which their ages could be ascertained. It 
is, however, probable that some of both sexes reproduce at the age _ 
of one year. In Frazer’s river, in the fall, quinnat male grilse of 
every size, from eight inches upwards, were running, the milt 
fully developed, but usually not showing the hooked jaws and 
dark colors of the older males. Females less than eighteen 
