38 MIGRATION OF ADULT SOCKEYE SALMON. 
The data here presented are quite too fragmentary to afford a basis 
for more than the most tentative conclusions; but the following points 
are suggested: 
1. The fish which go farthest up the main river before turning into 
the tributary in which they are to spawn run earlier in the season. 
2. The main bulk of the fish which pass up the main river past 
Hope are to be found in Puget Sound at the height of the season, the 
latter part of July and early in August. 
3. The fish entering the lower tributaries, Pitt River and the Harri- 
son Lake system, come largely from those fish constituting the last 
half of the run. 
In the event that it were deemed desirable to stop commercial 
fishing for part of the season, it would seem that the first part should 
be selected, as the indications are that there is a larger proportion of 
the upriver fish in the first half of the run, and these are the fish 
which have suffered more severely as a result of the disaster of 1913. 
SUMMARY. 
1. The experiment of tagging adult sockeye salmon in Puget 
Sound was initiated in an effort to determine the routes and rates 
of migration of Fraser River sockeyes in passing through the waters 
where commercial fishing is permitted. 
2. During July and August, 1918, numbered silver or aluminum 
buttons were attached to 4,494 adult sockeyes. Of these 1,199 were 
later recovered and data as to time and place of capture secured. 
3. The route most commonly followed passes from the Strait of 
Juan de Fuca across Washington Sound to the Salmon Banks and 
Whidbey Island, then through Rosario Strait and the southern part 
of the Strait of Georgia, past Point Roberts to the mouths of the 
Fraser River. 
4. The rate of migration as determined by the data for American 
waters is approximately 10 miles per day. 
5. The migration is more rapid during the last half than durimg 
the first half of the season. 
