788 
BULLETIN OF THE BUREAU OF FISHERIES 
off abruptly thereafter. The main portions of the runs of both areas are probably 
contributed by the Fraser and other northern rivers, but the second peak in the 
Birch Bay area may be composed largely of populations of such rivers as the Nicomekl 
and the Serpentine. 
A comparison of the data for the Point Roberts-Boundary Bay area with those 
from Rosario Strait indicate that the run in the latter area corresponds closely with 
that of the more northern dis- 
tricts; the somewhat earlier 
appearance here is probably 
due to the lesser distance of 
migration from the sea. 
There is a strong indication 
that a large part of the runs 
passing through Rosario 
Strait continues to Boundary 
Bay without entering Birch 
Bay. The Rosario Strait 
data are shown graphically in 
the lower section of figure 26. 
A comparison of seasonal 
occurrence of the Dungeness 
Spit-Middle Point area with 
that of the Admiralty Bay- 
Bush Point area indicates 
that the early appearance of 
cohos in the former area is 
consonant with its more sea- 
ward location. In this area 
a relatively heavy run follows 
the first group of fish, appear- 
ing during late August and 
early September, after which 
the intensity slackens. This 
is followed by the main run, 
Figure 26.— Seasonal occurrence of coho salmon in trap catches from the southern which reaches its peak in the 
part of Puget Sound. In the lower section of the figure occurrence in the principal w0ek in September and 
southern area is compared with that of one of the northern areas. R 
drops abruptly after the 
second week in October. It is evident that traps in this area fish a mixed popula- 
tion, but that the main run consists of fish bound for the southern areas. 
The Admiralty Bay-Bush Point run increases steadily from the last week in 
August to the third week in September, remains at a high level for three more weeks, 
and decreases steadily thereafter to the middle of November. Unlike the other areas, 
the run does not terminate here, but continues at a low level until early December. 
The data for seasonal occurrence in these areas is presented graphically for comparison 
in the upper section of figure 26. 
JUNE 
II 25 6 22 
AUG. SEPT. 
WEEK ENDING 
