SALMON AND SALMON FISHERIES OF SWIFTSURE BANK 
789 
In the lower section of figure 26, occurrence in the Rosario Strait area is com- 
pared with that of the Admiralty Bay-Bush Point area. Although both of these 
areas are immediately adjacent to waters in which their individual runs mingle, 
seasonal occurrence of cohos in Rosario Strait is considerably earlier than in the 
southern area, in fact the run in the former area has begun to decrease almost at the 
time that the southern run has first reached its peak. The Admiralty Bay-Bush 
Point runs also show a more prolonged peak of occurrence, and continue much later 
in the season. 
From table 38 it is evident that the Hope Island run is almost identical in occur- 
rence with that of the Admiralty Bay-Bush Point area, although the small catches 
in the early and late portions of the season appear only in the latter area. It is appar- 
ent that a large proportion of the Skagit River runs must pass around the southern 
end of Whidbey Island in the course of their migration. 
CHANGES IN ABUNDANCE 
CALCULATION OF TRAP INDICES 
It is apparent, from both trap and purse-seine catches, that there has been a 
considerable diminution in abundance of cohos in recent years. Inasmuch as traps 
and purse seines have been the principal types of gear catching coho salmon in this 
region, trends of abundance were determined from catches by both types of gear, and 
are presented together for comparison. 
In measuring abundance from traps, several difficulties are encountered which 
arise from the lateness of the coho runs in relation to those of other species taken by 
this type of gear. During early years, in certain areas where other species formed the 
principal catch, the traps were often removed from the water after fishing during only 
part of the coho season. Closed periods, which were imposed through legislation in 
many of the years after 1920, also prevented the traps in certain areas from fishing 
during the entire coho run. Years in which these closures were enforced cannot be 
compared directly to those in which fishing was not restricted unless some provision 
is made to offset the shorter fishing period. In most early years the traps were per- 
mitted to fish well into the winter months, while in later years legislation has often 
terminated the season before the entire run has appeared. For these reasons it was 
impossible to use the catches of any trap unless the opening and closing dates of its 
annual fishing seasons were known. This requirement sharply curtailed the available 
amount of data. 
In order to make the annual catch data comparable for both early and late years, 
they were weighted according to the length of the period fished. Inasmuch as the 
coho runs in the various districts are quite uniform in time from year to year, the 
average seasonal occurrences already presented were used as a basis for determining 
the time period of the runs in their respective districts. 
November 10 was arbitrarily selected as the end of the fishing season. The 
catches of traps which fished later in the year were reduced in proportion to the per- 
centage occurrence of the run after that date, and catches of traps which ceased 
