790 
BULLETIN OF THE BUREAU OF FISHERIES 
fishing before that date were similarly increased. For 1921, and for other late years 
in which closed periods have been in force, trap catches were increased by the average 
percentage occurrence during the closed periods in their respective areas. Catches 
of traps which fished for a lesser period of time than that in which 75 percent of the 
run for their district normally occurred were not included in the analysis. 
A certain amount of error is unavoidably introduced by empirically increasing 
or decreasing catches for particular years to compensate for irregular length of fishing 
season. However, catches which were decreased in size were confined almost entirely 
to early years when fishing was less restricted, and nearly all increases in catches were 
made in later years when closed periods were imposed and fishing seasons were short- 
ened by legislative action. Such error as may have accompanied these necessary 
corrections would tend to reduce the apparent level of abundance in early years and 
to increase it in later years. Any decline shown in the trend of abundance would 
thus be given added validity. 
Three particular districts were selected for analysis. The first was that extending 
from Sandy Point to the international boundary, and included the Birch Bay, Bound- 
ary Bay, and Point Roberts areas (see fig. 2). Because of the size of the district 
and the large number of traps situated therein, catches were used from all traps for 
which suitable data were available. Prior to 1910 the data were meager, for sockeyes 
were of such importance in this region that catches of other species were often not 
recorded. After the tremendous sockeye run of 1913 most of the traps were removed 
before the coho run. In 1932, unfavorable economic conditions sharply reduced the 
number of traps fishing. During the remaining years, suitable data for from 7-12 
traps were available. These traps, although but a small part of the total number 
fishing in the area, represent a considerable portion of those which were fished late 
enough in the season to intercept the coho migration. The number of traps available 
in this area, and their total catch for each year, are tabulated in the first two columns 
of table 39. 
The second area selected was Rosario Strait. Although the runs in this district 
are largely composed of the same populations which pass through the northern areas, 
fishing conditions differ considerably, for the area of water through which the runs 
must pass is much more restricted and the number of traps is very small. Three of 
these traps, located in strategic positions, have taken the bulk of the catch in this 
area. Data for the 16-year period, from 1919-34, when at least two of these traps 
fished every year, all three of them for fifteen years, are tabulated in the third and 
fourth columns of table 39. It is evident that the efficiency of Rosario Strait traps 
is greater than that of those in the northern area, and their index of abundance 
should provide a useful check on that calculated from the larger group. 
