660 
BULLETIN OF THE BUREAU OF FISHERIES 
The relation between these recoveries and the total recoveries in area 1 , expressed 
in percentage, is also given in table 6. It will be noted that the weekly percentage of 
recoveries in area 1, from the localities east of Cape Fox, Revillagigedo Channel, and 
Behm Canal, are highest in the beginning of the season and lowest toward the end. 
The weekly percentage of recoveries along the west shores of Gravina, Annette, and 
Duke Islands is, on the other hand, lowest at the first part of the season and highest 
toward the last. In other words, there is a seasonal increase in the percentage of 
tagged salmon caught in area 1 along the west shores of Gravina and Annette Islands, 
whereas there is a seasonal decrease in the percentage of the tagged salmon caught in 
other localities of the area. 
Since there are very few streams along the west shores of Annette and Gravina 
Islands in which salmon can spawn, most of the tagged salmon caught in this region 
were evidently bound either for other localities in area 1 , or in one of the other areas. 
During the first part of the season considerable numbers of tagged salmon were 
recovered in the north arm of Behm Canal and in area 3. No doubt many of those 
recovered from the west shores of Gravina and Annette Islands during the first part 
of the season were bound for those regions. In the latter part of the season, however, 
very few were recovered from the north arm of Behm Canal and area 3, but there were 
still large recoveries from the west shores of Gravina and Annette Islands. Where 
were these tagged salmon going at this time of the season? 
In discussing the percentage of recoveries of tagged salmon in area 2 from the 
weekly taggings it was pointed out that there was a seasonal increase in these recov- 
eries and that larger numbers were reported from the localities in the vicinity of 
Kasaan in the last part of the season than in the first part. Hence, it is not at all 
improbable that many of the tagged salmon recovered from the west shores of 
Gravina and Annette Islands during the latter part of the season were bound for 
the localities in area 2 in the vicinity of Kasaan Bay and southward. These salmon 
no doubt migrated across Clarence Strait to the west shores of Gravina and Annette 
Islands and then turned back to the localities on the east shore of Prince of Wales 
Island in area 2. Whether or not the salmon make this journey may depend to some 
extent upon the prevailing winds. If a southeast wind is blowing there seems to be 
a greater tendency for the salmon to migrate northward along the east shore of 
Prince of Wales Island rather than to cross over to the west shores of Gravina and 
Annette Islands. If a southwest wind is blowing the opposite course is more apt to 
be taken. No definite conclusions may be drawn because of the lack of sufficient 
information, at the present time, concerning the influence of wind direction on 
migration. 
SUMMARY OF CAPE CHACON EXPERIMENTS 
1. The results from the tagging experiments in Clarence Strait, both past and 
present, indicate that the pink salmon migrating into the strait at different times of 
the season follow rather definite migratory routes to the localities for which 'they 
are bound. 
2. Most of the first pink salmon to appear each season are bound for localities 
in Revillagigedo Channel, the north arm of Behm Canal, Ernest Sound, or the 
northernmost regions in Clarence Strait. These salmon enter Clarence Strait by 
way of Cape Chacon and, after migrating for a short distance northward along the 
east shore of Prince of Wales Island, leave this shore and either follow directly up 
the middle of the strait or turn eastward until they reach the west shores of Annette 
and Gravina Islands, From here they either continue northward into the north 
