MIGRATIONS OF PINK SALMON 
661 
arm of Bekm Canal and the northern regions of Clarence Strait, including Ernest 
Sound, or they continue eastward and southward into Revillagigedo Channel. 
3. This peculiar migration of the salmon along the west shores of Annette and 
Gravina Islands results in many of the salmon, bound for the north arm of Bekm 
Canal and the northern regions of Clarence Strait, being intercepted by the traps 
and purse seines operated along these shores. 
4. As the season progresses, more and more of the pink salmon entering 
Clarence Strait by way of Cape Chacon arc destined for the localities along the east 
shore of Prince of Wales Island below Approach Point. Most of these salmon con- 
tinue northward along the east shore of Prince of Wales Island instead of crossing 
the strait to the west shores of Annette and Gravina Islands. Many of those that 
do cross the strait turn back from the shores of these islands to the localities on the 
east shore of Prince of Wales Island. 
5. At the very close of the season many of the salmon migrating into Clarence 
Strait are bound for localities in Cordova Bay on the extreme southwest shore of 
Prince of Wales Island. 
The migration of the pink salmon along the west shores of Annette and Gravina 
Islands, especially the latter, makes these shores one of the most productive fishing 
areas in Clarence Strait. The fishing gear operated in this area intercepts the runs 
of pink salmon that are bound for practically all of the localities in Clarence Strait 
and its adjacent waters to the east. It is for this reason that so many of the pink 
salmon tagged in Clarence Strait were recovered along the west shores of these 
islands. 
PINK-SALMON TAGGING EXPERIMENTS IN SUMNER STRAIT, 1924-36 
It has long been known that the run of pink salmon in Sumner Strait appears 
earlier in the summer than the run in Clarence Strait. The time the salmon ap- 
peared in the commercial catches and the location of these catches in the strait 
indicated that most of these early migrants were bound for localities in the extreme 
eastern section and the adjoining waters of Zimovia Strait, Eastern Passage, Blake 
Channel, and Bradfield Canal. However, the extent to which this run penetrated 
the waters of Clarence Strait and other adjoining channels was not known. In 
order to determine more completely the distribution of the localities in which these 
pink salmon spawned, the Bureau laid plans for a number of tagging experiments to 
be carried on in various parts of Sumner Strait. This work began in 1924, was 
continued each year through 1927, and taken up again in 1935 and 1936. Since 
only one experiment was carried on in each of the latter seasons, their results will be 
discussed with those from the early experiments. 
The locations where the salmon were tagged in Sumner Strait from 1924 to 
1936 are shown in figure 1. The dots indicate the locations of early taggings and 
the triangle is that of the later ones. The early tagging was done at Cape Decision, 
Ruins Point, and Point Colpoys, and during 1935 and 1936 only at Point Colpoys. 
Although none of these experiments continued throughout the entire season of any 
year, they varied sufficiently in point of time so that, taken together, they give a 
picture of the movements of fish in this region over an entire season. Thus, experi- 
ments were carried on at Cape Decision and Rums Point in 1924, 1925, and 1927, 
between July 12 and August 3; those at Point Colpoys on July 10, 1926, July 26 to 
30, 1927, August 13, 1935, and August 16, 1936. A summary by geographic areas 
of the results from all experiments from 1924 to 1936 is given in table 7. The indi- 
vidual localities in the geographic areas, in which the tagged salmon were recovered 
