104 U.S. BUEEAU OF FISHERIES 



daily catches were available, hence a fishing day provided a constant 

 unit of time. From these data the result or catch of a constant unit 

 of fishing effort was calculated from the records of each trap, and from 

 the records of all of the traps for each year or portion of a fishing 

 season. Seasonal trends of the catches were also computed for the 

 individual years. 



Using standard methods for computing index numbers, the records 

 of the individual traps were combined into an index of size of catch 

 per unit of fishing effort for all of the traps during the years in which 

 data were available. Since these catches are the results of the use 

 of constant units of fishing gear it should be a reliable index of the 

 abundance of the Puget wSound sockeye salmon. The total amount of 

 fishing gear on Puget Sound has decreased during the years under 

 consideration. Therefore, any drop in average catch per unit of 

 effort is not due to increased competition between gear. The trend 

 of this index indicates a steady decline in the numbers of sockeye 

 salmon from 1917 to 1930. 



An outstanding feature of the seasonal trend of the fishery is a large 

 peak in midsummer. This period of great abundance has apparently 

 suffered the greatest depletion. 



From the results thus far achieved it appears that it may be possi- 

 ble to correlate certain regular seasonal fluctuations in abundance 

 with definite races of salmon entering the Fraser River. 



Catch records from the fishing seasons of 1931 and 1932 have only 

 recently become available and will be added to the data so far con- 

 sidered to complete this study. 



During 1929 an experiment was undertaken at the Bureau of Fish- 

 eries' hatchery at Birdsview, Wash., to determine the most effective 

 time at which to liberate young sockeye salmon. With this object 

 in view 24,000 fingerlings were liberated in the fall of 1929 after being 

 marked by the removal of their left ventral fins. On the following 

 spring another 24,000 fingerlings from the same brood were Hberated 

 and marked by the removal of the right ventral fins. 



Returns from these experiments were secured during the fishing 

 season of 1932. Approximately 600 fish were recovered, the spring 

 liberation returning in a ratio of about 4 to 1 to the fall liberation. 

 Full data regarding length, locality, and date of capture, and scale 

 samples were taken from these fish. 



All of the marked fish recovered after entering the Skagit River were 

 found in Grandy Creek where they were liberated, none being taken 

 at the weir on Baker River. This adds additional confirmatory evi- 

 dence to the theory that sockeye salmon return to their parent stream 

 to spawn. 



These data will be considered for the purpose of not only determin- 

 ing the proper time of liberation, but also of providing some infor- 

 mation on the rate of mortality of these fish in the ocean and the 

 importance of the Skagit River sockeye-salmon run to the commercial 

 fishery of Puget Sound. 



KARLUK RED-SALMON INVESTIGATION 



The scientific investigations which have been carried on for a 

 number of years at Karluk River were continued during 1932. Scale 

 samples were collected throughout the season for the purpose of 



