Canadian area C is arbitrarily included in the Puget Sound 

 district on the assumption that most of the catches reported 

 there were made north of La Push. Data reported for 

 Washington fishermen north of Barkley Sound and for the 

 following categories were not considered due to their geographic 

 dispersal and relative insignificance: 1) Tillamook Head to Cape 

 Elizabeth, 2) Tillamook Head to Barkley Sound, 3) Cape 

 Elizabeth to Barkley Sound. These data were included when 

 calculating statewide totals. Catches of pink salmon south of 

 Cape Johnson were insignificant and were not included in the 

 analyses. Because of the difficulty in standardizing effort for 

 Washington and Canadian trollers, data for each group were 

 analyzed separately. Sport catch was also analyzed separately 

 and is referred to by port of origin (Ilwaco, Westport, La Push, 

 Neah Bay). 



Because the ocean hook-and-line fishery is noncompetitive 

 according to Ricker's (1975) criteria, then theoretically CPUE is 

 a more valid index of abundance than is catch; but, in practice, 

 there are limitations to the use of sport CPUE and Washington 

 troll CPUE for this purpose. A restrictive (three salmon per day) 

 ocean bag limit for Washington fishermen was in force during 

 the period considered and the number of salmon caught per 

 angler trip has little quantitative meaning. The overall sport 

 catch is a better index of salmon abundance in this application. 

 Wright (1970) discussed the problems associated with the use of 

 Washington troll CPUE as an index of ocean salmon abun- 

 dance, most of which stem from the vague dimensions of the 

 landing as a unit of effort. For the Washington troll fishery, 

 effects of various factors on both catch and CPUE were 

 examined. Trends in Canadian troll CPUE have closely parallel- 

 ed those in troll catch, particularly for chinook and coho 

 salmon. In trend analysis of the Canadian fishery, it therefore 

 makes little difference whether catch or CPUE is the dependent 

 variable. 



The analytical procedure employed was linear regression, with 

 1) commercial ocean troll catch, 2) troll CPUE, or 3) ocean 

 sport catch in particular areas during specified periods being the 

 dependent variables. All regressions were calculated with Bio- 

 Med computer program BMD02R (Dixon 1968). When the 

 regression equation included an index of brood year abundance 

 as an independent variable, we used a log transformation based 

 on Ricker's spawner-recruit relationship, i.e., 



In J? 



In a - bS 



InS, 



where R is troll catch, etc. in year ;' and S is the index of 

 abundance in the brood year i—x, x being the age in years at 

 maturity. Ricker (1975) discussed appropriate data transforma- 

 tions for exploratory correlations involving physical factors. In 

 preliminary work, we found ranking to be the only useful altera- 

 tion when dealing with environmental variables. 



Pink Salmon 



Indices of brood year abundance. — Several indices (Table 1) 

 were based on net gear catch or CPUE for each of four major 

 stocks: 1) the Washington run, 2) the Canadian non-Fraser run, 

 3) the early Fraser run, and 4) the late Fraser run. For each, we 

 summed net gear catch and effort as indicated below, based on 

 information in Hourston et al. (1965) and Ward (1958). 



Washington run: 1) one-half of the net catch (or effort) in 

 Canadian area 20 during 15 July through 15 August; 2) one-half 



of the United States and Canadian net catch (or effort) in the 

 Strait of Juan de Fuca during 15 July through 15 August; 3) 

 annual trap, drag seine, and set net catch (or effort) in Puget 

 Sound; 4) annual net catch (or effort) reported for Admiralty 

 Inlet, West Beach, Skagit Bay, Port Susan, Port Gardner, and 

 Bellingham Bay, exclusive of that in the previous category. 



Canadian non-Fraser run: 1) and 2) as above; 3) net catch 

 (or effort) at Point Roberts, Rosario, Salmon Banks, and San 

 Juan Islands during 1 August through 15 August. 



Early Fraser run: 1) Canadian net catch (or effort) in 

 Canadian area 20 during 16 August through 31 August; 2) 

 United States and Canadian net catch (or effort) in the Strait of 

 Juan de Fuca during 16 August through 31 August; 3) net catch 

 (or effort) at Discovery Bay, the San Juan Islands, Rosario, and 

 the Salmon Banks during 16 August through 6 September; 4) 

 net catch (or effort) at Point Roberts during 16 August through 

 9 September; 5) net catch (or effort) in the Fraser River during 1 

 September through 20 September. 



Late Fraser run: 1) Canadian net catch (or effort) in Cana- 

 dian area 20 during September; 2) United States and Canadian 

 net catch (or effort) in the Strait of Juan de Fuca during 

 September; 3) net catch (or effort) at Discovery Bay, the San 

 Juan Islands, Rosario, and the Salmon Banks during 7 

 September through 15 October; 4) net catch (or effort) at Point 

 Roberts during 10 September through 15 October; 5) net catch 

 (or effort) in the Fraser River during 21 September through 20 

 October. 



To calculate CPUE for net gear, we standardized effort in 

 Washington gill net landings. Within an area, CPUE for each 

 EC 



gear type was calculated as 



£/ 



where C was catch in fish and 



/was effort in landings. Then an efficiency factor £ of 1.00 

 was assigned to gill nets. Efficiency factors for other gears (1, 



2 . . . ) were calculated as E = M — ), where N was CPUE for 



G 

 a gear type and G was CPUE for gill nets. Then the total stan- 

 dardized effort / std for a particular year in a specific area was 

 calculated as: 



/ 



si J 



/ 



gill net 



E,f, + E 2 f 2 



We calculated catch per standardized unit of effort as the total 

 catch for all net gears divided by the total standardized effort. 

 Efficiency factors are shown below. If CPUE for gill nets and 

 another gear was not significantly different, catch and effort 

 were added for both types. 



Area 



Gear 





E 



Puget Sound 



Gill net and set 



net 



1.00 





Drag seine 





4.13 





Indian trap 





16.29 





Purse seine 





10.84 





Reef net 





2.39 



Area 20 



Canadian gill net 





1.29 





Canadian purse seine 



23.89 



Predator-Prey relationship. — If the following assumptions 

 are met: 1) abundance of yearling coho salmon is proportional 

 to the number of hatchery yearlings released that year and 2) in- 

 creased numbers of yearling coho salmon contribute to reduced 



