HARGREAVES: PREY PREFERENCE OF PREDATORS 



a 2.54 cm (1 inch) diameter Quality Control Device 

 (QCD), manulactured by Northwest Marine Technol- 

 ogy^ (Shaw Island, WA 98286). Samples of tagged 

 fish were also dissected and examined visually and 

 microscopically to verify tag implantation. Three 

 thousand of these pink salmon were also marked 

 externally by amputating either the left or right pec- 

 toral fin. No other fins were amputated. The pur- 

 pose of the fin clipping was to determine which 

 method (nose tag or fin clip) was more effective for 

 identifying fish recovered from the stomachs of 

 predators. 



After tagging and marking, all pink salmon were 

 transferred back into the small, saltwater enclo- 

 sures. On 20 April the three thousand tagged but 

 unmarked pink salmon were transferred to a larger 

 (51 m'^) enclosure anchored near the southwest end 

 of Masset Inlet, and held until required for the sec- 

 ond experiment. This enclosure was shallowed by 

 hand once each week to check the condition of the 

 fish and cleaned with a high-pressure hose every 

 10-14 days to remove algae growing in the meshes. 

 The food supply of the pink salmon held in this 

 enclosure was not controlled and consisted of what- 

 ever came through the meshes. No supplementary 

 food was added. Under this regime the pinks re- 

 mained very active and appeared healthy. 



The first experiment was initiated at 1600 on 26 

 April, by releasing 3,000 tagged and fin-clipped pink 

 salmon at Marinelli Point. A sample of 99 fish was 

 removed, anesthetized with 2-phenoxyethanol, and 

 the live fork length of each fish measured to deter- 

 mine the size distribution of fish at the time of 

 release. A single beach seine set was made three 

 hours later, in the immediate vicinity of the release 

 site, to collect a sample of potential predators. 



The second experiment was initiated by releasing 

 1.800 tagged pink salmon at 1130 on 6 June, into 

 the boat harbor at Masset (Fig. 1). None of these 

 fish were fin-clipped. A sample of 100 pink salmon 

 was removed prior to the release, and each fish 

 measured to determine the size distribution of fish 

 at the time of release. Two beach seine sets were 

 made to collect potential predators prior to the time 

 of release, the first at 0930 and the second at 1015. 

 A total of sixteen additional sets were made after 

 the release, seven between 1245 and 1826 on 6 June, 

 and nine more between 0900 and 1400 the follow- 

 ing day. 



The beach seine used to capture potential pred- 

 ators was 46 m long and constructed of 6.4 mm 

 stretched nylon mesh. All potential predators were 

 examined immediately after capture. Each mea- 

 sured fish was anesthetized in 2-phenoxyethanol, 

 and the stomach contents obtained by either 

 hydraulic flushing or dissection. All fish remains in 

 the stomach contents were examined visually or 

 microscopically to identify prey to the species level, 

 using Hart (1973) as a general reference and Phillips 

 (1977) to identify juvenile salmon. All juvenile 

 salmon found in the stomachs of predators and any 

 live juvenile salmon captured along with the pred- 

 ators were measured (if possible), examined for 

 missing fins, and passed through the QCD several 

 times to determine if they were tagged. 



Yearling coho salmon, Oncorhynchus kisutch, cap- 

 tured on the first day of the second experiment were 

 used to obtain an estimate of the total population 

 of coho salmon in the vicinity of the release site. All 

 coho salmon captured on 6 June were retained, 

 anesthetized with 2-phenoxyethanol, and marked 

 using a hot-wire branding device. At 2000 on 6 June, 

 these coho were sorted to remove any that did not 

 appear healthy and fully recovered, and the remain- 

 ing 170 were released. All coho captured the follow- 

 ing day were inspected to determine if they were 

 marked. The numbers of marked and unmarked 

 coho were used to derive a simple (single census) 

 Petersen estimate of the total population of year- 

 ling coho salmon in the vicinity of the release site 

 using the following equation (Ricker 1975): 



A^ 



{M + 1) (C + 1) 

 (R + 1) 



(1) 



^Reference to trade name does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



where A^ = total number of coho salmon at time 

 of marking 



M = number of coho salmon marked and 

 released on 6 June 



C = total number of coho salmon captured 

 on 7 June 



R = number of marked coho salmon recap- 

 tured on 7 June. 



The 95% confidence interval for the population 

 estimate was obtained by substituting into this equa- 

 tion the fiducial limits of the number of recaptured 

 coho salmon, from tables of the Poisson distribution 

 (Ricker 1975). These figures were used to estimate 

 the total number of tagged pink salmon eaten by 

 coho salmon in the second experiment, using the 

 equation 



765 



