MURPHY ET AL.: SIZE AND DIET OF JUVENILE PACIFIC SALMON 



about 0.5 m"^/second. Its watershed is forested by 

 mature western hemlock, Tsuga heterophylla , 

 and Sitka spruce, Picea sitchensis. Annually, 

 5,000-75,000 adult pink salmon and 200-4,000 

 chum salmon spawn in the creek from late July to 

 October, and 250-600 adult coho salmon spawn 

 from late September to November (Koski 1984). 

 Pink and chum salmon fry typically migrate from 

 Porcupine Creek from late March to mid-May 

 (Koski^). Coho salmon smolts migrate from late 

 April to early June, but over 90% usually migrate 

 in late May (Thedinga 1985). 



METHODS 



Six stations, one each on the east and west sides 

 of the three basins (Fig. 1) were sampled by a 

 beach seine 37 m long, with 1.6 cm stretch mesh 

 on the wings, and a central bag of 6 mm stretch 

 mesh. The seine tapered from 2 m deep at the 

 central bag to 1 m deep at each end. In 1979, only 

 one station in each basin was seined about every 

 4 days from 16 May to 12 June. In 1981, all six 

 stations were seined biweekly from 26 May to 

 7 July and monthly thereafter through 11 

 November. Seines were set parallel to and about 

 40 m from shore by a skiff, and retrieved from 

 shore. Setting and retrieval were accomplished 

 within 10 minutes. 



All fish caught were identified and counted. 

 Fork lengths (FL) were measured to the nearest 

 millimeter from a random sample of <25 salmon 

 per species, station, and sampling period. Stom- 

 ach contents were collected only in 1981 from <10 

 salmon per species and station in May, June, and 

 July. Contents were collected from anesthetized 

 fish by flushing the stomach with water from a 

 syringe (Meehan and Miller 1978; Koski and 

 Kirchhofer 1984) and preserved in 5% formalde- 

 hyde. Prey were later identified, counted, mea- 

 sured, and weighed. 



For diet analysis, the index of relative impor- 

 tance (IRI) was calculated, where 



IRI = {% number + % weight) (% frequency of 

 occurrence) 



(Pinkas et al. 1971). Diet overlap between salmon 

 species was calculated (McCabe et al. 1983): 



2 2 X'Y, 



C = 



1 = 1 



2xf . J^Yf 



1=1 



1 = 1 



2K V. Koski, Northwest and Alaska Fisheries Center Auke 

 Bay Laboratory, National Marine Fisheries Service, NOAA, 

 P.O. Box 210155, Auke Bay, AK 99821, pers. commun. October 

 1985. 



where C = overlap coefficient and X, and Y, are 

 proportions of the total diet of salmon species X 

 and Y, respectively, contributed by prey taxon i of 

 s prey taxa. Diet overlap was calculated sepa- 

 rately for proportions based on prey number and 

 weight. Prey were also classified as epibenthic or 

 pelagic to assess overlap in foraging mode (Feller 

 and Kaczynski 1975). Epibenthic prey were poly- 

 chaetes, gammarid amphipods, harpacticoid cope- 

 pods, barnacle cyprids, and cumaceans. Pelagic 

 prey were calanoid copepods, euphausiids, barna- 

 cle nauplii, cladocerans, larvaceans, larval deca- 

 pods (Brachyrhyncha), hyperiid amphipods, and 

 fish (eggs, larvae, and juveniles). 



RESULTS 



Size 



In May 1979, pink salmon were the size of 

 newly emergent fry, about 32 mm FL (Fig. 2). 

 Average length increased 1.5 mm/day, to 73 mm 

 on 12 June 1979. In 1981, pink salmon averaged 

 73 mm FL in late May and early June. Changes 

 in average FL in 1981 could not be calculated 

 because most migration occurred before sampling 

 began. 



Average FL of chum salmon increased slower 

 than that of pink salmon. Mean FL of chum 

 salmon increased 0.4 mm/day in both years, from 

 39 mm to 51 mm in 1979, and from 60 mm to 78 

 mm in 1981 (Fig. 2). Chum salmon averaged 

 about 10 mm FL longer in 1981 than in the same 

 period in 1979. Chum salmon were not found in 

 the estuary after early July, except for two fry 

 caught in the outer basin in November. 



Average FL of coho salmon was nearly con- 

 stant, between 85 and 110 mm, throughout May 

 and early June in both 1979 and 1981 (Fig. 2). 

 Average FL of coho salmon in the estuary during 

 this period was influenced by an influx of Porcu- 

 pine Creek migrants, which averaged between 75 

 and 96 mm FL (Thedinga 1985). After the migra- 

 tion from Porcupine Creek in 1981 (=9 June), 

 average FL increased 1.6 mm/day to 165 mm by 

 20 July. Average FL then decreased to 85 mm in 



215 



