FISHERY BULLETIN: VOL. 86, NO. 2 



E 80 



E 



O 60 



u 

 -» 50 



Z 

 O 40 



< 30 

 0) 



ni 





._^k, ^^ (^ CHUM 



220 

 180 

 140 

 100 

 60 



(c) COHO 



3-^-tH^4 



>- 



4—.^ 



/ 



/ 



"n 1 1 1 1 r 



15 MAY 1 JUN 15 JUN 1 JUL 15 JUL 1 AUG 



Figure 2.— Length of salmon in 1979 and 1981. Data shown are 

 means and ranges for pooled samples from all stations on each 

 sampling date. Data for pink salmon in 1981 are omitted be- 

 cause of small sample sizes. (▲ = 1979, • = 1981.) 



August and 106 mm in September, after most 

 smelts had left and a few new smelts entered the 

 estuary. 



Diet 



A wide variety ef prey was eaten by the three 

 salmon species, but usually only one or two prey 

 taxa dominated the diet (Table 1). Pink salmon 

 ate mostly larval molluscs (Mesogastropoda) and 

 larval fish (mostly Gadidae) in May, and calanoid 

 copepods in June. Chum salmon ate mostly larval 

 molluscs in May; larval molluscs, larvaceans, and 

 cladocerans in June; and hyperiid amphipods and 

 larval decapods in July. Coho salmon ate mostly 

 fish and insects in May and June, and fish and 

 larval decapods in July. The identifiable fish prey 

 of coho salmon consisted of 53% Pacific herring, 

 Clupea harengus pallasi; 45% cod (Gadidae); and 



2% Pacific sand lance, Ammodytes hexapterus. No 

 identifiable pink or chum salmon were in the coho 

 stomachs. Catch of coho, but not that of the other 

 salmon, was significantly correlated (r = 0.46, 

 P < 0.001) with aggregate catch of herring, sand 

 lance, and cod, indicating that coho salmon con- 

 gregated near prey schools. 



Diet overlap was higher between pink and 

 chum salmon than between either species and 

 coho salmon (Table 2). Diet overlap between pink 

 and chum salmon was especially high in May 

 when both species ate large numbers of larval 

 molluscs. If based on prey weight, diet overlap 

 between pink and coho salmon was negligible. If 

 based on prey number, however, overlap was 

 >50% in June when both pink and coho salmon 

 ate large numbers of calanoid copepods. Diet 

 overlap between chum and coho salmon was con- 

 sistently low, especially when based on prey 

 weight. 



Of the 12 most important prey taxa in May and 

 June, when all 3 salmon species were present in 

 the estuary, only 4 differed significantly 

 iP < 0.05) in mean number per stomach between 

 pink and chum salmon, whereas 9-10 differed 

 significantly between the two species and coho 

 salmon (Table 3). Compared with pink salmon, 

 chum salmon ate more harpacticoid copepods, 

 cladocerans, and insects. Coho salmon ate fewer 

 small plankton and more fish than did the other 

 salmon species. Coho salmon averaged fewer than 

 20 total prey items, compared to more than 100 in 

 pink salmon and 200 in chum salmon. 



Coho salmon ate larger prey than did the other 

 salmon (Fig. 3). Median prey length for coho 

 salmon was 2.3 mm, compared with 0.4 mm for 

 pink and chum salmon. Coho salmon generally 

 selected larger individuals of each prey taxon — 

 particularly larger calanoid copepods, gammarid 

 amphipods, euphausiids, and larval decapods — 

 than did pink and chum salmon (Table 4). Offish 

 prey, coho salmon ate mostly juveniles, whereas 

 pink and chum salmon ate mostly eggs and lar- 

 vae. 



As they grew larger, all three salmon species 

 selected larger prey. Numbers of hyperiid am- 

 phipods, euphausiids, and fish larvae — all rela- 

 tively large prey — were positively correlated 

 with FL of pink or chum salmon, whereas num- 

 bers of cladocerans and larvaceans — both rela- 

 tively small prey — were negatively correlated 

 with chum salmon FL (Table 5). Numbers of 

 calanoid copepods and fish were positively corre- 

 lated with coho salmon FL, whereas the number 



216 



