Fishery Bulletin 97(3), 1999 



Chinook salmon stomach contents were removed 

 from the digestive tract, fixed in 109f formaUn, trans- 

 ferred to 40% isopropyl alcohol, and sorted under a dis- 

 secting microscope into major taxonomic groups. When 

 possible, stomach contents were identified to species. 



The best measure of dietary importance is one 

 where both the number and weight of a food category 

 are recorded (Hyslop, 1980) . Stomach content data 

 were summarized by four methods: 1) numerical, 2) 

 gravimetric, 3) frequency of occurrence, and 4) in- 

 dex of importance. Each food group was enumerated 

 and weighed from each stomach. Wet weight of prey 

 items was measured to the nearest 0.01 g. By using 

 these data, percentage of total number i'/cN), per- 

 centage of total weight (%W), and percentage fre- 

 quency of occurrence C^'AF) were calculated for each 

 food group. "Index of importance" (lOI) was calcu- 

 lated for each food group as follows: 



lOI. 



100 X HI ^, 

 I." HI 



where ///, 



17 



7(F + 7(W for food gi'oup a; and 

 the number of different food gi'oups 

 (Hannah, 1980; Grav et al., 1997). 



Figure 1 



Humlxildl Bav, Trinidad liav. and Crescent City. California. 



Because this index calculation is based on %F as well 

 as V( W, the bias towards heavier, infrequently found 

 prey items is reduced. 



Results 



Values for %N, %F, % W, and lOI for prey items en- 

 countered in the stomach analyses are listed in 

 Table 1. All stomachs examined (308) contained food 

 items except one from 1994 and two from 1995. Each 

 prey item was present in stomachs sampled from both 



1994 and 1995 except for octopi iOctopus riiheNcens) 

 (2), jacksmelt (Atherinopsis callfornlensis) ( 1 ), cottids 

 (1), pleuronectids (1), Pacific sandlance (108), and 

 rockfishes (45), which were observed only in 1994 

 samples and isopods (3) which were present only in 



1995 samples. Total values for %F did not equal 100% 

 owing to unidentifiable prey items in the diet, espe- 

 cially in 1995 (Table 1). 



In 1994, the lOI indicated that euphausiids were 

 the predominant food item, accounting for over 27%- 

 of the total. Euphausiids not only ranked highest by 

 %N and %F but also were the leading prey item by 

 % W. In addition, notable lOI values were observed 

 for crab megalopae. Pacific herring, surf smelt 

 (Hypomesiis pretiosiis). Pacific sandlance, northern 

 anchovy, night smelt (Spii-inchus starksi). and squid 

 {Loligo opaleHceus). Infrequently encountered prey 

 items included Pacific saury iCololabis saira), rock- 

 fishes, amphipods, jacksmelt, octopi, shrimp (mysid), 

 juvenile pleuronectids, and juvenile cottids. In 1995, 

 chinook consumed primarily northern anchovy, which 

 represented over 33% of the total lOI but also preyed 

 upon Pacific herring, squid. Pacific saury, surf smelt, 

 night smelt, euphausiids, and crab megalopae. Only 

 rarely were amphipods, isopods, and shrimp found 

 in stomachs. Large interannual variations in lOI can 

 be seen for euphausiids, crab megalopae. Pacific 

 sandlance, northern anchovy, squid, and Pacific saury. 



Seasonal variation of dominant prey items, for 

 1994 and 1995, is illustrated in Figure 2, A and B, 

 respectively. In 1994, 88 stomachs were examined 

 from May and June (late spring). lOI values for eu- 

 phausiids (34%), crab megalopae i259r }, and Pacific 

 herring (17% ) dominated all other prey items. Late 

 summer lOI values ( based on 108 stomachs acquired 

 in August and September 1994) indicated Pacific 

 sandlance (22*"^ ), surf smelt (21%), northern anchovy 

 ( 17% ), and euphausiids ( 14% ) to be major prey items. 

 In 1995, 26 stomachs were examined from fish col- 

 lected in June and 86 from September. The lOI val- 

 ues in our study showed that squid (45% ), surf smelt 

 (25%), euphausiids (19%), and Pacific herring ( 11%) 

 are important prey items in late spring whereas 



