100 



I 



Fish 

 Shrimp 

 Euphausiids 

 Other 



< 200 



Fish length (mm) 



Figure 3. — Percentage by weight of major prey categories in the 

 diet of Pacific whiting, Merluccius productus, for different length 

 groups of whiting at various locations in 1980. 



481500 501-520 521540 541560 561580 581600 601620 

 Fish length (mm) 



FIGURE 4. — Percent frequency of occurrence of Pacific herring, 

 Clupca harcnaus pallasi, in stomachs of different length groups of 

 Pacific whiting, Merluccius productus, taken late summer 1980 off 

 Washington and Vancouver Island. 



appears to level off at a 30% occurrence of Pacific 

 herring. 



The difference in Pacific whiting diets between the 

 years 1967 and 1980 is mainly seen as a greater 

 percent consumption of fish in 1980 by whiting 

 >450 mm. No clear explanation is available because 

 sample size in 1980 was small and samples were 

 obtained further offshore than in 1967 . The most dis- 

 tinct relationship noted in these results is the ten- 

 dency of Pacific whiting to consume prey species 

 which occur in patches or schools, usually euphau- 

 siids and pelagic fishes such as eulachon, northern 

 anchovy, and Pacific herring which also prey on 

 euphausiids (Wailes 1936; Barraclough 1964; 

 Barraclough et al. 1968). Likewise, the majority of 

 the invertebrate prey of Pacific whiting (e.g., pan- 

 dalid, sergestid, and pasiphaeid shrimp) are 

 documented predators of euphausiids (Renfro and 



Pearcy 1966; Pearcy 1970; Judkins and Fleminger 

 1972). Thus, euphausiids appear to be the unifying 

 factor, their presence attracting not only the Pacific 

 whiting but also other organisms which then become 

 available prey for Pacific whiting. 



Alton and Nelson (1970) reported the main food 

 items of Pacific whiting off northern Oregon and 

 Washington to be euphausiids, eulachon, and pan- 

 dalid shrimp. Similarly, Outram and Haegele (1972) 

 found euphausiids, pandalid shrimp, and schooling 

 fish (such as Pacific sand lance, Ammodytes hexap- 

 terus; Pacific herring; and eulachon) most frequently 

 in stomachs of Pacific whiting sampled off Vancouver 

 Island. They also found Pacific herring to be a more 

 important component in the diet of large Pacific 

 whiting. Off California, Gotshall (1969a) discovered 

 pandalid shrimp, euphausiids, flatfish, and school- 

 ing fish constituted most of the whiting's food. 

 Although Gotshall (1969a, b) showed that on the 

 average pandalid shrimp dominated the Pacific whit- 

 ing diet, in some months fish or euphausiids were 

 found to be the primary food items. This seasonal 

 change in diet may have resulted from the changing 

 size composition of the Pacific whiting population in 

 the study area due to northward migration of adults 

 in spring. 



Predator-Prey Size Relationship 



An examination of the relationship between a fish 

 predator's size and its choice of prey sizes may aid in 

 understanding, interpreting, and quantifying a pre- 

 dator's feeding habits (Ursin 1973; Werner 1974; 

 Agger and Ursin 1976; Werner and Hall 1977; Hahm 

 and Langton 1980). Calculation of the frequency dis- 

 tribution of predator weight to prey weight ratios, as 

 described by Ursin (1973), quantifies a predator's 

 food size preference through central tendency 

 measures of its distribution. These measures can then 

 be used directly as input parameters to models such as 

 Andersen and Ursin's (1977) multispecies Beverton 

 and Holt model or the various multispecies VPA mod- 

 els (Helgason and Gislason 1979; Sparre 1980). 



The basic method of calculating predator-prey size 

 ratios utilizes the total weight in grams ( W) and total 

 number (UN) of each prey typey in a collection of pre- 

 dator stomachs. The individual mean weight of each 

 prey type (»)) is calculated and compared with the 

 mean predator weight (;7»,). Ursin (1973) discovered 

 that for benthic-feeding fish the frequency distribu- 

 tion of the ratios of predator weight to prey weight 

 (wjw) was approximately log-normal in shape. There- 

 fore, a plot of In (SJ/uJ,-) vs. E/V, should be a normally 

 shaped curve. 



632 



